STRIKING GOLD - celebrating over 50 years of the Society of Technical Analysts

Page 1

The Society of Technical Analysts The Best of the STA Journal

STRIKING GOLD Celebrating over 50 years of the Society of Technical Analysts


First Published in 2022 by Society of Technical Analysts Paperback Edition Copyright © The Society of Technical Analysts ISBN 978-1-3999-3292-9 All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage retrieval system, without permission in writing from the publisher All articles were previously published in the STA Journal Design by Andrew Convery Printed in the UK


To all past, present and future members of the Society of Technical Analysts.


4

The UK’s professional body for Technical Analysts. Founded in 1968. The oldest of its kind in the world. The Society of Technical Analysts Dean House Vernham Dean Andover SP11 0JZ tel: +44 (0) 20 7125 0038 info@technicalanalysts.com www.technicalanalysts.com


5

8 60 98 134 186

CHAPTER ONE

Theory of Technical Analysis 08 to 33

CHAPTER THREE

Charting Types, Point & Figure & Candlesticks 60 to 79

CHAPTER FIVE

Moving Averages and Trends 98 to 117

CHAPTER SEVEN

Elliott Wave and Fibonacci 134 to 161

CHAPTER NINE

Psychology and Markets 186 to 201

34 80 118 162 202

CHAPTER T WO

Dow Theory, Wyckoff and Volume 34 to 59

CHAPTER FOUR

Pattern Recognition and Pattern Analysis 80 to 97

CHAPTER SIX

Indicators and Momentum 118 to 133

CHAPTER EIGHT

Gann Analysis, Cycles and Forecasting 162 to 185

CHAPTER TEN

Systematic Trading 202 to 217

Please note: The job titles and affiliations of the authors featured at the end of the articles were applicable at the time of original publication and as such may not be relevant today. Disclaimer: The Society is not responsible for any material published in this book and publication of any material or expression of opinions does not necessarily imply that the Society agrees with them. The Society is not authorised to conduct investment business and does not provide investment advice or recommendations. Articles are published without responsibility on the part of the Society, the editor or authors for loss occasioned by any person acting or refraining from action as a result of any view expressed therein.


6

Foreword The term ‘Technical Analysis’ means different things to different people: partly, it depends on whether the objective of the analysis is short-term trading or longer-term investment; but partly, too, it depends on training and experience. There is, nevertheless, a basic understanding about the body of knowledge that is known as technical analysis. This is that financial markets (and, indeed, economic indices) are not just random fluctuations within an uncertain environment. They are non-random and coherent phenomena that have the potential to generate observable entry and exit signals. There is, in other words, a functional order behind the apparent chaos. This means that it is entirely possible to offer more than just a subjective opinion on where a price or an index might be heading. An analyst who concentrates on the long-term has the power to decide whether or not a market is (1) in a trend; (2) undergoing a correction within that trend; and/or (3) undertaking a full trend reversal. Meanwhile, an analyst who focuses on the short-term (and who may not wish to include long-term trend considerations in his or her deliberations) will have a high degree of confidence about the likely outcome from a buy or sell signal. It seems extraordinary that as little as 50 years ago technical analysis was generally regarded as being a fringe pursuit. It might have been accepted that people would look at a price-time chart of a particular market before making a trading or investment decision, but such a chart was to be regarded as being just a picture of history. It was possible to see where a financial price stood in relation to that history, but there was very little appreciation of the fact that price behaviour might actually be telling you something about the future. It has taken the dedication of a relatively small group of people (some of whom are included in this book) to generate the background research that now informs the discipline. One line of analysis has been the role of crowd behaviour in the regular patterns and trends that dominate financial markets and economic indices. A great deal of research now confirms that crowd behaviour intensifies during the evolution of a trend. Initially, so-called ‘diversity generators’ spot practical investment opportunities but, as the trend progresses, an increasing number of ‘conformity enforcement’ techniques keeps the herd either bullish or bearish. Such techniques include propaganda through the mainstream media, physical proximity, and the choice of associates. The trend, in itself, is a reflection of competition between the bulls and the bears; but, eventually, the winning group will become overstretched. Once this happens, it takes very little to create a reversal. Reversals can, of course, be just corrections within an ongoing trend, or an actual reversal in the trend. I suspect that the latter involves either a reversal in whatever it is that constitutes fundamentals or in some form of an offset to those fundamentals. This is where the idea of boundaries is important. Many years ago, a number of technical analysts found that the Golden Ratio (38.2:61.8) played an important role in defining these boundaries. Specifically, markets tended to retain 61.8% of their previous impulse waves, and so could retrace 38.2% of any prior advance without triggering a full reversal. This idea can be used to distinguish between a correction and a new trend. As if this were not enough, the situation is complicated by the processes of evolution. First, a system of financial behaviour or economic activity has to recognise that something in its environment has changed. Second, the old system (including its beliefs and infrastructure) has to be downgraded in some way. Third, a new system (in terms of ideas and physical infrastructure)

has to be put in place. And fourth, this new system has to persist until the whole process starts again. All of this takes time - hence the development and persistence of trends. Moreover, evolutionary forces introduce a natural element of expansion and contraction hence the presence of cyclical behaviour. It doesn’t take a genius to recognise that the interaction of internal forces (i.e., forces that are specific to a particular index or market), and external forces (such as those that involve politics), can result in a very complex environment. One approach has been subjectively to simplify the complexity and to use models that are considered appropriate. Another approach, however, is to adopt the idea that a market or index already has all the relevant information contained within itself. This is the realm of technical analysis. It automatically simplifies a very complex environment. The task of the analyst is then to tease out the pertinent information from a specific chart, and use that information to draw conclusions about future probabilities. It is not possible to mention all the contributions that have been made over the years. Names such as Charles Dow, Ralph Elliott, and William Gann are now recognised with an immediacy that was not possible half a century ago. The point, however, is that progress has usually been based on the idea that what matters is trading and investment profitability, not the nicety of the theory. This is why a specific technical analysis tool can ‘work’, even though there is - at the time - no obvious reason for it. It is one of the great truths of technical analysis that it has accessed aspects of reality that have not been considered relevant by conventional thinkers. It has, in this sense, been in the vanguard of change. Technical analysis has accordingly pursued a variety of concepts: the idea that simple line charts might be complemented by phenomena such as bar charts, candle charts, and point and figure charts; the notion that various indicators (investor and marketgenerated) can be used to confirm a market’s progress; and the idea that collective human behaviour is influenced by its own past history, by internal cycles and rhythms, and by external influences such as the planets and natural geophysical forces. Above all there is the idea that specific patterns repeat themselves, often regularly and always frequently. So, when they emerge, these patterns can be used to anticipate future behaviour. Many ideas that might originally have been considered somewhat obscure have now been sufficiently progressed to become taken for granted. The research on these phenomena has been profound, and David Watts has undertaken the difficult task of collating that research, as published in the official journal of the Society of Technical Analysts - the Market Technician. The result is ten chapters, each of which deals with one of the main subject areas: 01

Theory of Technical Analysis

02

Dow Theory, Wyckoff and Volume

03

Charting Types, Point & Figure & Candlesticks

04

Pattern recognition and Pattern Analysis

05

Moving Averages and Trends

06

Indicators and Momentum

07

Elliott Wave and Fibonacci

08

Gann Analysis, Cycles and Forecasting

09

Psychology and Markets

10

Systematic Trading

This means that the book can be seen, not just as a history of progress within the Society of Technical Analysts, but also as an upto-date reference work. Readers can therefore either read the whole book sequentially, or delve into specific chapters as they need.

By Tony Plummer, FSTA


7

Introduction to the STA’s book of technical analysis articles Deborah Owen

The Society of Technical Analysts is marking its 50th anniversary by publishing some of the seminal articles that have appeared in its Journal over the years. The financial markets are rather like a tug-of-war competition in that, in any market, there will be buyers pulling the price up and sellers pushing it down. At any one point in time these conflicting pressures will be resolved into a single price. Academic studies have shown that prices tend to move in trends more often and for longer than could possibly be explained away by the laws of chance. Technical analysis is the study of past trends and patterns with the objective of predicting future price movements. It has long been axiomatic that markets reflect human behaviour and traders have been searching for a reliable way to accurately and consistently tap into the undercurrent of market sentiment since joint-stock companies were first formed in the 17th century. It is thought that Japanese rice traders in the 18th century used a form of technical analysis, known as candlestick charts, to plot movements in the price. In the 19th century, as a result of their (separate) systematic observation of historical price data, Charles Dow, Ralph Nelson Elliott and W.D. Gann each formulated a set of precepts which have, collectively, become the bedrock of technical analysis. For years, technical analysis was regarded as a fringe form of analysis by academics who espoused the Efficient Market Hypothesis (EMH) as the most convincing explanation of how markets behave. However, the 2007-09 financial crisis completely discredited the EMH as a model for describing market behaviour; markets are not perfectly efficient and decision-making by individuals cannot be assumed to be always rational. In the subsequent re-evaluation of market analysis, the ideas and tools of technical analysis have acquired a much wider following. It is a measure of the extent to which technical analysis has moved into mainstream financial analysis that the STA’s 50th Anniversary book includes articles from academics and a broad spectrum of market participants ranging from short term day traders to long term portfolio managers. The first issue of the Society’s Journal, Market Technician, was published in 1988. It was designed to provide a forum for technical analysts to put forward new ideas and exchange views on market trends and developments. I edited the Journal for 25 years and during that period of time witnessed a revolution in computer technology as well as a radical transformation in the financial

landscape. The core elements of technical analysis have stood up well to this changing environment. The collection of articles in this book represent not only the breadth of the subject but also the way that the discipline has evolved to encompass new ideas and technological developments. The underpinning concepts of technical analysis are covered in the first chapter, Theory of Technical Analysis. Charts are the starting point of all types of technical analysis and, when the Journal was first published, they were all drawn by hand. The articles included in chapter 3 elaborate on the techniques involved in Point and Figure charts, Candlesticks, Heikin-Ashi and Market Profile. Moving averages and indicators are the basic cornerstones of momentum analysis and these subjects are covered in chapters 5 and 6. Chapter 4 expands on the core concepts of pattern recognition. The articles in chapters 2, 7 and 8 have been selected because they build on the ideas encapsulated in Dow Theory, Fibonacci numbers, Elliott Wave analysis and Gann Theory. Extraordinary Popular Delusions and the Madness of Crowds published by Charles Mackay in 1841 is required reading for anyone who wants to understand the mob-psychology of markets and the articles in Chapter 9 focus on this important aspect of trading. Chapter 10 deals with the concept of systematic trading or money management. The role of editing the Journal was a stimulating and, at times, challenging exercise (some of the most successful technical analysts are unfamiliar with the basic rules of grammar!). The past half century has seen colossal changes in the financial markets. Collectively, the articles in this book provide an interesting historical record of developments over this period. Analysing the undercurrent of market sentiment will always be the key to profitable trading and investment and these articles also furnish the reader with a portmanteau of tools that can be used to navigate markets over the next 50 years.

To all past, present and future members of the Society of Technical Analysts. 50th Anniversary Party, June 2018 at the London’s Living Room, City Hall


8

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

CHAPTER ONE

Theory of Technical Analysis Articles in this chapter 10

A theoretical basis for Technical Analysis Tony Plummer FSTA

20

Towards a new understanding of Technical Analysis Tony Plummer FSTA

22

Order and Chaos in Financial Markets Tony Plummer FSTA

28

Art or science? Ruminations on the meaning of Technical Analysis Tim Parker MSTA


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

CHAPTER ONE INTRODUCTION

By David Watts BSC Eng MSTA, MICE

Introduction The Society has now been in existence for over 50 years with the journal being published for the last 34 years, such that over that time articles on every conceivable approach to Technical Analysis have been published. So before proceeding to look at the first chapter, what question should we first seek to answer? The perhaps “What is the basis for Technical Analysis? That is almost the title of the first article. We can then do no better than to start with the article “A theoretical basis of Technical Analysis by Tony Plummer” in which Tony considers a number of theoretical approaches. This will then begin to lay a foundation of understanding, in this statistics, Chaos Theory and then crowd or market behaviour all feature. Finally Tony then finishes with the mainstream price patterns and waves, which are part of the standard curriculum of Technical Analysis. The variety of behaviours inherent in a market are readily apparent to the student, in the way price actions react at key levels, these key levels being areas of support and resistance or accumulation and distribution. This demonstrates that the market observes a “market memory” that a Technician observes and then maps in his/her analysis. Then it’s worth noting market behaviours are quantified into categories like price ranges, breakout, pullbacks, short covering, Fibonacci retracements and even Elliott Waves to name more than a few, all which are apparent in a price chart. The way the crowd reacts is in fact the basis of the second article in this chapter “Towards a new understanding of Technical Analysis” also by Tony Plummer. Being able to read and understand the crowd behaviour apparent in price action as it develops, is a primary skill of a Technician. Then finally in this trilogy is the article “Order and chaos in financial markets” in which Tony develops a theoretical framework theory using Chaos Theory. I found this chapter aligns with my own understanding of price shocks, from a mathematical approach. In 1982, Benoit Mandelbrot published his classic book “The Fractal Geometry of Nature” which supports the findings of technical analysis, in his seminal study of cotton prices at various time frames. This resulted in many new indicators to classify such chaotic fractal behaviour. The final article for this chapter “Art or science, Ruminations on the meaning of Technical Analysis” by Tim Parker approaches the subject quite differently and from another perspective. What I particularly like about Tim’s article is his focus on “Context” or what I would describe as the wider market environment. So how a market behaves often depend upon this “Context” and many have thought they had a strict rule set leading to a pot of gold, only to find out that suddenly the market behaviour changed, just due to the market environment changing. Indeed, in the new age of computers in the

Psychology and Markets

Systematic Trading

1990’s simple computer models of the market were developed and then found to be wanting, just because there was no awareness of the dependencies that really needed to be included in a successful model. So, in conclusion, if you only take away from this chapter to think about the “Context” your market operates in, Tim would have done his job. It’s the professional technician that is able to consider the inter-market relationships and external influences and modifies his conclusions based upon that all important “Context”.

9


10

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

A theoretical basis for Technical Analysis Tony Plummer FSTA

Article originally featured in Market Technician 52 (October 2005)

Introduction I was recently asked to defend technical analysis to a UKSIP audience in a forum entitled “Technical Analysis: Astrology or Informed Analysis?” This was a challenge on a number of levels. First, it is not my view that the use of planetary alignments or planetary cycles to strengthen entry and exit rules implies ‘uninformed analysis’. On the contrary: there are many important books that suggest that life on earth is directly influenced by cosmic forces (Gaugelin, 1969; Lieber, 1979; Eysenck, 1982). In fact, since the giant planets (Jupiter, Saturn, Uranus and Neptune) usually pull the centre of mass in our solar system outside of the physical body of the sun (Landscheidt, 1989), it would be amazing if the influences were not actually quite significant. Furthermore, many successful traders - including W.D. Gann - have explicitly used astrology to finesse their timing. Second, there is the question of the degree to which astrology and technical analysis can actually be categorised together. Many technical analysts would argue that, even if astrology is a valid approach to markets, technical analysis is still a distinct and separate discipline because it uses different analytical tools. And this, in a way, points to the real challenge implicit in the forum’s title: is technical analysis just as a set of tools that sometimes work but sometimes don’t? Or does technical analysis actually have a genuine theoretical underpinning that justifies the use of certain analytical tools? What follows is part of my attempt to meet UKSIP’s challenge. However, there are two important qualifications. First, it only covers the phenomenon of price patterns, it does not cover the extraordinary influence of the Golden Ratio. Second, the analysis is offered only as a small part of an evolving process rather than as an unassailable body of theory.

Technical analysis and economic theory The most fundamental definition of technical analysis is that it is the study of past movements in asset prices in order to forecast future price movements in those prices. However, since past price movements are the outcome of various forms of investor activity, most analysts would also understand that the study of past price movements also includes - where possible - the study of various indicators relating to the supply of, and demand for, the assets in question. This allows us to include not only information - such as volumes, open interest, and put/call ratios - that can be derived directly from the various bourses, but also information - such as opinion surveys and cash holdings in funds - that can be gleaned from external sources. I generally use this broader definition, but want to make clear that excluding by definition does not mean excluding despite value. After all, if a specific indicator helps to make money, why exclude it? The basic idea underlying technical analysis is that human nature

has a consistency to it. This encourages certain general market phenomena, and certain specific price patterns, to reproduce themselves through time. Accordingly it is considered appropriate to formulate hypotheses about market behaviour on the basis of historical data and to use these hypotheses to anticipate the future. This process reproduces any other form of scientific enquiry. The problem, however, is that the assumptions about human behaviour used by technical analysts are profoundly different from the assumptions used by economic theorists. Technical analysts assume (usually implicitly) that the market behaves as a coherent whole. Economic analysts assume (very explicitly) that total market behaviour is no more than the arithmetic sum of random decisions by individuals. This difference in assumptions could not be better designed to foster mutual hostility between the disciplines. Technical analysts are accused of relying on some form of ‘hocus pocus’; economists are berated for ‘living in ivory towers’. The result is that technical analysts tend to ignore the influence of fundamental analysis on trends and that economists tend to ignore the power of technical analysis to forecast turning points. The truth is that both disciplines could usefully learn from one another. However, there is a need first to establish a common ground for understanding the phenomenon of financial markets.

The foundations of technical analysis In any market, prices fluctuate up and down. The implicit assumption of technical analysis is that these fluctuations are not random. The explicit claim of technical analysis, therefore, is that non-random fluctuations create patterns that repeat themselves. It is the repetition of patterns that allows us to forecast - or, better, to anticipate - price movements. So there are a number of questions that need to be answered: (1) (2) (3)

Are price movements random or non-random? If price movements are non-random, what patterns can we expect to emerge? If patterned price movements are present, how quickly can we decide which pattern is evolving?

Random or non-random behaviour So much has been written on the subject of the randomness or otherwise of price movements in financial markets that it seems dangerous even to attempt to address the issue. However, using a methodology suggested by the work of Baumol and Benhabib (1989), it is possible to look at the subject pragmatically, without going into a detailed discussion of statistical theory. Figure 1 plots each day’s percentage change (i.e., at time t) in the Dow against the previous day’s percentage change (i.e., at time


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

t-1), for every trading day since 2nd January 1990. In other words, movements in the Dow are plotted in what is called “t/t-1 phase space”. The result is exactly the sort of pattern that can be expected if daily movements were indeed random - that is, they are scattered widely throughout the phase space. Nevertheless, what is also relevant is that the movements are actually contained in a very limited area of that phase space. So, although the price movements appear random, they are in some sense contained. This becomes even more apparent if the scales on the chart are (say) quadrupled. See Figure 2. It is clear that price changes tend not to move out beyond a very specific twodimensional region. Should they do so, they tend to get pulled back in again towards the centre of that region.

Figure 1: Dow in t/t-1 phase space

8

6

4

DJ Industrial Average: Jan 1990 to Date (Daily close, t/t-1) Data source: Yahoo.com

2

t 0

-2

1 day percentage changes

-4

-6

-8 -8

-6

-4

-2

0

2

4

6

8

Figure 2: Dow in expanded phase space

DJ Industrial Average: Jan 1990 to Date (Daily close, t/t-1) Data source: Yahoo.com 1 day percentage changes

The operation of some sort of gravitational pull is particularly noticeable if the chart is extended backwards in time to (say) January 1946. By definition, the data set is a long one, not only covering an important period of economic and social evolution, but also including the 1987 equity crash. See Figure 3. Importantly, the basic region of attraction for the market remains unchanged, and the experience of 18th-20th October 1987 stands out as an idiosyncrasy. It is arguable that, in one sense, the 1987 crash was a truly random event. The point here is that the idea of randomness obviously has something to do with perspective: the longer the time perspective being taken (or, if you like, the broader the context), the less likely it is that fluctuations will seem random. Anyone who has traded in financial markets for any length of time will know when a price movement is ‘unusual’, based on his or her experience.

11


12

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Figure 3: Extended Dow data in t/t-1 phase space

32 28 24 20 16 12

19/20 Oct 87

8

DJ Industrial Average: Jan 1946 to Date (Daily close, t/t-1) Data source: Yahoo.com

4

t

0 -4 -8

1 day percentage changes

-12 -16 -20 -24 -28

18/19 Oct 87

-32 -32 -28 -24 -20 -16 -12 -8

-4

0

4

8

12

16 2 0 24 2 8 32

t-1

Strange attractors We can take this argument a stage further and look at the price changes over periods that are longer than one day. Figure 4 shows an example of the 5-day percentage changes in the Dow, in t/t-1 phase space. Here, the 5-day percentage change for a particular day is compared with the 5-day percentage change the previous day. What is clear is that the formerly circular ‘bubble’ containing price movements begins to spread out along an upward-sloping diagonal line. And this becomes even clearer for longer time periods. Figure 5 shows 20-day percentage changes in the Dow, in t/t-1 phase space. Oscillations in the Dow seem to be drawn towards the upward-sloping diagonal, where the rate of change at time t is equal to the rate of change at time t-1. Moreover, even though there are obviously unusual events, such as the 1987 equity crash, the oscillations are essentially bounded on the upside and downside. In the language of ‘Chaos Theory’, a strange attractor seems to be at work. It seems that, over longer time periods, changes in the Dow tend towards a stable path of change; but the changes are also persistently induced to accelerate and decelerate along this path. We can hypothesise that something encourages investors to chase the market up until some form of upper boundary is reached, and to chase the market down until some form of lower boundary is reached. This, in itself, is highly suggestive of a deep running ordered process at work. A related question is whether the acceleration/deceleration itself has a cyclical element of some sort. If it did, this would strengthen the case for non-random market behaviour. Figure 4: 5 day changes in the Dow

32 28 24 20 16 12 8

DJ Industrial Average: Jan 1946 to Date (Daily close, t/t-1) Data source: Yahoo.com

4

t

0 -4 -8

5 day percentage changes

-12 -16 -20 -24 -28 -32 -32 -28 -24 -20 -16 -12 - 8

-4

0

t-1

4

8

12

16

20

24

28

32


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

Figure 5: 20 day changes in the Dow

40 36 32 28 24 20 16 12

DJ Industrial Average: Jan 1946 to Date (Daily close, t/t-1) Data source: Yahoo.com

8 4

t

0 -4 -8 -12

20 day percentage changes

-16 -20 -24 -28 -32 -36

18/19 Oct 87

-40 -4 -3 -3 -2 -2 -2 - 1 -1 -8 -4 0 6 2 8 4 0 6 2

0

4

8 12 16 20 24 28 32 36 40

t-1

Groups and crowds Before we look more closely at the possibility of cyclical influences, we need to consider the basis for an ordered process in human behaviour. According to economic theorists, individuals make their decisions independently of one another. Group behaviour is then considered to be forecast-able because, according to probability theory, a large number of uncertainties somewhat spookily create a certainty. As James Surowiecki has recently demonstrated in The Wisdom of Crowds (Surowiecki, 2005), this is not only absolutely correct, but can yield answers and decisions that are amazingly accurate. The problem, however, is that once individuals start to be influenced by each other’s behaviour, then their expression of individuality is much reduced. Ultimately, deviation from some measure of average behaviour becomes minimal. This presents a problem for theoreticians because, although basic probability theory breaks down, it is highly likely that a large number of people doing the same thing will produce a forecast-able outcome. This, in a way, was one of the central findings of the late 19th century French sociologist, Gustave le Bon, whose now-famous book, The Crowd, analysed the French Revolution (le Bon, 1895). Le Bon observed that, when people came together in a common cause, the result was something different to just the sum of the parts. People behaved differently to the way that they would as individuals: they would focus on, and follow, the dictates of a recognised leader; they would act to protect their beliefs; and they would quickly see ‘non-believers’ as enemies. In short, the act of people coming together under a unifying belief system would foster conflict. The truth behind le Bon’s assertions is only too clear in the bloodied history of the twentieth century. There is, however, an important aspect of his analysis that is easily overlooked. This is that le Bon saw crowds, or groups, as psychological phenomena whereby people behave in a unified way once they adopt a shared set of beliefs. This is because belief systems - however unlikely and unreasonable they may seem - mobilise powerful inner emotions. Hence, crowds are held together, and energised by, emotions rather than by cold logic. Moreover - and this is important - the self-awareness of participating individuals is reduced (Neumann, 1990). The psyche is “invaded”by the values of the collective.

Consequently, the ability of individuals to make rational choices, evoke moral judgement and engage in active reality testing is suppressed. This, of course, would help to explain why outsiders have such difficulty in understanding and interpreting group behaviour. Importantly, though, it might also help to explain why stock markets bubble and crash, and why economies boom and slump. Somehow, emotionally laden beliefs are increasingly distributed throughout a market place and throughout an economy.

Co-operation The suggestion here is that warfare and violence may actually be a specific outcome of the more general tendency towards co-operation amongst human beings. Indeed, this suggestion becomes seriously compelling once account is taken of contemporary ideas relating systems theory, group psychology, and natural evolution. Nature organises itself hierarchically into ever-greater wholes: lower-order parts contribute to higher-order wholes, and the wholes organise the parts. As a result, each whole is qualitatively different to the mere summation of the parts. In human beings, this organisational force finds expression in the need to merge psychologically into a group. At one level this can be explained as the need to reduce the sense of personal isolation. At another level it can be explained as the need to have a sense of purpose. Either way, the outcome is ‘natural’ in the genuine sense of that word. However, the inner need to merge into greater wholes takes on a different imperative when competition or conflict is involved. The associated threat to each individual’s psychological security is reduced when others are involved in meeting that threat (Trotter, 1947). Quite obviously, the greater the threat, the more urgent it is that each individual’s resources are directed towards meeting the purposes of the whole. A competitive environment therefore demands conformity from individuals (Bloom, 2000). Non-conformity is punished by exclusion from the group: individuals are left to meet the threat alone. The point is that, once account is taken of the inner dimension of human existence, the need to do things together becomes a viable explanation for a part of human behaviour. Evolution can then be seen in terms not just of random mutation and survival of the fittest individuals but also in terms of the perpetuation of the

13


14

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

most adaptive groups in the face of external threats. Nature is, ultimately, a collaborative enterprise.

Conformity enforcement Importantly, the idea that group beliefs can induce conformity has found a great deal of support from independent researchers. In the 1950s, Stanley Milgram was able to show that more than 60 per cent of ordinary people could be induced to deliver massive electric shocks to apparent patients just by being told by an authority figure in a white coat that it was alright to do so. Fortunately, the ‘patients’ were actors and no electric shock was actually involved. In the late 1970s, Ed Diener at the University of Illinois published evidence that, in a group setting, people strongly identified with other group members, had little sense of personal identity and tended to act without prior thought. Other evidence - notably from Bristol University in the UK and from Harvard in the US - showed that people’s perception of non-group members and, indeed, of reality itself, could all too easily be influenced by pressure from other group members. In the psychological arena, therefore, a group can be less than the sum of its parts. In recent years, the central ideas of self-organisation, group conformity enforcement, and non-rational collective behaviour have been used by a number of important developmental theoreticians to explore the processes of history. Foremost amongst these have been Arthur Koestler (1978), Erich Jantsh (1980), Fritjof Capra (1982), Ken Wilber (1983), and Howard Bloom (1997 and 2000).

Non-rational behaviour What theory and research are both pointing to is the very real possibility that economic and financial activities are, at heart, less rational than many might want to believe. This does not mean that such activity is always irrational just that it is essentially activated by deeply held psychological needs. These needs are orientated towards obtaining security and meaning, and are universal. Hence, non-random behaviour in economic and financial markets may be the outcome of genuine group influences rather than just the outcome of statistical interactions. This, in turn, would mean that excesses (which are, in any case more easily observed after the event than before it) could well be part of a forecast-able spectrum of behaviour instead of the unforecastable outcome of ‘keeping up with the Joneses’ or of ‘animal spirits’. The problem, however, is that a clearly defined body of theory that covers this spectrum of economic and financial behaviour is not currently available within the academic community. This suggests that some form of paradigm shift is almost certainly looming as a result of the aftermath of the financial bubble of the late 1990s. After all, how long can academics continue to ignore the tendency of markets to diverge from, and oscillate around, fundamental values? However, it also implies that we need to look for guidance outside of the current ‘rational expectations’ paradigm that embraces economics.

Financial markets One way forward is to look for patterns that regularly emerge in financial markets. There are three reasons for this. First, largerscale movements in financial markets basically reflect the evolving mood that embraces all economic and social behaviour within a community. Second, financial markets provide a continuous flow of uncontaminated data. Market price action therefore provides a marvellous testing ground for hypotheses about human behaviour. Third, recurring price patterns would (if found) imply recurring behaviour. In other words, the patterns could be interpreted. This, indeed, is the route that technical analysts have chosen to take. The result is a large body of industry literature confirming

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

that (a) markets oscillate in reasonably regular cycles, that (b) markets spend time in base, top or holding patterns before entering a significant trend, and that (c) these price patterns tend to incorporate certain predictive price configurations, such as ‘head and shoulder tops’. The overriding impression is that market behaviour is not random.

Bubbles and crashes A valid point of access for an analysis of market price patterns is to look at investor behaviour during periods of emotional extremes. Traditional economic theory regards such extremes as aberrations. However, insofar as such extremes are actually part of a spectrum of behaviour, they are likely to reveal the basic energies that drive all behaviour. Extremes of behaviour are often noteworthy for their clarity of purpose. Shown in the Figure 6 are the loci of US price action in the Dow Jones Industrial Average from January 1921 to July 1934 and in the NASDAQ from September 1995 to September 2002. The time periods involved are different: the former covers a period of seven years, the latter covers a period of three years. However, when the time elapse relating to the Dow is placed on the lower horizontal axis and the time elapse relating to the NASDAQ is placed on the upper horizontal axis, something very interesting emerges: the patterns of the acceleration into the peak and the subsequent collapse are very similar. This similarity has been recognized by Didier Sornette, who is Professor of Geophysics at UCLA. Sornette has shown how non-linear mathematics can track and predict a stock market ‘bubble and crash’ (Sornette, 2003). There are two distinct conclusions. The first is that the price acceleration into the final peak is curvilinear, and that the time-elapse of oscillations around that accelerating trend gets progressively shorter. The second is that this specific phenomenon only works because of the impact of “cooperative self-organization”. In other words, non-linear mathematics can predict the timing of the peak (because the oscillations become so fast that they effectively converge on zero), but such non-linear mathematics only work because stock markets are ‘natural’ systems.

Homogeneity These conclusions are a dramatic confirmation of the impact of group behaviour in financial markets; and it is almost no accident that they have been generated outside of the discipline of economic theory. Nevertheless, they need to be placed in a wider context of investment behaviour. Professor Sornette notes that, as a stock market bubble accelerates into a peak, investors take more and more notice of what others are doing. Hence, behaviour becomes increasingly homogeneous and local information has long-distance effects. Ultimately, of course, the market becomes satiated, or ‘overbought’, and extremely vulnerable to small perturbations. So only a small amount of profit taking can initiate a full-blown ‘crash’. What Professor Sornette is, in fact, describing is a specific - and particularly dramatic - example of a more general mechanism. This is the mechanism that induces conformity from market participants and thereby produces oscillations in financial markets.

Conformity enforcement in financial markets Financial markets are characterized by inter-group conflict: it is a contest between the bulls and the bears. Seen in this light, financial markets are not just processes that encourage prices to converge on fundamental induced values. They are reflections of a collective movement between the opposite polarities of optimism and pessimism. Hence, prices are likely to overshoot fundamental values in both directions. One implication is that market participants actually pay less attention to ‘fundamental’ values


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

Figure 6: The Wall Street Crash and the NASDAQ Collapse

The Wall Street Crash & the NASDAQ Collapse (Weekly closes) Data sources: Economagic & Datastream

than is usually thought. Partly, this is due to the fact that such values are very difficult to calculate in advance. Partly, though, it is due to the competitive process. So much attention is ultimately given to ensuring that one particular view is right (e.g., prices are going up) that participants lose sight of fundamentals. Energy is spent in generating propaganda to colleagues, clients, other members of the profession and the media. The sub-conscious intention is always to ensure that financial resources continue to underwrite one’s own view. What is missed, however, is that this is also exactly what others are doing. Conformity enforcement is a very subtle process - which is why it is usually deemed not to exist. Nevertheless, it is conformity enforcement that lies at the root of all price trends. In modern financial markets, the pressure to conform has become institutionalised. Market professionals are consistently monitored against peer group performances: indices are constructed to show the sector allocations of other funds, and deviations from that ‘norm’ are monitored for success or failure. For the individual fund manager it is a risk to take a marginally different view, let alone an alternative view. So, when a market either begins to run ahead of perceived valuations, or even begins to ‘bubble’, there is huge pressure to join in. Not to do so is a direct risk to personal wealth and personal income. Ultimately, however, it is this threat to personal status that provides the main pressure to conform. Individuals participate in markets for reasons of wealth, power and prestige - in other words, to enhance control over future resources. To be out of the market when it is going up, or in the market when it is going down, threatens this control and generates fear. Collective behaviour, of course, reduces fear. So, for the majority, it is easier to trade on the evidence of actual price movements, than it is to invest on the basis of theoretical valuations (which may, in any case, be wrong). For whatever reason, the end result is that the psychological environment of the market becomes dominated by a limited set of uncritically held beliefs, known as memes. In a bullish market the meme is that prices are going up; in a bearish market, the meme is that prices are going down. A meme is the glue that holds otherwise disparate individuals together in groups and crowds.

The basic mechanism Figure 7 shows some of the basic principles involved. It

demonstrates the most likely pattern that will be traced out by a broad financial market price index during the course of a complete bull-bear cycle. No account is taken at this stage of the timescale involved. Starting at the lower left-hand side of the chart, a market will be very oversold, probably after some form of crisis. There will then be a bear squeeze of some sort as short positions are covered. This need not entail a significant proportion of investors suddenly becoming bullish - it just needs some investors to close bear positions. This will cause the market to jump sharply. The rise may continue for a little while because investors do not all respond simultaneously. Crucially, they will respond, not so much to ‘fundamental’ considerations, as much as to the fact that prices are rising. In other words, price movements - particularly sharp movements - are a critical item of information. At some early stage, however, such technical buying dries up, and the market begins to retrace back towards the lows. This is a ‘re-test’ of the lows and occurs while those who missed the initial rally will be considering whether or not they now need to react. Investors will take into account the fact that prices have rallied but they will also necessarily re-assess fundamentals. Some may even decide that the market had originally over-discounted fundamentals or that the fundamentals might be shifting. This triggers another bout of buying, which eventually takes the market out of whatever holding pattern it has been in. In other words, a trend starts to materialize. As time progresses, either new information becomes available that confirms that fundamentals are improving and/or the rally in the market enters a feedback relationship with fundamentals such that the latter improve anyway. Figure 7: The basic mechanism

15


16

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

There are a number of very important points that emerge from this simple model of market behaviour. First, a reversal materialises once a market has gone too far. The market is, in some sense, satiated. Second, the reversal is an information shock to the market. It reveals that the market may be out of line with fundamentals and that the market can no longer attract sufficient investor energy to continue the old trend. There is thus an ‘energy gap’. Third, the subsequent re-test of the low occurs while investors absorb the implications of what has just happened. Investors ‘learn’ that something has changed. Fourth, the market ‘signals’ a trend move by breaking out of the holding pattern. Since investors have, in effect, learnt that fundamentals have changed, all subsequent information will be seen in the context of that learning. Data that confirm the trend will increasingly be acted upon; data that contradict the trend will increasingly be ignored. Fifth, market participants will increasingly focus their attention on price action rather than fundamentals. Finally, the market will run ahead of fundamentals and will become overbought, or satiated. This presents the conditions that will trigger a reversal. The whole process then begins in reverse.

The role of prices We thus have a three-phase mechanism that accounts for all the basic behaviour within a financial market trend, whether up or down. We also have a specific mechanism that accounts for market reversals. Consequently,we can hypothesize the existence of a six-wave pattern in a full market cycle - three waves up and three waves down. This is an important conclusion. However, there are other important inferences that need to be drawn. The first, which has already been mentioned, is the important role of prices. Individuals can process complex information, but a group can only react to simple information. The most important piece of information to a financial market group is the actual behaviour of prices. The response of the group will be greater, the faster and more pronounced is the change in prices. In a sense, therefore, prices will fulfil the leadership role in a psychological group. The group will accordingly react to this leadership and it will chase trends. The process is dramatically enhanced when high profile individuals confirm their own personal commitment to the trend. This, of course, stands economic theory on its head. In economics a price is determined by the behaviour of buyers and sellers. This is true, but is only part of the process. When prices generate information, prices also determine behaviour. So a feedback effect is involved. As the biologist and philosopher Gregory Bateson has argued, feedback is one of the characteristics of any living system (Bateson, 1979).

The process of learning This brings us to the second inference from the market model. In any system that is oscillating in a feedback relationship with its environment, any new information from that environment has to be assimilated and absorbed. A process of learning is therefore involved. So it hardly seems accidental that the mechanism just described mirrors the process of learning that can be found in the human brain. In the early 1960s, Henry Mills found that, although people would initially be very quick at picking up the mechanics of a new task, they would inevitably go through a stage where their ability to apply their new learnings would slow (Mills, 1967). Only after this slowdown could activity speed up again. Normally this ‘slowdown’ is missed because it is only temporary. Nevertheless, it is a real phenomenon, which reflects something important. At some stage during the learning process, information is transferred from short-term memory in the forebrain to longterm memory deeper within the brain. Learning thereby moves from the conscious into the sub-conscious (Hebb, 1949). This is both automatic and necessary, and frees up consciousness for

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

other tasks. However, energy has to be diverted away from other processes in order to facilitate this adjustment, and the ability of a person to do a conscious task actually deteriorates temporarily. After the adjustment, people can apply their learnt techniques to the new task and not even think about it very much. Markets are exactly the same. At a top or bottom, markets will go through a process of learning that fundamentals have changed. They respond to new information, hesitate while that information is absorbed, and then automatically apply the resulting learning during the thrust of a trend.

Markets as collective learning processes Financial markets (and, with them, whole economies) can be viewed as natural self-organizing systems that learn from their interaction with their environment. They are a particular form of what Howard Bloom of New York University calls “collective learning machines” (Bloom, 2000). As such, they organize their lower-order parts in a coherent fashion, oscillate rhythmically, and express themselves in terms of a limited matrix of patterns. This is essentially why the discipline of technical analysis has the power regularly to generate effective buy and sell signals. Analysts will look at indicators of investor energy in order to estimate the intensity of the market’s hold over investors. They know that the stronger the market’s grip, the nearer the market is to a turning point. Analysts will also look at the periodicity of historical oscillations in order to forecast the timing of likely turning points in the future. If a cycle has made itself felt in the past, it is likely to continue into the future. And, finally, analysts will look at the current evolution of price patterns, in the knowledge that certain patterns reproduce themselves. Once the market’s position in the context of a specific pattern is known, it is possible to estimate what might happen in the future. Quite obviously, the most powerful signals are going to be generated when each of the three lines of analysis coincide.

Price cycles Central to technical analysis procedures is the phenomenon of price cycles. It was earlier noted that financial market oscillations might be driven by a natural learning mechanism and that inflexion points might be triggered by an energy gap that arose out of investor satiation. The important point here is that an energy gap reverses the polarity of the market from bullish to bearish, or from bearish to bullish. There are, however, two important questions: First, does this mean that bear markets are inevitable? Second, what determines the difference between a big bear market, such as those that emerge in the form of a ‘crash’, and minor setbacks? In the context of financial market cycles (and, indeed, of economic cycles), it is necessary to understand that downswings are as important as upswings. Nature cannot evolve without periods of rest, because it needs to replenish its energy. Hence, any period of activity will be followed by a period of rest. So, despite the best attempts of governments, bull markets will always be followed by bear markets, and economic expansions will always be followed by recessions. What then determines the extent of a downswing? One answer, of course, is the amplitude of the upswing: the bigger the upswing, the bigger (potentially) the correction. However, it also depends on the time span of the cycle: the longer the cycle, the longer (potentially) the correction. Technical analysis has the capability of determining the difference between big moves and small ones by putting all moves into the context of history and accepting that this history has a valid and vital role to play. Hence, for example, if there is strong evidence that the Dow Jones Industrial Average has, for a very long time, oscillated with a rhythmic periodicity of about 11 years (which it has), then there is every reason to suppose that the oscillation will continue. This will give a strong indication of when an important reversal can be


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

expected and what order of magnitude that reversal might take. The primary presumption of cycle analysis, therefore, is that this time it will definitely not be different.

Price patterns within cycles An important inference is that cycles can be defined by their patterning as well by the precision of their periodicity (Plummer, 2003). This means that the observed variability in cycle periodicities does not invalidate the forecasting potential of cycle analysis because the evolution of a current cycle can be tracked in real time against the pattern of a previous cycle. An example of this is shown in Figure 8, which compares the pattern of fluctuations in the Dow Jones Industrial Average between September 1990 and September 2001 with the pattern of fluctuations between September 1957 and May 1970. Both periods represent one beat of the 11-year cycle in the Dow, and both periods embraced rapid change in the US economy: 1990-01 covered the revolution in information technology, and 1957-70 covered the social revolution of the ‘Swinging Sixties’. In a sense, therefore, the two periods are directly comparable. When the price patterns of the two periods are overlaid on one another, by the simple expedient of plotting each beat of the cycle on a separate time axis, a remarkable similarity emerges. This is not unusual. Once the coincidence of patterns is found, it becomes a simple matter of tracking a new cycle beat against an earlier comparable one and tracking that cycle beat into its final low. Any variations in the periodicity will not matter.

Figure 8: Patterns in the Dow

US Dow Jones Industrial Average (Monthly closes, 1 year % change) Data Source: Bloomberg

This, in a sense, is where technical analysis brings such great strength to market analysis. It focuses directly on the patterns of market behaviour - both in terms of price movement and investor activity - because its working assumption is that such patterns are both non random and meaningful.

The fallacy of the rational individual The small and simple shift in emphasis - from the individual to the group - creates a massive shift in our understanding of economic and financial motivation. Economic theory cannot properly explain why a large number of people, who are assumed to be making rational decisions independently of one another, end up (for example) buying red cars or trying to move house, all at the same time.“Mood” may be regarded as being part of the answer; but, then, by what mechanism can a change in mood be made to swarm through a population of separate and rational individuals? Economic theory also cannot properly explain why particular patterns emerge in financial markets. If individuals really do make decisions independently of one another, then prices should just jump about in a random fashion. There is no mechanism for explaining why markets should generate the specific behavioural patterns that have so far been analysed. Nor is there any mechanism for explaining why specific patterns recur.

Bearing this in mind, it is now appropriate to look at some of the findings of technical analysts regarding specific price patterns.

Derivative price patterns If a single beat of a cycle (of whatever length) contains a simple sixwave (three-up / three-down) pattern, then it should be possible to isolate that pattern from whatever trend is driving the market. Or, to put the same thing another way, since a cycle beat consists of six basic waves, these waves can only present themselves in a small number of ways when they are subjected to a trend. Moreover, each trend will itself be part of the six-wave structure of a higherlevel cycle beat. If these conclusions are correct, then not only are price patterns non-random and meaningful, but also they are limited in number. This is a significant claim. It means that technical analysis - when properly approached and applied - is an extremely powerful method of interacting with financial markets. This, indeed, is what some of the leading proponents of technical analysis in the last one hundred years have argued. One of these analysts, whose research spanned the traumatic years of the first half of the twentieth century, was Ralph Nelson Elliott (Elliott, 1938). Elliott’s work is usually ignored by economists for the very reasons it is so powerful - namely, it assumes non-random, patterned, group behaviour. However, Elliott made two significant observations.

17


18

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

First, all impulsive movements, whether up or down, consist of fivewaves (three in the direction of the trend, interspersed with two corrections). Second, all corrections consist of three waves (two in the direction of the corrective trend and one contra-trend move).

Five-three patterns and investor behaviour At first sight, Elliott’s five-three pattern appears to contradict the hypothesis of a three-three archetype. However, the two are entirely consistent, once (a) the role of trends is included and once (b) some allowance is made for the possibility that Elliott himself may not have had all the answers. Shown in Figure 12 is the formation generated when a rising trend is applied to an otherwise balanced three-three cycle. The cycle itself is notated as 1-2-3 up and A-B-C down. Once a trend is applied, however, wave B (which is theoretically a ‘re-test’ wave) actually makes a new high. In other words, the basic three-three profile incorporates a five-wave movement. Figure 12: The effects of a trend

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Three-wave corrections This particular model does not, of course, automatically generate a three wave correction. There was, however, something missing from Elliott’s original exposition. Elliott was intrigued by the patterning of markets and did not look too closely into their causes. However, the foregoing exposition places great store on the impact on investors of sharp price movements. Such movements are information ‘shocks’. The impact of energy gaps has already been discussed. These are contra-trend shocks. There is, however, another class of information shocks. These are protrend shocks, caused by a higher-level trend, where prices either move into, or extend the length of, an impulse wave. The presence of a pro-trend shock is recognizable by sharp price movements, increases in trading volumes, price gaps between one day’s close and the next day’s open, and rising open interest. All shocks have to be absorbed by the market and therefore generate a price retracement of some sort. The contra-trend shock produces a re-test of the high or low, the pro-trend shock generates some form of subsequent holding pattern. Figure 13 shows the influence of a pro-trend shock. A break out from the base pattern indicates to investors that a trend is now developing. This is an item of information to which they respond, and buying volumes increase. Eventually, however, satiation sets in and the market hesitates (and, in effect, waits for ‘fundamentals’ to catch up). The market then moves ahead again, reaches its climax and turns down into a relatively deep correction.

Figure 13: Pro-trend information shock

This is a very important insight and justifies a lot of the work that technical analysts conduct in relation to market satiation. There are two critical phases after a trend has developed. The first is when it is ‘overbought’ or ‘oversold’. This is the potential point of inflexion in a cycle, and may be captured by indices such as overstretched momentum or by indicators representing panic buying or panic selling. The second critical phase, however, is when a ‘fifth wave’ extends the market into a new high or new low. There are two possibilities. First, the fifth wave is not, by its very nature, truly impulsive. In this case, it may not be ‘confirmed’ by indicators of investor enthusiasm for the trend. Momentum may be weaker, volumes may be lower, and open interest in the relevant futures markets may be falling. Second, the fifth wave may be dynamic enough to create an investor panic. In this case, investment positions are finally driven to satiation. It is quite obvious that new highs or lows that are either not confirmed or generate excesses (or both) could be followed by a sharp reversal. Note that some capitulation is likely to occur at the end of the third wave. It can also occur at the end of the fifth wave. On rare occasions, it may happen at the end of the first wave. Capitulation is therefore likely to occur between one and three times in a specific trend.

The combination of simple distortions to lower- level cycles and the more dramatic effect of information shocks produced by highlevel cycles yield the basic five-three pattern observed by Elliott. Underlying this pattern, however, is the operation of a three-three cycle.

Other patterns This analysis is very brief and does not do justice to the forces involved. Nevertheless, it already confirms two other aspects of technical analysis. First, it confirms the influence of the famous ‘head and shoulder’ pattern that so often defines a reversal either out of a market high or away from a market low. Figure 14 shows the head and shoulders top formation that was implicit in Figure 13. The top of wave 3 becomes the left shoulder, the peak of wave 5 is the head, and the end of wave B is the right shoulder. A line linking


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

the lows of waves 4 and A become the neckline and a sell signal is generated when prices fall through that neckline. The reason why the pattern works so effectively is that in moving into a trend, the market has probably responded to a higher level information shock. This means that the qualitative structure of the market has shifted - it has incorporated new information and has evolved. Therefore, when a correction occurs at the end of the trend generated by the information, that correction necessarily comes from a ‘higher’ level. It necessarily is longer in price and time than those that preceded it. This, indeed, is what Elliott found.

Figure 14: The head and shoulders formation

Psychology and Markets

Systematic Trading

The role of technical analysis The discipline of technical analysis has been developed only gradually, over a very long period of time. Its main driving force has been one of profitability rather than theoretical nicety, and this has often militated against effective communication with practitioners in other areas of research, such as economics. Despite its apparent lack of theoretical rigour, technical analysts have observed, catalogued and used a huge volume of very effective tools and predictive techniques. There are, for example, no known price patterns that lie outside of Elliott’s classifications. Once these tools and techniques are seen in the context of selforganizing groups, which learn (both from their environment and from their own behaviour) then much of it begins to make sense. Markets evolve in a non-random fashion, according to patterned cycles. Different styles of investor behaviour can be identified at each stage of these patterns, and can therefore provide a strong clue as to where a market is in any particular cycle. More to the point, technical analysis has the power - sometimes predictively, but always quickly - to signal serious price reversals. We have shifted from forecasting based on the arcane workings of statistics to forecasting based on the extraordinary forces of Nature.

Second, the analysis substantiates the presence of genuine trends in markets. This, alone, is a hugely important conclusion. Markets enter a trend when investors have learnt that circumstances have changed: they are only applying ‘learnt’ behaviour. Trends basically continue until markets have run too far ahead of fundamentals.

Bibliography Bateson, G (1979) Mind and Nature: An Essential Unity. Wildwood House, London. Baumol, W and Benhabib, J (1989) “Chaos: Significance, Mechanism, and Economic Applications” in Journal of Economic Perspectives. Bloom, H (1997) The Lucifer Principle. Grove Press, New York. Bloom, H (2000) Global Brain: The Evolution of the Mass Mind From the Big Bang to the 21st Century. John Wiley, New York Capra, F (1982) The Turning Point. Wildwood House, London. Elliott, R N (1938) The Wave Principle. Elliott, New York. Reprinted in Prechter R. (ed.) (1980) The Major Works of R N Elliott. New Classics Library, New York. Eysenck, H and Nias, D (1982) Astrology: Science or Superstition? Temple Smith, London. Gaugelin, M (1969) The Cosmic Clocks. Paladin, London Hebb, D (1949) The Organization of Behaviour. John Wiley, New York. Jantsh, E (1980) The Self-Organizing Universe. Pergamon, Oxford. Koestler, A (1978) Janus: A Summing Up. Hutchinson, London. Le Bon, G (1895) Psychologie des Foules. Felix Alcan, Paris. Reprinted (1922) as The Crowd. Macmillan, New York. Lieber, A (1979) The Lunar Effect. Corgi, London. Mills, H R (1967) Teaching and Training. Macmillan, London. Neumann, E (1990) Depth Psychology and a New Ethic. Shambhala, New York. Plummer, T (2003) Forecasting Financial Markets. Kogan Page, London. Sornette, D (2003) Why Stock Markets Crash. Princeton University Press, Princeton. Trotter, W (1947) The Instincts of the Herd in Peace and War. Ernest Benn, London Wilber, K (1983) Up From Eden. Routledge & Kegan Paul. London.

19


20

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Towards a new understanding of Technical Analysis Tony Plummer FSTA

Article originally featured in Market Technician 4 (February 1989)

One of the most important ideas to have been developed in the natural sciences in recent years, is that order develops out of chaos. This idea is partly based on the simple - yet profound notion that every part of the natural world contains both an element of randomness and an element of predictability. One of the easiest ways of describing the phenomenon is in terms of probability theory: it is not possible to forecast individual events with a great degree of confidence, but it is possible to forecast group events with only a small margin of error.

Movements in prices determine who is right and who is wrong. Emotions are closely involved with the result. The winners will feel pleasure and the losers will feel displeasure, especially fear. Furthermore, these emotions will be intensified by communicating with like-minded individuals. Ultimately, of course, being part of the “wrong” crowd contains no benefits whatsoever. Fear therefore triggers evasive action. Who has not experienced sweaty palms during their trading activities and done precisely the wrong thing as a result?

The concept of a “duality” of characteristics has a great deal of validity for Technical Analysis because it can be used to explain the presence of an underlying order in financial markets. Such order has been known to exist ever since stock prices were first plotted on graph paper. However, the reasons for it has essentially eluded all except analysts such as William D. Gann.

At the beginning of a new trend, most players are still locked into the emotions and thought processes of the old trend. However, there will always be a large minority who have been successful in recognising the turn. These people will be in a position to look at the fundamentals rationally and will therefore be the stimulus for the new trend. As this new trend develops, more and more people will join it on the basis of changing fundamentals. However, towards the end of that trend, the vast majority will tend to look only at the most recent trend in prices and forget the fundamentals. The process involves a subtle shift from rational to non-rational behaviour. More precisely, since behaviour is always a mix of rational and non-rational, it shifts from a mix of high rational/ low non-rational to a mix of low rational/high non-rational.

The source of the current myopia is to be found in the methodology and assumptions of economic theory. The analytical procedures of theorists such as Rene Descartes and Adam Smith have concentrated on the “parts” of a system rather than on the system itself. It has also encouraged an unyielding belief in’ the dominance of rational behaviour in human affairs. Economies are therefore seen as nothing more than the simple summation of all activities, and events such as panics and crashes are regarded as exceptions, rather than as a special case, of a general rule. The equity crash of 1987 has, however, presented the theorists with a problem. First, it was found that “Random Walk” did not operate during the Crash. Second, the theorists could not make the Crash fit into the rigorous theoretical framework of modem economics. The “solution” has (so far) been to assume that everybody responds rationally to irrelevant information (which creates a “bubble”) only to adjust their expectations very quickly when the information is found to be wrong (which creates a crash). Nobody seems to have asked why it is that large numbers of rational people do not recognise that the information is incorrect in the first place? The truth of the matter is that the basic assumption of rational behaviour used by modem economic theory is only applicable as a special case. If every part of nature has both an identity of its own and identity derived from its participation in greater wholes, then it follows that every person can be seen both as a unique individual and as a member of large socio-economic groups. Gustave Le Bon was one of the first analysts to recognise that people in a group, or “crowd”, behave non-rationally (and sometimes irrationally). He argued that crowds are psychological phenomena, where common beliefs encourage people to behave alike. It is not necessary for people to be standing shoulder to shoulder. Nations, for example, are “crowds” because the members share the same belief system. Crowds develop in financial markets because participants end up having common beliefs about the future trend in prices.

Most people would intuitively accept this argument. However, the important implication is that the processes involved are not random. Crowds can be analysed just like any other “natural” system. They therefore respond to new information in an elementary way, and they oscillate rhythmically. It is these two simple points which fully justify the assertions of generations of technical analysts that not only do specific price patterns recur, but that it is possible to forecast both the extent and the duration of a potential price movement respond to new information in ,an elementary way, and they oscillate rhythmically. It is these two simple points which fully justify the assertions of, generations, of technical analysts that not only do specific price patterns recur, but that it is possible to forecast both the extent and the duration of a potential price movement. Part of the logic involved can be demonstrated with the use of a very simple diagram. It is a limit cycle. Limit cycles are now regarded as being an accurate description of most of Nature’s dynamic processes. In the case of financial markets, one of the relevant limit cycles relates price changes to the level of investor sentiment. Price changes and sentiment move together during the main part of a trend, with no clear cut direction of causation. Volume and open interest rise during this stage. At turning points, however, the relationship breaks down. The limit cycle is biased to the right. Changing prices cannot stimulate further investor involvement because the market is overbought or oversold. This is the first reason why analysts have been able to use the concept of “non-confirmation”. However, behaviour does not stay on the limit cycle all the time.


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

As profit-taking develops, prices change direction and a shock is delivered to the limit cycle system. There is a “once-off” change in prices and sentiment inwards from the limit cycle - this is what happened during the 1987 equity “crash”. Next, because of the self-organising processes involved, the path of adjustment involves a spiral back to the limit cycle. This spiral involves a re-test of the earlier turning point. This is what is currently happening in the equity markets.

Psychology and Markets

Systematic Trading

length in my new book “Forecasting Financial Markets: The Truth Behind Technical Analysis” which is to be published by Kogan Page this Spring.

The re-test may involve a new high in a bull market, or a new low in a bear, but it is not usually accompanied by a noticeable improvement in the sentiment and breadth indicators. This, of course, is the second type of, “non-confirmation”. Such a re-test is invariably followed by a move into a new trend. This is specifically why Elliott discovered that all bear markets are three-waves. The implication of the model, however, is that bull markets are also three-waves. This appears to contradict Elliott’s assertion that all bull markets are five-waves affairs. In fact, five-waves develop when the limit cycle of next higher degree forces a re-test to extend into new territory. Five up-waves are therefore a sign of basic strength. Indeed, it is entirely logical that, during the growth phase of the capitalist system, equity bull markets will always tend to be five-wave affairs. Such a conclusion will not, however, hold true when the economy begins to transfer its belief system into another structure. There are many other important implications of this simple model. It can be used to explain all the price patterns which are recognised by Technical Analysis. It enables an analyst to measure prospective price movements with a degree of accuracy which is simply astounding. Furthermore, it confirms the presence of rhythmic oscillations thereby enabling the analyst to forecast the duration of price movements. These features are explored at

Tony Plummer was formerly a Director of Hambros Bank Ltd. and responsible for their trading positions in the London Gilt-Edged market.

21


22

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Order and Chaos in Financial Markets Tony Plummer

Article originally featured in Market Technician 9 (October 1990)

Everybody has to make assumptions about the nature of the world in which he or she lives.; it is impossible to conduct personally all the research which would be necessary to arrive at a perfect understanding of reality, even if such an understanding was possible. As a result, each person has to learn about reality from other people - particularly from parents, teachers and friends. Furthermore, since the only practical check on the validity of learnt beliefs is whether or not other people subscribe to the same belief, there is a natural tendency to accept as being true only that which a number of other people accept as being true. In practice, therefore, our understanding of reality progresses by a type of common consensus. Furthermore, those individuals whose task it is to test for the nature of reality will formulate their experiments within the context of the common consensus.1 This limitation on the flow of genuinely new ideas means that existing theories are not easily replaced: they tend to evolve without any radical change to the underlying structure. 2 A revolutionary “paradigm” shift usually only materialises either, when a persistent and increasing number of observations no longer accord with consensus beliefs, 3 or when there is a major social crisis. 4 At any point in time, therefore, there could be two or more theories which are consistent with the available evidence, but only one will actually be “acceptable”. Not surprisingly, very little effort will be made to disprove the consensus theory, and significant efforts will be made to repudiate alternatives. This would not necessarily matter if all that was at stake was the accuracy of our understanding of reality. However, understanding has implications for derived behaviour. Behaviour based on an incorrect interpretation of reality may eventually lead to a mismatch between expectations and outcomes. It is the stress inherent in this mismatch which creates psychological disturbances in individuals and disorder in economic systems. The problem, therefore, is not just one which is an interesting subject of debate amongst academics: it is of profound importance to the way that we conduct our lives and to the way that Governments involve themselves in our lives. The argument of this essay is that our inability to solve the apparently intractable problems of the current era - unemployment, inflation, global pollution, and environmental destruction - is largely the consequence of a significant divergence between economic theory and the underlying economic “reality”. This divergence has essentially occurred for two reasons. First, the analytical procedures which are used to establish the nature of reality in the economic sphere are actually biased against being able to do so. Second, and as a result, there is a genuine misrepresentation of - if not an actual misunderstanding about - the nature of economic systems. The analytical procedures used in economics are the same as those that were originally established for the natural sciences

by men such as Isaac Newton and Rene Descartes. These procedures presume that it is possible to understand all aspects of any complex phenomenon by “reducing” that phenomenon to its constituent parts. The process (which is known as “reductionism”) works very well in the context of everyday life. Indeed, the fund of knowledge is initially enhanced as differentiation increases, and so the process is self-justifying. However, the process eventually causes analysts to lose sight of the whole system: and knowledge of the parts ultimately becomes useless without an understanding of the relationship between the parts and the whole. This, indeed, is the criticism that can be levelled against modern economics. 5 It fails to recognise clearly that the combination of all economic activities transmutes into a cohesive system with its own set of characteristics; it effectively assumes instead that an economy, is no more than the simple aggregate of its mutually exclusive parts. 6 The problem has materialised because of an inflexible belief in the dominance of rational behaviour in human affairs. Indeed, in the last twenty years or so, the idea that economic activity is driven by rational behaviour has been progressively refined into the hypothesis of rationally formed expectations. 7 This hypothesis is based on three interlinked assumptions. The first is that individuals do not behave irrationally - that is, they do not put their hand in the fire (unless, of course, they get pleasure from doing so!). Second, individuals learn from their mistakes - that is, if they burn themselves by putting their hand in the fire, they do not willingly put their hand in again. And

This implies that scientific observations and the data obtained from those observations are themselves dependent on beliefs which are intrinsic to the theory being tested. See T. Kuhn, “The Structure of Scientific Revolutions”. University of Chicago Press, Chicago, 1962. I

Not surprisingly, the problem of confronting consensus beliefs is one that has constrained the advance of knowledge from time immemorial. It is a problem which effectively ruined the lives of men such as Galileo Galilei and Nikolei Kondratieff. 2

3

T. Kuhn, Op. Cit.

The last two paradigm shifts in economic theory were caused by major socio-economic crises. “Keynesianism” was born out of the Great Depression of 1929-32 and “Monetarism” was adopted after the Great Inflation of the early ‘Seventies. It is worthwhile observing that monetarist theory co-existed with Keynesian theory throughout the ‘Sixties, but was not embraced generally by the economics profession until a crisis occurred. 4

Modem economics is broadly referred to as “neo-classical” economics. - See also Notes 4 and 7. 5

See, for example, D. Patinkin, “Money, Interest and Prices”. Harper and Row, New York, 1965. 6

Modem economic theory is based on two hypotheses: rationallyformed expectations (after J. F. Muth) and the existance of a natural rate of unemployment (after M. Friedman). James Tobin has called it “Monetarism II”. See J. Tobin, Stabilisation policy ten years after, in “Brookings Papers on Economic Activity”, No. 1, 1980. 7


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

third, individuals arrive at their decisions independently of one another - that is, they do not put their hand in the fire simply because someone else does so, or tells them to. On this analysis, group behaviour is the simple aggregate of rational behaviour by all individuals.

a momentous phenomenon as a fact of mathematics which needs no further explanation. This is particularly true when it is applied to the behaviour of people. Why should an individual’s economic and financial behaviour converge on “average” behaviour without any apparent reason?

As presented, the rational expectations hypothesis is easy to accept, partially because it accords with most people’s idea of how individuals behave. That is, it derives from generally accepted socio-cultural beliefs about personal motivation. Nevertheless, evidence of the need for a change is already beginning to accumulate. Two examples will suffice.’ First, the phenomena of economic booms and slumps are not easily explicable within the context of rational behaviour. Why do people tend to do the same thing at the same time? Why do people’s expectations about the future converge on a common consensus which de-stabilises the economy? Edward Dewey has asked the simple question: “Are people depressed in a depression because business is poor or is business poor because people are depressed?” 8 In other words, is there a significant psychological dimension to economic activity which is a cause rather than an effect?

In fact, there appears to be a very good reason for the validity of the “law of large numbers”, which is already being explored in theoretical physics. It has been found that only a part of the identity of an electron can be obtained by looking at it in isolation. 14 In effect, an electron obtains the rest of its observed identity from its relationship with other electrons. Hence, an electron has a duality in its basic nature: on the one hand, it has a tendency towards some form of individuality; but on the other hand, it has a tendency towards participation in some form of group. 15

Second, modern economic theory cannot deal with the phenomenon of speculative excesses in financial markets. The official academic response to the 1987 equity crash is called the “Theory of Rational Bubbles”, which attempts to maintain the integrity of the rational expectations hypothesis. 9 The theory argues that a bubble can be built on irrelevant information, and that a crash will materialise when the error is discovered. However, no-one seems to have asked why a majority of market participants many of whom devote massive financial resources to research - do not recognise that the information is irrelevant in the first place! Furthermore, how can events such as the Dutch “Tulip Mania” of 1634 be regarded as the outcome of rational decisions taken by independent individuals? 10

The concept of a duality of characteristics is an extremely important one which can be applied throughout the hierarchy of nature - from the sub-atomic level to the level of the cosmos. Every part of nature has its own identity and (therefore) an apparent randomness in relation to the “whole” of which it is a part. However, each part also participates in the working of a greater whole; and, in order to achieve its own purposes, each greater whole devolves some form of order onto its lower parts. 16

E. R Dewey (with 0. Mandino), “Cycles: The Mysterious Forces that Trigger Events”. Hawthorn, New York, 1971. 8

See, for example, B. T Diba and H. I. Grossman, The theory of rational bubbles in stock prices, in “The Economic Journal”, September 1988. 9

It is almost impossible to visualise the craze to own rare tulips in a country which actually produces them en masse, but it happened. Homes and properties were sold to take advantages of the speculation. See, for example, H. B. Neill, “The Art of Contrary Thinking”. Caxton, Idaho, 10

Indeed, it has intrigued philosophers in general for thousands of years. 11

The problem of personal motivation is one that has received detailed attention from psychologists in recent years. 11 The rich variety of responses - as a given stimulus is processed through the dimensions of logic and intuition on the one hand, and through emotions on the other - is well attested. Indeed, the possible responses to any given situation are almost infinite. It is therefore virtually impossible for an outsider to predict accurately another person’s response to a particular event. 12 Hence if a person receives a windfall sum of money, it is not possible to specify exactly whether the money will (in whole, or in part) be spent, saved, given away, or even destroyed. How then do economists justify their assertion that large numbers of people respond in the same way to given stimuli? The technical answer to this question is mathematical. According to the so-called “law of large numbers” (or “law of probability”), a large number of uncertainties produces a certainty - that is, a large number of apparently random and unrelated events produces an overall outcome which is predictable. Hence, for example, it is possible to calculate very precisely the “half-life” of a group of atoms (that is, the time it takes half the atoms in a group of identical atoms to decay), even though it is impossible to predict with any accuracy when a particular atom in a group will decay. By the same analysis, it is possible to calculate the response of groups of people to particular stimuli, even though it is impossible to predict how any one particular individual will respond. 13 Interestingly, however, no mathematician (or economist for that matter) has yet to be able to explain exactly why the “law of large number” works, it is a “law” that cannot yet be deduced from first principles in mathematics. Indeed, it is assumed to be a first principle, and it is left at that. Nevertheless, the “law of large numbers” hints at something which is both amazing and important. It suggests that order develops out of random fluctuations: and there is something intrinsically unsatisfactory about treating such

Unless, that is, some sort of constraints are imposed on an individual’s allowable response. We shall discuss the nature of these constraints shortly. 12

It is, nevertheless, concluded that individual responses will tend to reflect the overall behaviour of a representative sample of people. 13

The reason is that it is impossible to locate a specific electron at a precise point in both space and time. 14

The idea that sub-atomic phenomena have a duality of characteristics also implies the presence of a third force to reconcile them. The nature of this third force (or forces) is still being investigated. 15

As a result, each unit in any hierarchical natural system has both an element of randomness and an element of predictability. If the same analysis is applied to human behaviour, it is immediately clear that unpredictable rational behaviour by individuals converges on predictable behaviour within groups, not because of the mechanistic workings of mathematics, but because of the controlling influence of the groups themselves. The influence of groups, or “crowds”, is a universal and persistent phenomenon: However, despite some remarkable research, its importance is still not fully appreciated. As long ago as the late 1890’s, Gustave Le Bon recognised that crowds were not necessarily physical structures which were constrained to a particular place and time. He argued 17 that they were psychological phenomena which were independent of location and time, and which were something different to the simple aggregate of the participants. Le Bon realised that a crowd would emerge once its members had a common cause, and that the resulting crowd would control the behaviour of its members. Subsequent theorists have validated Le Bon’s conclusions and

23


24

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

have extended his basic analysis. Sigmund Freud deduced that group members always submit to the authority of some form of leader. Arthur Koestler 18 belief observed that group members also accept group’s system uncritically, identify with other members of the same group, and consider themselves different from members of other groups. In this way, groups of people become differentiated from one another, and conditions of tension (and of potential conflict) therefore emerge. Conditions of tension encourage people to subordinate their own needs to that of their own group: their behaviour is thereby altered. Nor has the analysis just been theoretical. Research by Stanley Milgram 19 has confirmed the extremes to which people will go in obeying authority figures, while research by Henri Tajfel 20 in Bristol and Ed Diener 21 in Illinois has confirmed the influence of crowds in altering people’s perceptions of events and in altering people’s behaviour. These conclusions have quite startling implications for our understanding of economic “reality”, and hence for our understanding of economic systems and their associated problems. First, it means that the assumption of rational behaviour by independent individuals is inapplicable at the macro level. A significant part of a person’s behaviour will almost always be determined by the constraints imposed by the need to conform to common beliefs and objectives within groups. This element of behaviour is, by definition, non-rational: it is behaviour which is determined by what other people are saying and doing, rather than by internal reasoning processes. 22 Second, conformity to common beliefs and objectives within groups implies that common emotions are likely to be experienced by group members. The important point here is that individuals always compare new sense data with their beliefs. These beliefs will generally act as a barrier to the intellect, 23 so that emotions are more easily triggered. It follows, therefore, that shared beliefs will tend to stimulate shared emotions. Patriotic fervour is a very good case in point. If these concepts are applied to aggregate economic behaviour, it begins to become rather clear why widely dispersed numbers of people will tend to do the same thing at the same time. People are in fact psychologically linked to one another through social and business ties, through newspapers and television reports, and even through advertising. The resulting psychological environment generates (relatively) uniform beliefs, and stimulates (relatively) uniform emotions, about both the present and the future. Uniform emotions then translate into (more or less) uniform activity. Nevertheless, crowd psychology also requires two other factors - namely a crowd leader and conditions of tension. How do these emerge in economic and financial activity? The important point about leadership is that it need not reside in a person or persons. Specific leaders can, of course, emerge in times of intense stress. For example, the predictions of a well-established stock market forecaster may trigger a panic when the equity market is already on the verge of a sharp movement. 24 However, generally speaking, the leadership function can be satisfied by any indicator which reflects the crowd’s belief system. 25 In economic and financial activity, the appropriate indicator is the change in the prices of goods and services. This is so because such prices represent the quantitative well-being of the economic participants in one way or another. Consequently, changes in prices will stimulate crowd behaviour. There are three points to be made here. The first is that changing prices provide leadership because (for most people) thought processes are linear - that is, it is usually assumed that what has just happened will continue to happen. Therefore, a change in prices is a clear instruction for the crowd to assume that the “trend” will continue, and to respond in an appropriate manner.

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

The second point is that crowds can only recognise relatively simple items of information. Consequently, the crowd dimension to economic activity will become effective when price changes become large. The third point is that the process increasingly becomes self-fulfilling. Crowd activity stimulates a movement in prices; the movement in prices confirms the crowd’s beliefs and stimulates further activity; and this activity generates a further change in prices. It does not take a significant leap in imagination to recognise that this “feedback” effect is precisely the mechanism behind such phenomena as stock market booms, hyper-inflations, and house price spirals. Finally, conditions of tension or stress emerge in the economic and financial sphere in response to some form of fear. 26 Fear is aroused whenever a person’s perception of events does not match his or her expectations or desires concerning those events. Fear is significant in that, at extremes, it closes the mind to rational behaviour and allows people to concentrate their energies on immediate evasive action. Fear - or even fear of fear - therefore triggers group behaviour. It encourages people to “herd” together (whether physically or psychologically) for mutual protection, and to label other people as “enemies”. In the world of modern economic theory, where rational expectations prevail, the influence of fear is simply deemed not to exist. However, in reality, it is easy to see that it is almost always present. Wage negotiations are an obvious example, where conflict is based on the employees’ fear of inadequate pay rises against the employers’ fear of excessive rises. More generally, the overwhelming fear of losing a particular standard of living, or of not enjoying an improving one, can be seen as an important influence behind such phenomena as inflationary spirals, house price booms and depressions. Indeed, fear of economic disadvantage can also be seen as the main cause of pollution and the over-exploitation of natural resources. Quite simply, fear of others obtaining an advantage encourages companies to ignore the global implications of their activities. 27 And of course in stock markets, fear specifically, the fear of being left out of a rise or of being left in a fall (i.e., the fear of being wrong) - is an ever-present force. Consequently, fear and the associated crowd behaviour is a primary feature both of speculative booms and of stock market crashes. The argument that economic and financial excesses are characterised by group behaviour is an important step forward in our understanding of economic reality. However, this does not, in itself, make it predictable. Fortunately, there is a number of developments in the fields of physics and biology which point the way towards a resolution of this problem. The clue derives from the fact that Gustave Le Bon originally argued that a crowd essentially had a mind of its own. Much more recently, Gregory Bateson 28 deduced that the systemic processes which could be found within the human brain were paralleled by similar types of processes in the metabolism of all living organisms. He argued that therefore the concept of “mind” could be applied to metabolic processes as well. Erich Jantsch and others subsequently extended the concept to cover all aspects of life, including group (or crowd) behaviour. This latter is accordingly seen as having metabolic activity analogous to that of any other living organism. The new Bateson/ Jantsch methodology (which actually builds on the pioneering work of Ludwig von Bertalanffy 30 and Ervin Laszlo 31 ) is called “systems theory”. Systems theory focuses attention specifically on dynamic processes rather than physical structure. Each natural system is seen as being hierarchical and as being able to control its parts for its own purposes - i.e., it is “self-organising”. Each such system is responsive to disequilibrium, is open to the environment for the transformation of information and energy, and is able to process that information and energy. In applying these concepts to economic and financial systems, we find that there


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

are two important aspects to activity - both of which are in contradistinction to the presumptions of mainstream economic theory. The first is that activity does not automatically converge on static equilibrium. 32 Instead, the system exists in a condition of dynamic disequilibrium where it is continuously responding to discrepancies between itself and its own particular niche in the environment. The second important aspect is that activity involves oscillations. The oscillations materialise because of the influence of “feedback” effects. Indeed, feedback is regarded as the main source of overall stability in any self-organising system. It is the way that different parts of the system transmit information to one another so that they can co-evolve. Such co-evolution occurs at the same level of a hierarchy (such as in the case of, say, car and steel manufacturing) and at different levels of a hierarchy (such as in the case of the car manufacturing and total economic activity). The result is what has been called “order through fluctuation”. 33 Order through fluctuation is the basic concept which allows the apparently random fluctuations of lower-level systems to co-exist with the observed order of higher-level systems. It is the essential reason why the “law of large numbers” actually works. In theory, of course, this analysis should apply to all levels of a particular dynamic structure. That is, every level should appear random to higher levels, but should appear stable to lower levels. However, some confusion has recently emerged because of the fact that, in some lower-level systems, the hierarchical relationship seems to be reversed. That is, the higher-level system seems to devolve chaos onto the lower-levels. These chaotic systems have been the focus of a great deal of research in recent years, and are now included in a loosely-knit discipline known as “Chaos Theory”. 34 Whereas systems theory chooses to analyse natural processes from the perspective of stability (i.e. “top down”), Chaos Theory analyses processes from the perspective of the ‘inherent fluctuations (i.e. “bottom up”). The original Chaos Theory was that a small initial perturbation to a system could, if left unchecked, lead to ever-increasing fluctuations. 35 This, in turn, could lead to the actual breakdown of some systems. 36 However, subsequent work on fluid dynamics has revealed that the transition from obvious stability to apparent chaos need not involve a systemic collapse. 37 Rather, chaos merely involves an alteration in the way that a system expresses itself. Furthermore, genuinely chaotic conditions in nature can only arise if the self-organising processes involved actually break down. At this stage, genuine evolution through revolution takes place. 38 Finally, research into the geometry of natural structures - known as fractal geometry - confirms that apparently random patterns can reproduce themselves continuously at different hierarchical levels of a particular structure. 39 The important point, then, is that genuine chaos does not exist in Nature, and only appears to exist at points of transition. This implies that, except at these points of transition, financial and economic activity should essentially be forecast-able. It is of particular relevance that the processes involved can be demonstrated with the help of a very simple diagram. It is a limit cycle. 40 Limit cycles describe the co-evolution through time of two interdependent variables. If the limit cycle is stable then oscillations converge on it from a wide range of initial states. Chart 1 shows the limit cycle between two variables - x and y - without any reference to time. Chart 2 shows the same cycle extended into the third dimension of time. Chart 3 shows the cycle in terms of only one of the variables, plus time. 41 When these concepts are applied to financial markets, it is immediately possible to obtain important new insights into the way that they behave. One of the relevant limit cycles in any financial

Psychology and Markets

Systematic Trading

Chart 1

market is that relating price changes to investor sentiment. 42 See Chart 4. Price changes and sentiment move together during the main part of a trend, with no clear cut direction of causation. Analysts recognise this situation because indices of investor activity, such as volumes, tend to rise. At turning points, however, the relationship breaks down. The limit cycle is biased to the right. Changing prices cannot stimulate further investor involvement because the market has become overbought or oversold. This is the first reason why market analysts refer to the concept of “non-confirmation”, which heralds a price reversal. However, behaviour does not stay on the limit cycle. As profittaking develops, prices change direction and a “shock” is delivered to the limit cycle system. There is a “one-off” change in prices and sentiment inwards from the limit cycle. See Chart 5. This is what happened during the 1987 equity crash. Next, the shock sets up a series of potentially de-stabilising oscillations, caused by negative feedback. See Chart 6. These oscillations translate into a spiral through time. See Chart 7. The spiral continues until the self-organising limit cycle processes re-assert themselves. See Chart 8. In financial markets, the spiral movement involves a “re-test” of the price levels from which the shock occurred. This re-test may actually involve a new high in a bull market, or a new low in a bear market. Whether it does or not, however, it is not normally accompanied by a strong improvement in the indicators of investment sentiment, such as volumes. This is the second reason why analysts refer to the concept of “non-confirmation” to herald a price reversal. The subsequent move will be a dynamic thrust in the direction indicated by the limit cycle. 43

34

See, for example, J. Gleick, “Chaos”. Sphere Books, London, 1987

This is the so-called “Butterfly Effect”, where it is argued that the “flap of a butterfly’s wings in Brazil might set off a tornado in Texas”. E. N. Lorenz, Address to the American Association for the Advancement of Science, Washington, 29th December 1979. 35

M. J. Feigenbaum, “Universal Behaviour in Non-linear Systems”. Los Alamos Science, No. 1, Summer 1980. Feigenbaum, together with R. M. May, explored the effects of “period-doubling” on the road to genuine chaos. They found that the periodicity of a fluctuation could double every so often, until the system broke down. 36

25


26

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Chart 3 37

See J. Gleick, Op. Cit.

See Fritjof Capra, “The Turning Point”. Wildwood House, London, 1982. 38

B. Mandelbrot, “The Fractal Geometry of Nature”. Freeman, New York, 1977. 39

A limit cycle is defined as the (isolated) periodic oscillation between two variables, and is represented graphically by an (isolated) closed non-linear path. See, for example, D. W. Jordan and P. Smith, “Non-linear Ordinary Differential Equations”. Oxford University Press, Oxford, 1977. 40

Chart 3, therefore, demonstrates how rhythmic cycles (i.e., with regular peaks and troughs) occur. 41

Sentiment in this context is difficult to define precisely. However, it can theoretically be measured as the ratio of the number of bulls to the number of bears. 42

The implication obviously is that the equity markets still have to negotiate an important bear phase. 43

This simple model, which develops into a basic “three-up, threedown” construct, can be used to explain a large number of wellknown market phenomena. It can, for example, be used to explain all the price patterns which have been known to traders for many years. This occurs because of the “nesting” effect of the threephase pattern at all levels. The “chaos” aspect of self-organising systems argues that these patterns will be difficult to forecast while the markets are in the process of changing trend. However, the “fractal” aspect of the same systems suggests that basic rules will still always apply.

Chart 4

Finally, once a reversal pattern has resolved itself, the operation of the spiral mechanism means that the subsequent price movement can be measured with an incredible degree of accuracy. This is because spirals can be mathematically defined: and in the case of financial markets, the most important spiral is the “Golden Spiral” based on the number 1.618. And of course, the model confirms the presence of rhythmic oscillations, so that it is possible to forecast the duration of price movements.44 Significantly, if this is true of financial markets, then it must also be true of the markets for goods and services. There is therefore tremendous scope for research into economic activity, using the hypothesis of nonrational behaviour.

Chart 2

Chart 5

Pattern Recognition and Pattern Analysis

Moving Averages and Trends


Indicators and Momentum

Chart 6

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

The structure of price patterns, the extent of price movements, and the duration of price trends are discussed in T. Plummer, “Forecasting Financial Markets: The Truth Behind Technical Analysis”. Kogan Page, London 1989. 44

The conclusions which follow from this analysis are simple, though profound. First, economic activity has a very significant psychological dimension which is almost completely ignored by economic theory, and hence by economic policy. Second, the aim of economic analysis should be to incorporate the phenomenon of financial and economic excesses into a model of reality which sees those excesses as part of a spectrum of behaviour. The way forward may well be to research into the behaviour of stock markets, where ordered fluctuations have been recognised and measured by traders for many years, but where economic theory presumes that chaos exists.

Chart 7

Third, excessive fluctuation in economic variables such as prices and employment, can be avoided if efforts are made to minimise those policy changes which might trigger and perpetuate a crowdtype environment. The aim of Government, therefore, should be to concentrate on long-term policy objectives. Fourth, and for the future, it will be necessary to reduce the fear of “missing out” on a quantitative improvement in living standards, and increase the fear of triggering a global catastrophe. Without this, the exploitation of natural resources and the associated pollution may well take us all to a logical conclusion!

Chart 8

Tony Plummer was formerly a Director of Hambros Bank and responsible for currency and sterling fixed interest positioning.

27


28

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Art or science? Ruminations on the meaning of Technical Analysis Tim Parker MSTA

Article originally featured in Market Technician 65 (July 2009)

A summary to a talk given to the STA on 10th March, 2009 This article offers some thoughts on the philosophy of technical analysis which have been tucked away in the back of my mind for a number of years while I’ve been making a living, day-to-day as a technical analyst and salesman. I should declare at the outset that philosophy is not my strong point and neither is the statistical manipulation of data, nor the dissection of financial statements: my educational background is in English literature as a graduate of Cambridge University. After that, my rite of passage into adult life was a three-year stint as a soldier, some of which was spent on the streets of West Belfast and some on guarding Buckingham Palace. It may seem an unlikely background to a career in technical analysis but, after 20 years of eye-balling charts and engaging institutional investors in the attempt to understand the dynamics of financial markets, it is clearer to me than ever that academic achievement does not guarantee investment success. No-one, of course, has a monopoly on the best way to make money in the financial markets but some people have better tools than others or, more accurately, they use the same tools as others - but more effectively. One of these tool-sets, and the nature of the practitioners who use it, is the subject of this article. Is technical analysis an art or a science. Are the people who use it artists or mechanics? The point about discipline is frequently made. It means obeying rules and following a process of controlled behaviour after a period of training. It is essential to master one’s emotions while trading financial assets. Technical analysis is often seen as exactly that - a system of rules that, once learnt, guards the investor against the inconvenient influences of prejudice and emotional involvement while trading. In behavioural jargon, discipline confronts and negates the biases inherent in all of us. You might even argue that the discipline of technical analysis explicitly involves a degree of punishment because, if you disobey the rules and allow your heart to hold sway over your mind, you will get hurt financially. The definition of technical analysis shown in the box above is rather a mouthful and fairly meaningless to those not already well versed in the subject. However, there is one word in the definition that stands out because it challenges the whole notion of technical analysis being an inviolable rules-based discipline and that is ‘suggest’. The ‘suggestion’ of future price activity, rather than an unambiguous formulaic output is what shifts the whole philosophy of technical analysis out of the realm of science and firmly into that of art. It also casts the analyst as more of an artist than a scientist. There are two types of price pattern on the charts, reversal and continuation. It is obviously important to know which is which. For example, Chart 1 of BAE Systems could at one point have been interpreted as a pennant, indicating a bearish continuation pattern yet in the event it turned out to be a reversal.

Chart 1: Ambiguous patterns

Of course good analysis does not end with the recognition of a pattern on one chart. Many other factors such as different time periods and supporting indicators can be brought to bear, as well as Elliot Waves, Gann, Fibonacci and inter-market analysis. A meaning or forecast can emerge quite powerfully from such comprehensive analyses, but the point is still that it does not necessarily follow that it will always emerge - an ambiguity will often exist. Information derived from the study of market action can have forecasting value but judgement may be needed to extract the best of this value and could well make the difference between winning and losing. The judgement of information is as important as the information itself. An analyst’s judgement is subjective and depends on his or her skill and intuition in reading and interpreting the lines and patterns on a chart, much like the critic of art (of which more later). An analyst’s judgement is also, however, dynamic and constantly changing, which is highly appropriate for the inherently dynamic environment of the markets themselves. On 29th December 2008 John Kay wrote an article in the Financial Times which included the following paragraph: ‘Economic systems are also dynamic. Dynamic in the sense that they evolve - which makes the mathematics harder. But also dynamic in the sense that the structural relationships constantly change. Some economists believe there is a deep underlying structure from which laws of economic behaviour that are universal in time and space can be deduced. I think that search is a wild goose chase and that the best we can do is to identify empirical regularities that apply to particular contexts. Whoever is right, it is evident more work needs to be done in understanding the relationships.’


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

A technical analyst might argue that any technical approach is a collection of laws about price behaviour “that are universal in time and space” and reflect a “deep underlying structure”. But, if we consider one of the most respected techniques of technical analysis, Dow Theory, it turns out to be flawed when incorporated into mechanised trading systems that involve, for instance, automated trend-line drawing and automated trend-break trading signals. One of the reasons for the disappointing results derived from these systems is that, intuitively, we know that in bull markets resistance lines do not work well and in bear markets support lines do not contain selling pressure. In addition, a trend may end and be followed by a sideways market of unknown duration and unknown future direction - such as the one, some would say, that hangs over range-trading equities right now. In these instances reversal indicators may simply fail while volume and momentum just fade away. Overall, trend followers are usually late and contrarians are usually early. In all these circumstances what we need to identify early on is the ‘particular context’, or what type of market we are in, so that we can adopt the appropriate technical approach. In other words, if the markets are displaying momentum characteristics, then we could or should use Dow trend theory but, if markets are ranging, an array of suitable oscillators would be more appropriate.

Psychology and Markets

Systematic Trading

Chart 2: A perfect double top?

Chart 3: Imperfect pattern

The point is that a more flexible, dynamic and intuitive response by the analyst as market interpreter is needed to identify the inflection points between bull and bear markets.

The study and recognition of price patterns Patterns are at the epicentre of technical analysis. They help us identify vital inflection points when markets are transitional or in an intermediary phase. The recognition of chart patterns is absolutely key to raising the probability of successfully identifying turning points but it is subjective - as expressed by that phrase ‘suggests future activity’ in the definition of technical analysis. Experienced practitioners recognise a small number of perfectly formed patterns on a regular basis, such as the double top on the S&P 500 index shown in Chart 2, with often satisfying results. But what about the rest of the time? At any one time more than 80% of price charts will be frustrating and confounding efforts to define classic patterns from the data. I believe that a large proportion of these charts are displaying ‘partial patterns’ which can prove to be very profitable. The outcome of the partial and imperfect ‘head and shoulders’ pattern shown on Chart 3 was, for example, unambiguous. The theory goes that a pattern raises the probability that a price is likely to move in one direction or another after completion of the pattern. But the process of recognition, or rather the experience of the analyst scrutinising a partial pattern, is also a key factor in raising the probability of success. In fact, it is true to say that recognition of the imperfect pattern - which only a handful of market participants see and understand - is better and more fruitful than the highly visible pattern, which is more universally recognised and therefore invalidated or arbitraged away. The BAE Systems example shown in Chart 2 is a good example, as is the V-reversal pattern which did not work on Carnival (Chart 4). What I term the ‘recognitive process’ relies on the experience and intuition of the analyst from which we can infer that the analyst has an interpretative skill which he brings to bear. He exercises judgement as well as calculation, and therefore practises as much as an artist as a scientist. In these cases the rules of the analysis in other words the tool-set - are more than matched by the quality of the experience and intuitive power of the analyst.

Chart 4: Imperfect result

The evolution of patterns As John Kay and other analysts have observed, economic systems and markets constantly evolve or, as the Latin origin of the word would have it, constantly ‘unroll’. In any process of gradual development there are always vestiges of a past form in the present form. It is for this reason that we can expect to find value in recognising certain patterns of historic data which, more often

29


30

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

than not, have led to predictable outcomes. Evolution, however, does not stop in the here and now. It goes on and on, ‘unrolling’ and ‘unravelling’, constantly developing new forms. It has occurred to me that recorded patterns of historic data do so as well; they go on evolving. Technical analysts should, therefore, be looking for patterns of data that have developed from earlier patterns, or for mutations or even for completely new patterns. ‘Partial’ pattern recognition becomes ‘new’ pattern recognition, with new rules and new identifiable characteristics. The chart of Tullow Oil (Chart 5) is an example of a pattern which we have labelled a ‘CRAB’, an acronym for ‘Confirmed Rally After a Base’. Similar to the ‘Cup and Handle’ it offers a good entry point for a long position after the break out of a base formation. An attempt at back-testing might help secure credibility in a new pattern, but all too often this is self-defeating. Rather like stresstesting the banks in the US, it would all depend on the arbitrary and subjective selection of the rules of the back-test itself. The concrete and standardised definition of the new pattern would be elusive, especially if you tried to programme it in a formulaic way. Chart 5: Tullow ‘CRAB’

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

the analyst can apply a technique to the empirical regularities of what he observes, he must first recognise them and identify the context in which they operate. Markets are dynamic, yes; they evolve, I think so; they are repetitive, well, I’m not so sure... Let us examine one of the three premises on which John Murphy claimed the technical approach is based.

History repeats itself When a young, creative and ambitious artist studies the painting of, for example, one of the Renaissance Masters, he doesn’t just copy the work. A copy would look nice and sell for a modest sum but would condemn the artist to a life of plagiarism. The true artist copies first, then adapts, creates, innovates and paints something new. In doing so he undermines the conventional wisdom of the moment and discovers something different - just like the high-scoring technical analyst. It is no coincidence that the artist Damien Hirst, the arch capitalist, is very, very rich; richer than 99% of technical analysts! He has discovered something, or expressed something, very new, and has created (perhaps incredibly) enormous value. My contention is that markets are dynamic in an evolutionary sense and that technical patterns are too - evolving according to a constantly adapting set of rules. The TD POQ pattern, introduced by Trevor Neil in a talk to the STA earlier this year (see Market Technician Issue No 64) was a perfect example - a quasicandlestick pattern without a name. Of course a common denominator in financial markets is the human nature of the participants and it is the human activity, with all its rational and irrational behavioural elements, which creates the technical patterns that technical analysts attempt to interpret. It follows, therefore, that a consideration of one or two aspects of behavioural psychology might be helpful in this enquiry.

From patterns to psychology Dynamics The concept of dynamics in pattern recognition was mentioned earlier but it merits further examination. ‘Dynamics’ is a branch of mechanics concerned with the motion of bodies under the action of forces; or the forces which stimulate change within a system or process. I would argue that the ‘assumption’ that patterns will work well in the future is misplaced. Does the idea of a ‘dynamic’ market imply that markets produce the dynamics of repetition in price movements and, therefore, are eminently predictable or that the dynamics of the market produce a constantly moving evolution? The unrolling process of gradual development from one form to the next, explicitly denies the concept of simple ongoing, repetition. In the real world - and practical experience to some extent backs this up - patterns repeat themselves from time to time, but they can then change as the evolving market absorbs, understands and arbitrages away the full predicted outcome of the pattern. In this process the role of the analyst is paramount. In trying to define in a semantic sense what a technical analyst really is, it could be argued, that as dynamics are a branch of mechanics (and therefore scientific), and that as the market is a system in which change is stimulated by the forces of dynamics so, therefore, the analyst is truly a scientist and an expert in the systematic study of the structure and behaviour of the physical world. However, before

One of the most powerful arguments against the notion that new patterns are constantly forming in price data is that human psychology does not really change over time, and that the investor and the investment crowd react, time and again, in the same semiferal way. It is true that human psychology does not change very quickly, indeed the process of evolution of the human brain is somewhat slower than the evolution of technology, but the study of psychology is always changing and evolves quite quickly. The process of ‘study’ is the effort to try and find truth, meaning, vision or enlightenment or - more prosaically in the financial world being the first to find the answer to the question of how to understand and beat the market, and therefore make money. The constant effort of study means that the market has to stay one step ahead. As conventions are studied, understood and arbitraged, so they change because the brain has recognised new things, such as, new patterns. Conventions change just as patterns change. I will concede that it is likely that some parts of our psychological make-up do not change. As far as I understand it, there are embedded processes in the brain which are NOT learned... fear and flight, appetite and greed, for instance, are obvious ones derived from our anthropological beginnings as hominids on the savannah concerned primarily with survival and replication. These responses to the environment are largely instinctive and emotional, and do not originate from, or allow, the luxury of rational


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

thought and considered action. In this sense our psychological make-up can be dangerous to our wallets as we sometimes get carried away by our own emotions or the collective emotions of the crowd. Technical analysis attempts to excise emotion from our assessment of price action by reducing it to the so-called objective analysis of price and volume as it is delivered by the market. But, as I have implied earlier, I have a problem with the ‘objectivity’ of the process: it is, in fact, very subjective. The best way of thinking about this theory is to reverse it. Instead of saying that, as technicians, we eradicate emotion by objectively observing and recording price and volume data, we should say that we extract emotional output from the analysis of price and volume data. In other words, understand the emotional output of the market crowd. That may sound odd, but there are a number of market players who do just that. My old colleague and friend David Fuller, for example, calls his approach ‘behavioural technical analysis’, closely linking the manic/depressive nature of crowd psychology with price action itself. I think his linkage between psychology and price action is apposite but would define it differently. You have, on the one hand, a collection of investor emotions which can be deduced from the price action on the charts but, on the other, the deductive process itself involves subjectivity and an emotional response in the analyst himself... “am I scared by the break of the short term uptrend, or am I sanguine about the persistence of the primary trend on the longer term chart?” Therefore, there are two levels at which information about emotions play a part - the data itself and the analyst. It is in the assessment or judgement of these different and complex levels of emotion that the analyst succeeds or fails - and therefore it is not a science, it is art. Of course, the idea that technical analysis is an ‘art’ can be, perhaps, pushed too far, particularly when we consider the psychology in a bit more detail. Is ‘emotion’ the right description of what constitutes investor and market sentiment? For example, the embedded processes in our brain that are not learned could perhaps be described as ‘feral responses’ rather than ‘emotions’ and it is often our hardwired primitivism that shines through in the way we behave in markets. In fact the key emotions in the market place are probably fairly few and are more to do with fight, flight and hunger than the finer gradations such as expressions of love, melancholia or humour. These latter emotions are truly in the realm of high art rather than high finance. Our in-built responses I would characterise as ‘innate’, while our learned responses are ‘intuitive’. As an illustration of how the brain does both of these things at the same time consider the following. On the ‘innate’ side, or the side of embedded processes, is normal eye perception. The human eye moves, on average, three times a second and forgets every move. However, through a process called ‘confabulation’ the brain recognises patterns in what it sees and creates a narrative that purveys meaning. In the same way good technical analysts see, perceive or recognise patterns in price and volume data which he or she will understand as a narrative with a forecastable ending. A price chart plotting data three times a second would appear meaningless second by second but, at some point, the brain will make sense of it and discover narrative value. The better analyst, of course, would have a better story to tell. To underscore the point, the brain is constantly seeking a story by inference and by pattern recognition. The market is a narrative, but the narrator is elusive; the plot is largely unpredictable and the characters are many and varied. The simple psychological set-up is that the investor is always trying to anticipate the response of every other investor. In this sense the investor is being subjective, like an artist, rather than objective, like a scientist. Cognitive scientists call this the ‘mirror’ system in the brain: a neuronal system that operates by allowing us to

Psychology and Markets

Systematic Trading

understand others by what is going on in our own brain; in other words to get inside other people’s heads.

Psychology Another psychological function that is worth exploring is that of the mental short cuts we make, or ‘judgement heuristics’ to use the jargon. These are the principles or methods by which we make assessments of probability simpler. We automatically tend to make mental short-cuts to save time and effort but these can often lead to biases which tend to persist such as the overvaluation of dotcom stocks in 1999. We knew they were overvalued but we believed we would make money if we stayed in them because prices were going up despite increasingly ludicrous valuations. The ‘new paradigm’ had arrived. Going into a little more detail I would like to outline why heuristics are often useful to us but sometimes lead to systematic errors. For example, some types of judgement heuristic are as follows: 1. The representativeness heuristic - an event is judged to be probable to the extent that it represents the essential features of its generating process. So, for example, we might simply extrapolate performance of an extreme move in a share price. This might systematically lead one to make poor judgements by making the assumption that extreme instances are representative of future instances. This is true in any bubble from tulips to houses. 2. The availability heuristic - one’s judgement about the relative frequency of an event often depends on the availability of events in the process of perception, memory or construction in the imagination. Again, the use of this heuristic can systematically lead one to make poor judgements by making events that easily come to mind seem more likely than they are. 3. Adjustment and anchoring phenomena i.

Conservatism is sometimes recommended when adjusting our beliefs or methods in the light of new information. A well established belief/method should be overthrown only when one has solid evidence against it. An otherwise reliable method should be changed only when it meets significant failure. In socio-economic terms we can think of Communism disintegrating only when proof of its failure was beyond doubt and American-style Capitalism now seriously questioned but only after significant recent failures.

ii. Insufficient adjustment due to anchoring can lead to mistakes: • Sometimes reasoners hold fast to some piece of information and ignore the consequences of additional information. E.g. The Madoff whistleblower and the SEC. •

When solving a problem involving probabilities, reasoners may start with an initial value and adjust it to reach a final value. The anchoring phenomenon occurs when their results are biased toward the initial value.

• Sometimes reasoners anchor to an initial problem-solving method when a completely different approach would be more effective. 4. Risk and loss aversion - Many people are averse to taking risks. They tend not to bet £500 on a 50% chance of winning £1,000, even though that is the fair price. But studies show that people would rather take a risk than suffer a loss. Equivalent problems get different responses depending on

31


32

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

whether the problem is framed in terms of losses or gains. For example: Choose between: a) a sure gain of £3000, and b) an 80% chance of winning £4000 and a 20% chance of winning nothing.

Also:

Pattern Recognition and Pattern Analysis

world, investors). He or she recognises and re-orders the emotions in others, which is a process in which his or her own identity and personality are central. By contrast, the scientist comes up with a hypothesis and tries to find truths about the world independent of the person. They find truth in the world which is ‘out there’ and objective, that is, outside the realms of people’s emotions. Scientists are, therefore, explicitly not trying to influence other people in the same way. I would suggest that financial markets are both ‘inside’ and ‘outside’ people and technical analysis is an attempt to bridge the gap with both art and science playing their parts.

Choose between: a) a sure loss of £3000, and b) an 80% chance of losing £4000 and a 20% chance of losing nothing.

The point about these examples is would you have altered your choice based on the different ways the question was framed despite the identical probabilities?

Technical analysis and the inter-relationships of psychology, markets and economics Finally, it is useful to look at how psychology interrelates with the markets and economics. Human psychology tries to mechanise survival in a chaotic physical environment while technical analysis attempts to mechanise human psychology in a chaotic market environment. In this sense it is possible to see the link between technical analysis and a chaotic environment - it is trying to create order and extract information and meaning from it. Therefore the mechanising process of technical analysis is similar in function to the emotional responses of survival. Both cause action, action causes change, and the economic effect of change is profit or loss. If you can assess accurately the probability of how things will change you will make money. The economic forces of supply and demand are the changes derived from the mechanics of human emotions, or more accurately, the mechanics of psychology. And the link from economics to technical analysis is through supply and demand. Intrinsic value is based on a balance between these forces, at which point a price is established. It then goes on to fluctuate as the balance shifts. The purpose of technical analysis is to suggest, through recognition of price trends and patterns, when inflection points have been reached and therefore when the assessment of intrinsic value may change.

Technical analysis: art or science There is a crucial difference between the functions of ‘analysis’ and the functions of an ‘analyst’. The former derives from received theory and the attempt to make hard and fast rules in order to formalise or standardise the methods of the discipline. The latter, being human, does the opposite. He or she is attempting to break things down into their component parts in order to explain them and it is this process of exploration and explanation that is truly artistic. In literature this has been explicit in movements such as the Structuralists in the early 1980s, who attempted to break down writing and reconstruct it as a mechanical series. The difference between artists and scientists is that artists get people to discover the truth by means of emotional responses in other people, that is, they influence people (or, in the financial

Moving Averages and Trends

Tim Parker, PH Partners Ltd www.phanalysis.com


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

33


34

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

CHAPTER T WO

Dow Theory, Wyckoff and Volume Articles in this chapter 36

Dow Theory Bill Adlard

40

The Dow Theory and other Sell Signals Peter Beuttell

42

Wyckoff Laws and Tests Dr. Hank Pruden and Dr. Bernard Belletante

45

Unravelling the DNA of the Market: Applying the Double Helix Framework to Wyckoff and Elliott Henry O. “Hank” Pruden

52

The warning signs of Major Market Tops (Wyckoff) Paul Desmond


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

CHAPTER T WO INTRODUCTION

By Luise Kliem FSTA

Introduction Charles Henry Dow (1851-1902), generally acknowledged to be the ‘father’ of technical analysis in the western world, did not set out to develop a ‘Theory.’ As a financial journalist, he went into partnership with Edward Jones to form Dow, Jones & Company, which by 1885 was publishing The Wall Street Journal. During this time, Charles Dow created the first ever stock index, which he called an average, and which was subsequently split into the Dow Jones Industrial Average and the Dow Jones Railroads Average. The latter, of course, is now known as the Dow Jones Transportation Average. The combination of journalism with the creation of these averages allowed Dow to write extensively about stock market behaviour in his WSJ editorials, but the eventual Dow Theory evolved through the work of Samuel Nelson, William Peter Hamilton and Robert Rhea. And it was Rhea who in 1932 published The Dow Theory - with the rather indigestible subtitle An Explanation of its Development and an Attempt to define its Usefulness as an Aid to Speculation. Only here do we see the final version of the six basic principles: the Averages discount everything; there are three trends in the market; major trends have three phases; the Industrial Average and the Transportation Average must confirm each other; volume must confirm trend; and: a trend remains in effect until a reversal has been signalled by both Averages. The requirements for trend change confirmation have often given rise to the criticism that Dow Theory signals are ‘late’. But critics tend to forget the crucial point: Dow signals are intended to warn of the emergence of bull and bear trends - and trends take a little time to develop. Criticisms in the wake of Robert Rhea’s publication by the economist Alfred Cowles (in a 1933 issue of Econometrica) are well known, but fortunately so is a 1998 study by Brown, Goetzmann and Kumar (a NYU Stern School of Business paper) which completely contradicts Cowles. Applying more modern portfolio analysis in their re-examination of Hamilton’s 1902 - 1929 WSJ editorials they conclude that the suggested timing strategies yield high Sharpe ratios and positive alphas. Martin Pring, in his highly regarded book Technical Analysis Explained, calculates a quite remarkable Dow Theory success rate for the 1897 - 2009 period, although warns that this partly results from his particular interpretation. A contemporary of Dow was Richard Demille Wyckoff (1873 - 1934), who began his Wall Street career in 1888 as a runner. Having successfully traded his own account for several years, he opened a brokerage house and began to publish research in 1909. The creation of The Magazine of Wall Street was followed by two books on his methodology: Studies in Tape Reading (1910) and How I Trade and Invest in Stocks and Bonds (1924). He developed guidelines involving market cycle analysis not dissimilar to Dow: the Wyckoff market cycle describes accumulation, mark up, distribution and mark down phases. But he goes further than Dow’s suggested retracement areas: he works with price projections involving bar charts and point and figure chart horizontal counts. Many more aspects of the Wyckoff approach, too many for the purposes of an introduction,

Psychology and Markets

Systematic Trading

can be found. The Wyckoff Institute website may be worth a visit: https://wyckoffsmi.com The focus on Dow and Wyckoff in this chapter makes it an equity-orientated one, logically leading on to an examination of the role of volume and market breadth as indicators of ‘the health of the market’. Since Dow’s day, volume has been considered as something that should underpin the prevailing trend. Post Dow, with the ready availability of computer software, technicians have been able to develop increasingly complex tools involving money flow and market breadth. The former essentially involves the combination of price with volume, while the latter measures the number of stocks participating in an up or down move of the market, thus offering evidence of trend strength or weakness.

Our chosen Articles Our articles on Dow Theory begin with the remarkably informative and insightful Dow Theory by Bill Adlard. Bill, a technician with many years’ experience to draw on, provides the reader with a detailed and very ‘digestible’ description of Dow, interspersed at times with his own insights and experiences. Students new to technical analysis arguably need not go much further with their initial Dow studies! This is followed by Peter Beuttell’s contribution, entitled The Dow Theory and other Sell Signals. Here we are offered an interesting reminder that even experienced technicians can differ in their interpretations of Dow signals. Luckily, Peter puts us right with regard to the signal under discussion at the time. Richard Wyckoff then receives well deserved attention from the eminent Dr. Hank Pruden and Dr. Bernard Belletante, both experienced market technicians as well as holders of senior positions in the world of academia. An article as detailed and useful to the student of Wyckoff as the first one is on Dow, this authoritative piece (entitled Wyckoff Laws and Tests) also examines the ‘three Wyckoff laws’ and ‘nine classic tests for accumulation.’ A forecast for the Dow is offered, and there is the promise of a ‘what actually happened’ report the following year. The forecast in this 2004 article was that the Dow would continue its bull market, with potential to reach 14,400. So what did happen? The Dow reached an intra-day high of 14,198 on 11th October 2007. Subsequently, the financial crisis saw the index collapse to below 6500 - but I don’t think we can blame the authors for not forecasting that in 2004! Dr. Pruden then goes on to provide us with an article which he tells us was inspired by the 2014 IFTA conference, hosted in London by the STA. The theme for the conference had been ‘Unravelling the DNA of the Market’ and this inspired the author to consider the application of the double helix framework to Wyckoff and Elliott. Put more simply, he provides us with an article outlining how the combination of Wyckoff with the Elliott Wave Principle can be a powerful investment tool. We move on to measures of the ‘health of the market’ with Paul F. Desmond, president of the highly respected Lowry Research Corporation. In a superb article entitled The warning signs of major market tops he fully explains a number of equity market indicators that we would all be wise to understand: new 52-week highs, the advancedecline line, and market divergences to mention just three. The subsequent articles, Volume and Volatility, Money Flow, Volume leads Price, and Pattern Recognition are a little older but nonetheless offer important insights and immensely useful research to our readers. In the world of technical analysis, especially, age is no barrier to usefulness - proof of that, if proof were needed, is the undiminished influence of Charles Dow and Richard Wyckoff.

35


36

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Dow Theory Bill Adlard

Article originally featured in Market Technician 53 (July 2005)

Summary of a presentation given to the STA on 9th March, 2005 Charles Dow was born in 1851, in Sterling, Connecticut, the son of a farmer. He became a journalist, and in the late 1870’s began specializing in financial reporting, particularly as a mining expert. He moved to Wall Street in 1880, and met Edward D. Jones while working for the Kiernan News Agency. In 1882 they set up Dow Jones and Company as a financial news agency. In 1885 Dow became a member of the New York Stock Exchange. From 1885 to 1901 he was a partner in a firm of stockbrokers, and was for several years the firm’s floor broker. In 1889 Dow Jones & Co founded the Wall Street Journal and Dow became its first editor. He died in 1902. The general perception in those days was that bonds were a better bet than common stocks, i.e. equities. There was, in fact, very little information available to the public about common stocks and the companies behind them. Daily tables of stock prices did not exist. Information about a company’s balance sheet was rarely published and, if it was, the management would attempt to obscure the full value of their company for fear of a takeover. There was a recognition that you could make a lot of money from stocks, but a feeling that they were volatile and unpredictable, and you could easily lose your shirt. Bonds had the great advantage that they were secured by the assets of the business, there was a fixed and regular coupon, and they were redeemable at a fixed point in time. You did not need to know a great deal about the business, only what its assets were. The challenge, and opportunity, therefore, for Dow Jones & Co was to find a way of making information about companies and stock prices more widely available and predictable. Dow Jones and Co was based near the New York Stock Exchange, and its original product was a handwritten news-sheet called the Customers’ Afternoon Letter which was distributed around Wall Street daily by messengers. The Letter was revolutionary in that it not only contained daily stock price tables, but also made public quarterly and annual financial information regarding companies - information that only insiders had available to them before this. The Letter evolved into the Wall Street Journal when the first issue of the WSJ appeared on July 8th 1889. For many years this remained one of the most important, and only, sources of financial information for investors. It would not be until the Securities Act of 1934 that companies would be required to file quarterly and annual reports that all investors could look at. Dow’s analytical techniques probably arose, therefore, from the need to find a way to forecast the economy, and therefore stock prices, in an era when important data was only available to a small number of insiders, who would keep it secret and use it to manipulate the market to their own advantage. The market was unregulated, and there was nothing, short of outright fraud, to

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

prevent them doing so. Dow had experimented with averages since the 1870’s, and probably invented the Dow Jones Average in 1884, as an average of 11 companies, mainly railroads. Not many industrial companies were then publicly quoted. The big growth businesses of the day were the railroads. In 1896 he split it into two separate averages, one consisting only of railroad companies and the other only of industrial companies. These are still with us today as the Dow Jones Transportation Average and the Dow Jones Industrial Average. The thinking behind this was that if knowledgeable insiders were enthusiastic about prospects for goods production, stock prices of goods production companies would rise. But if that was not matched by similar rises, caused by similar buying by insiders, in the prices of railroad companies (then the main means of distribution), it would warn of an impending change in economic conditions. He only used closing prices, probably because of fear of “manipulation” in intraday dealings. Dow discovered that the trends in the market averages did indeed lead the economy. He put it thus: “The price trend is not saying what the condition of business is today, but what it will be months from now.” This is still as true today as it ever was. We should not let the plethora of fundamental information we have today blind us to this simple truth. As editor of the newly founded Wall Street Journal, Dow wrote a series of editorials between 1899 and his death in 1902 interpreting the movements of his averages and forecasting the economy. His writings were widely studied after his death, and “Dow Theory” emerged. A couple of historical footnotes: Dow Jones & Co. employed Clarence W. Barron, from Boston, as an out-of-town correspondent. In 1902, on Dow’s death, he bought the company, and later launched the monthly Barron’s magazine, which, like the WSJ, is still with us. Dow Jones & Co was also responsible for the invention and introduction of the “tape” for a more instantaneous read-out of intraday stock prices. Dow thought that the reason the averages lead the economy was because people with price sensitive knowledge would act in their own interest and cause the market to be priced accordingly well before that information become public, and that was undoubtedly true in his time. However, the underlying emotional state, or mood of people generally, is also a very influential component, from short to ultra long time-frames. Mood and market prices are very closely related. Rising prices mean optimism. Euphoria and full commitment to the market means topping prices. Falling prices mean pessimism. Despair and capitulation mean bottoming prices. Prices are the first thing to respond to changes in mood. The fundamentals always take longer. If a negative mood influences an investor, he can reach for his telephone or his computer and sell shares. However, if the same mood influences the CEO of a multinational, he may make decisions accordingly, but it will take months or years for the effects those decisions to be reflected in fundamental data. This is another reason that the markets always lead the economy. In fact, I think the reason that the fundamentals do so badly in predicting major tops or bottoms in stock markets is because the inherent optimism or pessimism of the markets becomes embedded in the fundamentals. Then that optimism or pessimism becomes the basis for comparison. Thus comparing optimism with optimism at market highs leads to perceiving optimism as normality. Comparing pessimism with pessimism at market lows leads to perceiving pessimism as normality. Therefore, if this is right, at major tops we should in fact see analysts fully bullish, and at major bottoms, we should see analysts fully bearish. The data


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

actually bears this out. Investors Intelligence (www.investorsintelligence.com) have been polling investment advisors since 1963, and plot the results against the market. Advisors’ sentiment rises and falls with the market, but always lagging. Chart 1: www.investorsintelligence.com

% Bullish Advisors

Dow’s first principle in interpreting the market averages, and perhaps his way of explaining this phenomenon, was that the market averages discount everything. The value of the average at any point in time represents the sum total of all the knowledge, and of all the hopes, fears and expectations based thereon, of all market participants at that time - including those all important knowledgeable but selfish insiders. Dow’s next great observation was that the market averages move in trends. An up trend is in being when the market continues to make higher highs and higher lows. A down trend is in being when the market continues to make lower highs and lower lows. The trend should be assumed to be continuing until the contrary is signalled by the price. This observation is applicable to any chart of any instrument in any market, and should be the starting point of any attempt at technical analysis. Dow classified trends into three types: the primary trend, the secondary trend and the minor trend. In Dow’s time, the markets were unregulated, and as we have already seen, it was widely believed, and probably true, that large operators combined together to manipulate the market to their own advantage. Dow held that the primary trend was not capable of manipulation, and that the primary trend was in fact the most reliable leading indicator of the economy. The primary trend would last for a year at least, or possibly several years. The chart above shows that the market always begins to rise in anticipation of the end of a recession, bearing out Charles Dow’s canny observation 100 years later. It refutes completely any notion that the economy picks up, and then the stock market reacts to that. The secondary trend was a partial retracement of the primary trend which would occur from time to time, and would retrace between one third and two thirds of the preceding movement in the direction of the primary trend, and last several months. The minor trend was the fluctuations in both primary and secondary trends, which would last weeks.

DJIA

William Peter Hamilton’s 1922 book “The Stock Market Barometer”, is still in print. It was so called not, as one might think today, because he had found some profitable means of predicting the stock market, but to show that the stock market was a “barometer”, i.e. an advance warning signal for the economy. Hamilton demonstrates that by reviewing the trends in the economy and in the averages from 1900 to 1921. This includes, of course, the upheavals of the First World War. Dow held that the two market averages should confirm each other. Therefore one can re-state the earlier rules more fully. An up trend is in being when the market continues to make higher highs and higher lows and both averages confirm each others’ higher highs and lows. They may not do so at exactly the same time. A down trend is in being when the market continues to make lower highs and lower lows and both averages confirm each others’ lower highs and lows. They may not do so at exactly the same time. The trend should be assumed to be continuing until the contrary is signalled by the price and that is confirmed in both averages. When one average makes a new high or low and that is not confirmed by the other average, it is a warning that the continuation of the trend is not confirmed, and therefore a warning of possible trend change. In my opinion, this still holds good today. There have been two major non-confirmations in recent times which heralded major trend change, marked 1 and 2 in my chart. There is another one potentially forming marked 3, which requires the Dow Industrials to make a new all time high for a major non-confirmation to be avoided.

Trend change A change in the primary trend should not, however, be inferred until the final phase of a bull or bear market has been reached (see below), and until valuations are cheap, for a bear market low, or high, for a bull market top.

37


38

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Chart 2: DJIA 1950-2004 showing the ECRI leading indicator (lower line) and the recessions (grey stripes) Finally, distribution. During this phase, the public’s enthusiasm for stocks knows no bounds, but knowledgeable strong investors - Dow’s insiders - are selling, effectively offloading their stock. Speculation reaches a height, and there is the opposite of a “flight to quality”: an aversion to “boring” solid stocks, and an indiscriminate willingness to accept greater and greater risk. In the distribution phase, indicators will show bearish divergence. The same three phases apply in reverse to a primary bear market. First, distribution. This overlaps with the final phase of the bull market. As the market trend rolls over from bull to bear, volume begins to diminish on rallies, and rise on down moves.

Lines A change from primary trend to secondary reaction would often be signalled by the formation of a ‘line’, i.e. a sideways consolidation within a range of about 5% above and below a mean figure, followed by a breakout, which had to be confirmed by both averages. If the breakout was in the direction of the primary trend, that trend was continuing. If the breakout was in the opposite direction, a secondary reaction had started.

Next, panic. Prices may drop vertically, and volume mounts to climactic proportions. Finally, capitulation. In this last stage, even top quality stocks decline, and cause distress to those who have held through the panic stage, forcing sales. By the time all the bad news and pessimism is fully priced in, the market will bottom. It may do so on a day of climactic volume and huge negative breadth, ending, or followed very soon after, by a strong rebound. Before the bottom there may be several days of 90% plus negative breadth. During the capitulation phase, indicators will begin to show bullish divergence.

This is another observation of general usefulness - trading ranges are very common, and the direction of the breakout generally determines whether the trend is continuing or correcting.

Strength of Dow Theory

Volume

It is the ultimate fallback technique. If all else fails, look to see whether you have a pattern of rising highs and lows, or falling highs and lows. Check the volume to see whether it rises with the up moves or the down moves. Look for sideways consolidations and look for a breakout to establish the direction of next move.

Dow also noticed that volume tends to increase in the direction of the trend. Therefore in an up trend, the volume will increase as the price rises and diminish as the price falls. In a down trend, the volume will increase as the price falls and diminish as the price rises. In my view, since 2000 the volume pattern has continued to show falling volume under a rising market and rising volume under a falling market - which is not good news for the bulls.

Dow identified three phases to a primary bull market. First, accumulation. Prior to this phase, the general public has capitulated and is extremely bearish. The financial news is at its worst. Distressed sellers are in the market, but knowledgeable and strong buyers are buying, and are prepared to raise their bid to acquire stock. Next, public participation. During this phase, the fundamentals improve, the public begins to join in, and a strong broad rally results.

Limitations of Dow Theory There are two principal limitations. First, because you have to wait for an out of sequence high and low for a signal of trend change, you may miss the highest high or lowest low by a considerable amount. Dow Theory signals “late” but reliably. Second, there is no telling how long a new trend may last - it may only last one high and one low before the previous trend continues. As Edwards and Magee put it: “A reversal in trend can occur any time after that trend has been confirmed.” The answer to this conundrum lies in Elliott wave theory. Ralph Nelson Elliott studied the work of Robert Rhea on Dow Theory after the 1929 crash, and developed his wave theories on the foundations so ably laid by Charles Dow.


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Chart 3: Dow Industrials (scale right). Dow Transports (scale left). CQG

Dow Industrials (scale right)

Dow Transports (scale left)

Chart 4: DJIA and Total Volume

Psychology and Markets

Systematic Trading

39


40

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

The Dow Theory and other Sell Signals Peter Beuttell

Article originally featured in Market Technician 33 (October 1998)

Earlier this year I took up the offer via the IFTA to join the Market Technicians Association e-mail list. Whilst a good percentage of the mail is only of passing interest, there have also been some very worthwhile discussions. One which occurred in the first half of August was over whether a Dow Theory Sell signal had been given. A number of contributors to the list were uncertain. Having already consulted both the Murphy and Edwards & Magee books, I had come to the conclusion that one had occurred. It was therefore with some relief that I read Robert Colby’s posting, responding to a contributor’s earlier e-mail as follows:

In mid-July the Industrials made new all-time highs, but these were not matched by the Transports, and this non-confirmation was the first “early warning” that a Sell Signal might occur. On July 29th, the Transports closed at 3,244.93 putting in place the first half of the signal. On August 4th, the Dow closed at 8,487.31 completing it (Figure 1). This was the first Bear signal since the August 1982 Bull market indication, which netted a 7,647-point gain (910%). Figure 1: Dow Industrials & Transports

“It is not my OPINION that the Dow Theory gave a new bear market sell signal last week. It is a FACT, according to the clear and strict definitions published by Charles Dow and William Hamilton early in this century.” So that settled that. Since there was obvious confusion amongst our American colleagues, which left me concerned about my own interpretation, I therefore hope that members will find it useful to see a basic synopsis of the Theory, and see what other evidence there is for bear market conditions now existing in the US. The main points of Dow Theory are: • There are three degrees of trend, or wave: Primary, Secondary, Minor. • Primary trends last a year or more, with a 20% move being the criterion for a Bull or Bear market. • Secondary trends normally last from three to thirteen weeks. • Minor trends are ignored. •

To produce a change-of-trend signal, both indices must break the extreme of the previous Secondary reaction on a closing basis. They do not have to do it simultaneously. Murphy refers to “failure swings” but the basic theory does not require these for a signal.

Since 1897, there have been 23 Bull market signals and 24 Bear market signals. Of these signals, 21 Bull and 20 Bear were correct an impressive performance. Refer to the Encyclopedia of Technical Market Indicators by Colby and Meyers for full details of the Theory and its record. In practice, the recent Sell signal unfolded as follows. The May/ June correction in the Industrials lasted just over a month and so fell into the required Secondary category. Its low was 8,627.93. The Transports corrected for slightly longer and bottomed at 3,259.30.

There were numerous other bearish signals visible in late July. A 5-wave advance off the October 1997 low was completing, and there were the usual weekly and daily momentum, breadth and volume divergences appearing. However, one or two other ways of looking at the market were warning that at least an intermediate, if not major. top was forming. Time will tell which, but the signals suggested that what was coming was worth avoiding. Firstly, my Elliott wave count (oft revised on the way up) indicated that, if the 1990’s Bull market was wave (5) of the post-1974 pattern, in percentage terms wave (5) would be .618 of the net travel of (1) through (3) at 1,194.54 in the S&P 500. The intra-day high in July was 1,190.58 - close enough. The longest advance since 1960 was 3 years and 7 months, finishing in late 1965. The 1984-87 period was 3 years and a month. The 1990-94 advance was 3 years and 3 months. As the bull run from the December 1994 low entered July, it was the longest advance uninterrupted in both price and time terms for over 30 years. It had also risen at the same semi-log rate as the pre-Crash period. It was both long-in-the-tooth, and on an unsustainable trend (Figure 2).


Indicators and Momentum

Elliott Wave and Fibonacci

Figure 2: S&P 500

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

Figure 3: S&P 500 & Total US Cum. Aid Line

Whilst the NYSE Cumulative Advance/Decline Line was making new highs in Q1 1998, the Total US equivalent (NYSE + NASDAQ + AMEX) was not, and a divergence warning was building (Figure 3). When the S&P rallied in wave v in July, the Total US CADL was already collapsing, and the NYSE CADL finally produced a divergence warning.

Peter Beuttell is a Director of MTS Research Ltd.

41


42

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Wyckoff Laws and Tests Dr. Hank Pruden (l) and Dr. Bernard Belletante (r)

Article originally featured in Market Technician 51 (November 2004)

Wyckoff is a name gaining celebrity status in the world of technical analysis and trading. Richard D.Wyckoff, the man, worked in New York City during a “golden age” for technical analysis that existed during the early decades of the 20th Century. Wyckoff was a contemporary of Edwin Lefevré who wrote The Reminiscences of A Stock Operator. Like Lefevré, Wyckoff was a keen observer and reporter who codified the best practices of the celebrated stock and commodity operators of that era. The results of Richard Wyckoff’s effort became known as the Wyckoff Method of Technical Analysis and Stock Speculation. Wyckoff is a practical, straight forward bar chart and point-and-figure chart pattern recognition method that, since the founding of the Wyckoff and Associates educational enterprise in the early 1930’s, has stood the test of time. Around 1990, after ten years of trial-and-error with a variety of technical analysis systems and approaches, the Wyckoff Method became the mainstay of The Graduate Certificate in Technical Market Analysis at Golden Gate University in San Francisco, California, U.S.A. During the past decade dozens of Golden Gate graduates have gone on to successfully apply the Wyckoff Method to futures, equities, fixed income and foreign exchange markets using a range of time frames. In 2002 Mr. David Penn, in a Technical Analysis of Stocks and Commodities magazine article named Richard D.Wyckoff one of the five “Titans of Technical Analysis.” The Wyckoff Method has withstood the test of time. Nonetheless, this article proposes to subject the Wyckoff Method to the further challenge of real-time-test under the natural laboratory conditions of the current U.S. Stock market. To set up this “test,” three fundamental laws of the Wyckoff Method will be defined and applied.

Three Wyckoff Laws The Wyckoff Method is a school of thought in technical Market analysis that necessitates judgment. Although the Wyckoff Method is not a mechanical system per se, nevertheless high reward/low risk opportunities can be routinely and systematically based on what Wyckoff identified as three fundamental laws (see Table 1): Table 1: Wyckoff three fundamental laws

1. The Law of Supply and Demand states that when demand is greater than supply, prices will rise, and when supply is greater than demand, prices will fall. Here the analyst studies the relationship between supply vs. demand using price and volume over time as found on a bar chart. 2. The Law of Effort vs. Results divergencies and disharmonies between volume and price often presage a change in the direction of the price trend. The Wyckoff “Optimism vs. Pessimism” index is an on-balanced-volume type indicator helpful for identifying accumulation vs. distribution and gauging effort. 3. The Law of Cause and Effect postulates that in order to have an effect, you must first have a cause, and that effect will be in proportion to the cause. This law’s operation can be seen working as the force of accumulation or distribution within a trading range works itself out in the subsequent move out of that trading range. Point and figure chart counts can be used to measure this cause and project the extent of its effect.

Past: position of the U.S. stock market in 2003 - Bullish Charts 1 and 2 show the application of the Three Wyckoff Laws to U.S. Stocks during 2002-2003. Chart 1, a bar chart, shows the decline in price during 2001-02, an inverse head and shoulders base formed during 2002-2003 and the start of a new bull market during March-June 2003. The upward trend reversal defined by the Law of Supply vs. Demand, exhibited in the lower part of the chart, was presaged by the positive divergencies signalled by the Optimism Pessimism (on-balanced-volume) Index. These expressions of positive divergence in late 2002 and early 2003 showed the Law of Effort (volume) versus Result (price) in action. Those divergences reveal an exhaustion in supply and the rising dominance of demand or accumulation. The bullish price trend during 2003 was confirmed by the steeply rising OBV index; accumulation during the trading range continued upward as the price rose in 2003. Together the Laws of Supply and Demand and Effort vs. Result revealed a powerful bull market underway.


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Chart 1: Wyckoff Laws: Law of Effort vs Result. Law of Supply and Demand.

Psychology and Markets

Systematic Trading

help define the onset of a bear trend out of top formation following a significant advance.) These nine classic tests of Wyckoff are logical, time tested, and reliable. As the reader approaches this case of “Nine Classic Buying Tests,” he/she ought to keep in mind the following admonitions from the Reminiscences of a Stock Operator (See Appendix): Table 2 Wyckoff Buying Tests: Nine Classic Tests for Accumulation Nine Buying Tests (applied to an average or a stock after a decline)*

Chart 2: Law of Cause and Effect

Indication:

Determined From:

1) Downside price objective accomplished

Figure Chart

2) Preliminary support, selling climax, secondary test

Vertical and Figure

3) Activity bullish (volume increases on rallies and decreases on reactions)

Vertical

4) Downward stride broken (i.e., supply line penetrated)

Vertical or Figure

5) Higher supports (daily low)

Vertical or Figure

6) Higher tops (daily high prices rising)

Vertical or Figure

7) Stock stronger than the market (i.e., stock more responsive on rallies and more resistant to reactions than the market index)

Vertical Chart

8) Base forming (horizontal price line)

Figure Chart

9) Estimated upside profit potential is at least three times the loss if protective stop is hit

Figure Chart for Profit Objective

* Adapted with modifications from Jack K. Hutson, Editor, Charting the Market: The Wyckoff Method (Technical Analysis, Inc., Seattle, Washington, 1986), page 87.

“The average ticker hound - or, as they used to call him, tapeworm - goes wrong, I suspect, as much from overspecialization as from anything else. It means a highly expensive inelasticity. After all, the game of speculation isn’t all mathematics or set rules, however rigid the main laws may be. Even in my tape reading something enters that is more than mere arithmetic. There is what I call the behavior of a stock, actions that enable you to judge whether or not it is going to proceed in accordance with the precedents that your observation has noted. If a stock doesn’t act right don’t touch it; because, being unable to tell precisely what is wrong, you cannot tell which way it is going. No diagnosis, no prognosis. No prognosis, no profit.

The “Nine Classic Buying Tests” of the Wyckoff method The classic set of “Nine Classic Buying Tests” (and “Nine Selling Tests”) was designed to diagnose significant reversal formations: the “Nine Classic Buying Tests” define the emergence of a new bull trend (See Table 2). A new bull trend emerges out of a base that forms after a significant price decline. (The “Nine Selling Tests”

“This experience has been the experience of so many traders so many times that I can give this rule: In a narrow market, when prices are not getting anywhere to speak of but move within a narrow range, there is no sense in trying to anticipate what the next big movement is going to be - up or down. The thing to do is to watch the market, read the tape to determine the limits of the getnowhere prices, and make up your mind that you will not take an interest until the price breaks through the limit in either direction. A speculator must concern himself with making money out of the market and not with insisting that the tape must agree with him.

43


44

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

“Therefore, the thing to determine is the speculative line of least resistance at the moment of trading; and what he should wait for is the moment when that line defines itself, because that is his signal to get busy.” Point 4 on the charts identifies the juncture when all Nine Wyckoff Buying Tests were passed. The passage of all nine tests confirmed that an uptrending or markup phase had begun. The passage of all Nine Buying Tests determined that the speculative line of least resistance was to the upside.

Future: A market test in 2004 The authors as academics are intrigued by the natural laboratory conditions of the stock market. A prediction study is the sine quo non of a good laboratory experiment. The Wyckoff Law of Cause and Effect seemed to us to provide an unusually fine instrument of conducting such an experiment, a “forward test.” Parenthetically, it has been our feeling, shared by academics in general, that technicians have focused too heavily upon “backtesting” and not sufficiently upon real experimentation. The time series and metric nature of the market data allow for “forward testing.” Forward testing necessitates prediction, followed by the empirical test of the prediction with market data that tell what actually happened. How far will this bull market rise? Wyckoff used the Law of Cause and Effect and the point-and-figure chart to answer the question of “how far.” Using the Inverse Head-and-Shoulders formation as the base of accumulation from which to take a measurement, of the “cause” built during the accumulation phase, the point-andfigure chart (Chart 2) indicates 72 boxes between the right inverseshoulder and the left inverse-shoulder. Each box has a value of 100 Dow points. Hence, the point-and-figure chart reveals a base of accumulation for a potential rise of 7,200 points. When added to the low of 7,200 the price projects upward to 14,400. Hence, the expectation is for the Dow Industrials to continue to rise to 14,400 before the onset of distribution and the commencement of the next bear market. If the Dow during 2004-2005 comes within + or 10% of the projected 7,200 points we will accept the prediction as having been positive.

Conclusions In summary, U.S. equities are in a bull market with a potential to rise to Dow Jones 14,400. The anticipation is for the continuance of this powerful bull market in the Dow Industrial Average of the U.S.A. through 2004. This market forecast is the “test” to which the Wyckoff Method of Technical Analysis is being subjected. Part (B) of “Wyckoff Laws: A Market Test” will be a report in year 2005 about “What Actually Happened.” As with classical laboratory experiments, the results will be recorded, interpreted and appraised. This sequel will invite a critical appraisal of the Wyckoff Laws and in particular a critical appraisal of the Wyckoff Law of Effort vs. Result. The quality of the author’s application of the Wyckoff Laws will also undergo a critique. From these investigations and appraisals, we shall strive to extract lessons for the improvement of technical market analyses. Irrespective of the outcomes of this market test, we are confident that the appreciation of the Wyckoff Method of Technical Market Analysis will advance and that the stature of Mr. Richard D.Wyckoff will not diminish. Dr. Hank Pruden is a professor in the School of Business at Golden Gate University in San Francisco, California and a visiting professor at Euromed-Marseille Ecole de Management, Marseille, France. Dr. Bernard Belletante is a Professor of Finance and Dean of the Euromed-Marseille Ecole de Management.

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

References Forte, Jim, CMT, “Anatomy of a Trading Range,” Market Technicians Association Journal,” Summer-Fall 1994. Hutson, Jack K., Editor, Charting The Market: The Wyckoff Method, Technical Analysis, Inc., 1986. Lefervé, Edwin, Reminiscences of a Stock Operator, Wiley Press (original, Doran & Co, 1923). Penn, David, “The Titans of Technical Analysis,” Technical Analysis of Stock & Commodities, October, 2002. Pruden, Henry (Hank) O.,“Wyckoff Tests: Nine Classic Tests For Accumulation; Nine New Tests for Re-accumulation,” Market Technicians Association Journal, Spring- Summer 2001. Pruden, Henry (Hank) O.,“A Test of Wyckoff,” The Technical Analyst, February 2004. Charts, courtesy of Wyckoff/Stock Market Institute, 13601 N. 19th Avenue 1, Phoenix, Arizona, U.S.A. 85029-1672.


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

Unravelling the DNA of the Market: Applying the Double Helix Framework to Wyckoff and Elliott Henry O. “Hank” Pruden

Article originally featured in Market Technician 82 (March 2017)

Preface • • •

DNA was first isolated by the Swiss physician Friedrich Miescher in 1869. In 1953, James Watson and Francis Crick suggested what is now accepted as the first correct double helix model of DNA structure. A metaphor is a figure of speech that describes a subject by asserting that it is, on some point of comparison, the same as another otherwise unrelated object. (Wikipedia)

So DNA, which is in itself a kind of metaphor; is one more, and perhaps the ultimate, way to consider how markets possess a kind of life of their own. This is useful in encouraging the analyst to identify ever more basic structural components, how they interact, and ultimately to predict outcomes... Robert Miltner, Scientist, Chemist and Entrepreneur, Larkspur, California

Introduction The following series of visuals were inspired by the theme of the IFTA 2014 Conference in London: “Unravelling the DNA of the Market.” I found the topic particularly appealing because for years in both active trading for my own account or in teaching classes at Golden Gate University, I had found synergy in combining the Wyckoff Method with the Elliott Wave Principle. The two approaches working together created something that was greater than the sum of their two respective parts. I believe that Wyckoff and Elliott represent ever more basic structural components of the market. I further believe that the double-helix framework of DNA is a very useful metaphor for combining Wyckoff and Elliott for better, more profitable market timing decisions.

45


46

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Wyckoff is a straight-forward price and volume method for analysing the present technical position and probable future trend of price behaviour in stocks, bonds and commodities. Figure 1: The Double Helix Framework

Figure 1 is an abstract of the double helix structure of DNA. This shall be used metaphorically as the market structure that combines or binds together the analytical components of the Wyckoff Method of Market Analysis with the Elliott Wave Principle of Market Analysis.

The Wyckoff Method Figure 2: The Wyckoff Method Strand

Figure 2, The Wyckoff Method Strand, is defined in Table 1.


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

Table 1: Distinctive Characteristics of the Wyckoff Method

• Wyckoff is a straight-forward price and volume method for analysing the present technical position and probable future trend of price behaviour in stocks, bonds and commodities. The method is a collection of the best practices and concrete experiences of the old time pool operators observed and recorded by Mr Richard D. Wyckoff. He gave primary emphasis to price and volume behaviour reflected on the ticker tape and shown on charts. Mass behaviour (the public) was generally on the other side of the trades from the “smart money” operators. Wyckoff condensed the “smart money” into a construct he named the Composite Man.

• The Wyckoff Method is a judgmental approach to interpreting the behaviour of the market. Mr. Wyckoff and his associates condensed the patterns of market behaviour they observed into three laws, nine tests and several schematics, plus additional principles and procedures.

• It was a “bottom up” approach based upon the best practices of actual traders and not a top down set of hypotheses deduced from a grand theory.

Figure 3: Schematic of the Wyckoff Cycle

Wyckoff Theory

Distribution Markdown

Re-accumulation

Re-distribution

Markup

Accumulation

Figure 3 is a schematic of the Wyckoff Cycle. This is a drawing of the price action depicting the key Wyckoff stages of Accumulation, Markup, Distribution, and Markdown.

47


48

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Figure 4A: Illustration of Wyckoff Applied

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Figure 4B: Illustration of Wyckoff Applied

Source: September 1998, Technical Analysis of Stocks & Commodities, pg.77

Source: September 1998, Technical Analysis of Stocks & Commodities, pg.77

Figure 4A, is an idealised illustration of the Wyckoff Method applied to the stock market behaviour using the vertical or bar chart.

Figure 4B, continues the idealised illustration of Wyckoff applied using a figure or point and figure chart.

Figure 5: The Elliott Wave Principle Strand

Figure 5: The Elliott Wave Principle Strand, is defined Wikipedia, the free encyclopaedia, as follows: The Elliott Wave Principle is a form of technical analysis that some traders use to analyse financial market cycles and forecast market trends by identifying extremes in investor psychology, highs and lows in prices, and other collective factors. Ralph Nelson Elliott (1871–1948), a professional accountant, discovered the underlying social principles and developed the analytical tools in the 1930s. He proposed that market prices unfold in specific patterns, which practitioners today call Elliott waves, or simply waves. Elliott published his theory of market behaviour in the book 'The Wave Principle' in 1938, summarised it in a series of articles in Financial World magazine in 1939, and covered it most comprehensively in his final major work, ‘Nature’s Laws: The Secret of the Universe’, in 1946. Elliott stated that: “because man is subject to rhythmical procedure, calculations having to do with his activities can be projected far into the future with a justification and certainty heretofore unattainable.”


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

Figure 6: Schematics of the Elliott Wave Principle

Figure 6 is an assembly of Elliott Wave Principle cycles in three different degrees of refinement, thus wave 1 in the first level, top schematic that is the first of five-waves found in a bull market. Wave 1 in turn is composed by another five smaller wave bull movement, illustrated immediately below it. The third level schematic is in turn sub divisible into 21 sub waves that reflect the five wave bull movement of the immediate higher degree.

Wyckoff and Elliott: Partners in command •

Partners in Command (New York, The Penguin Press, 2007) was written by Mark Perry to review the remarkable relationship forged between U.S. Army Generals George Marshall and Dwight Eisenhower. That partnership in command helped lead the Allied Forces to victory during WW II. In this acclaimed book: “Perry shows that Marshall and Eisenhower were remarkably close colleagues who brilliantly combined strengths and offset each other’s weaknesses in their strategic planning, on the battlefields, and in their mutual struggle to overcome bungling, political sniping and careerism of both British and American Commanders that infected nearly every battle and campaign”[ I]. Marshall and Eisenhower were titans in war and peace.

• In a loosely parallel fashion, the teachings of Richard D. Wyckoff and Ralph N. Elliott can be brought closer together to benefit the analyst-trader. Wyckoff and Elliott can combine strengths and offset each other’s weaknesses. As David Penn had written in the Technical Analysis of Stock and Commodities magazine [2], both Wyckoff and Elliott were titans of Technical Market Analysis. Then in a more recent TSAA Review article [3], I wrote about the ways Wyckoff and Elliott were sufficiently independent, yet complementary. They are powers. When used together; Wyckoff plus Elliott generate synergy or the famous 2+2=5 formula.

49


Theory of Technical Analysis

50

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Applying Wyckoff Plus Elliott A trade in the DJIA Illustrates the Power of Wyckoff Plus Elliott •

Please see Figure 7 for a Wyckoff Analysis and Figure 8 for an Elliott Wave Analysis of the 12th June 2008 DJIA. The analyses of these charts presume that the reader has a reasonable familiarity with the rudiments of both the Wyckoff Method and the Elliott Wave Principle to follow the interpretations presented below.

In Figure 7, the one-minute bar chart of the DJIA shows a classic Wyckoff sign of weakness breakdown and a pullback rally to a last-point of supply set-up around 12:45-1:00pm at DJIA 12,225 on 12th June 2008. A put or a short ETF position could have been entered. The DJIA then systematically and steadily worked its way downward until about 3:10pm. That steady decline ended with a vertical plunge to the level of prior support at 12,074. That plunge appeared climactic and also created an oversold condition by overshooting the supporting parallel line of the down channel. The DJIA entered a Wyckoff oversold condition that made it vulnerable to a rally. A bear-trader would have been alerted to exit for the day. But, the real clincher for exiting was given by Elliott on the next and final rally of the day.

Figure 7: Wyckoff Analysis

Legend for interpreting the Wyckoff principle appearing in Figure 7 • SOW: Sign of Weakness, which will usually occur on increased spread and volume, as compared to the preceding rally. Supply is showing dominance. Fall through the Ice or breaking of support. • LPSY: Last Point of Supply: After a SOW, a feeble rally attempt on narrow spread shows us the difficulty the market is having in making a further rise. Volume may be light or heavy, showing weak demand or substantial supply. At LPSYs the last waves of distribution are being unloaded before markdown is to begin. LPSYs are good places to initiate a short position or to add to already profitable ones. • Climax = Selling Climax: The approaching exhaustion of supply or selling is evidenced in preliminary support (PS) and the selling climax (SC) where a widening spread often climaxes and where heavy volume or panicky selling by the public is being absorbed by larger professional interests. Once these intense selling pressures have been expressed, an automatic rally (AR) follows the selling climax. A successful secondary test on the downside shows less selling than on the SC and with a narrowing of spread and decreased volume. A successful secondary test (ST) should stop around the same price level as the selling climax. The lows of the SC and the ST and the high of the AR set the boundaries of the TR.


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

Figure 8: Elliott Analysis

In Figure 8, the Elliott Wave Principle revealed a clear five-stage C-wave down to the low at 12,074. Furthermore, the fifth wave itself revealed a five-wave pattern with a classic tiny triangle in the fourth wave. Elliott was flashing warning signs to get out. Finally, the Elliott pattern was reinforcing the forgoing Wyckoff interpretation. Together Wyckoff and Elliott were saying “get out” to the trader near the bottom of the day. The final rally of the day was a five-wave upward impulse wave that broke the downtrend line in Figure 7 while recovering 100% of the preceding down wave. This powerful bullish indication warned the trader that more strength would follow; this bullish impulse wave was warning the trader not to carry her short sale position overnight. In conclusion, Wyckoff and Elliott conducted a command performance for the astute trader on 12th June 2008. WE are partners in command!

Conclusion This article presents the technical analyst and technical trader with the metaphor of the double-helix framework for grasping a more profound look into a basic DNA structure of the stock market. The double helix structure can be used to combine the independent powers of the Wyckoff Method and Elliott Wave Principle. Together Wyckoff and Elliott forge a partnership that combine their strengths and offset each other’s weaknesses. That powerful synergy of Wyckoff and Elliott was illustrated with the case-study of intraday market action. That action was first explained with the Wyckoff Method, and then the Elliott Wave Principle. Together, Wyckoff and Elliott made a compelling case.

51


52

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

The warning signs of Major Market Tops Paul Desmond

Article originally featured in Market Technician 79 (September 2015)

Introduction The Law of Supply and Demand is universally recognised in both academia and the business world as the foundation, the starting point of all economic analysis. The Law of Supply and Demand states that when demand for a freely traded commodity exceeds the supply of that commodity, its price will rise. And, if the supply of that commodity exceeds the demand for it, the commodity’s price will fall. Notice the lack of equivocation. No mights or coulds or shoulds, just will. It is the Law. And, since common stocks are a freely traded commodity, their price movements are dictated by the Law of Supply and Demand - the starting point of common stock analysis (and, therefore, stock market analysis). As such, it is difficult to imagine why it is not at the core of every investor’s portfolio strategy.

Background analysis of market tops Since 1938, Lowry Research Corporation has measured the factual, unbiased forces of supply and demand at work on the New York Stock Exchange. Our database of those measurements now extends back 87 years - from mid-1925 to present. Recently, we have expanded our analysis to encompass 24 major stock exchanges around the world. In essence, we measure the daily movements of money into and out of the stock markets. These movements reveal changes in investor psychology over time, moving through repetitive cycles of hope, fear and greed - which are the direct result of the purchases and sales of billions of shares of stock by millions of investors. Since the actions of individual investors are usually heavily influenced by group psychology and economic cycles, security prices generally move in relatively well-defined trends, commonly known as bull markets and bear markets. Gradual changes in both buying enthusiasm and in the desire to sell provide a series of progressive warning signs - signs that can help investors anticipate important trend changes, shifting to a more aggressive strategy near the start of bull markets, and shifting to a more defensive strategy near the start of bear markets - markedly enhancing their longer term portfolio performance. Most investors will concede that the most difficult part of managing a portfolio of stocks is identifying the formation of a major market top before it is too late. This is undoubtedly due to the universal enthusiasm for stocks, and generally positive economic news that usually dominates investor psychology at such times. But, the warning signs are nevertheless present for those willing and able to see them. The period leading up to a major market top shares a number of similarities with the autumn season as it makes the transition into winter. That is, in autumn the leaves begin to fall from the trees in a very gradual process - nearly imperceptibly - one at a time, until the trees are eventually bare at the onset of winter. It is no different near major market tops. Individual stocks begin to roll over into their own bear markets, one at a time, usually beginning with the less noticeable small-cap and midcap stocks. An important consideration in this gradual process of erosion is that small-cap stocks generally make up about 40% to 50% of the stocks traded on most major global equity markets, while mid-

caps typically make up about 30% to 40%. Big-caps generally account for only about 10% to 15% of common stocks traded on most of the large world markets. In the emerging equity markets, the percentage of small-cap stocks is even more dominant. Thus, as a mature bull market rallies through a series of higher highs in the big-cap price indices (such as the DJIA and S&P 500), investors must be able to see that a growing majority of stocks may already be in downtrends. Otherwise, at the final high in the big-cap indices, a relatively small number of heavily weighted bigcap stocks can deceive investors into believing that the broad market is still in a healthy uptrend. The most important warning signs commonly found near bull market tops involve evidence of increasingly extreme selectivity. Just as there are a wide variety of ways to recognise the changing conditions from autumn to winter, there are a number of ways to observe the gradual process of a bull market devolving into a bear market, all of them involving the Lowry measurements of supply and demand. A detailed review of each of these indicators could easily occupy a lengthy dissertation. Instead, this paper offers an overview of the importance of these measurements in terms of avoiding the ravages of recurrent bear markets throughout the 87 year history of the Lowry Analysis.

1. New 52-Week Highs Usually, the first warning sign that a bull market is losing some of its upward momentum occurs when the percentage of stocks rising to New 52-Week Highs begins to contract. This warning typically occurs as much as six to twelve months in advance of the final highs in the big-cap indices. It is not so much a sign of trend weakness as it is a sign of fading strength - somewhat like an athlete who discovers she cannot jump as high, or run as fast, as she could in her youth. The contraction in New Highs serves as a gentle reminder that all bull markets eventually come to an end. It also shows that fewer stocks are participating in the bull market, prompting investors to begin reviewing their portfolios more closely, looking for holdings that have stopped making new highs. If a stock again fails to make a New 52-week High during subsequent market rallies, its lack of strength may indicate it is time to cull that stock and reinvest the proceeds in a stronger stock that is consistently rising to new highs. Later, as the


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

Chart 1: Example of 52-week highs peaking ahead of the market. Canada 2010-11

Table 1: Examination of Trading at Fourteen Peaks in the Dow Jones Industrial Average Bull Market Top Day

% Stocks @ New Highs

% at or <2% of New Highs

% off 20% or more

% off 30% or more

09/03/1929

2.30%

15.62%

31.84%

18.77%

03/10/1937

6.05%

21.34%

5.94%

1.06%

05/29/1946

8.59%

30.44%

6.30%

0.86%

04/06/1956

5.32%

23.36%

1.92%

0.42%

01/05/1960

1.60%

5.83%

23.25%

7.67%

12/13/1961

3.56%

11.83%

25.29%

11.60%

02/09/1966

9.66%

19.04%

9.52%

2.68%

12/03/1968

9.43%

20.12%

9.51%

2.36%

01/11/1973

5.30%

11.82%

34.22%

20.51%

09/21/1976

10.97%

22.88%

21.65%

10.09%

04/27/1981

7.09%

15.18%

28.01%

9.39%

08/25/1987

6.23%

15.23%

17.37%

7.44%

07/16/1990

5.35%

18.11%

37.31%

22.74%

01/14/2000

3.54%

6.31%

55.33%

32.45%

10/09/2007

10.77%

11.03%

26.51%

16.51%

Average

6.38%

16.54%

22.26%

10.97%

percentage of stocks rising to New 52-week highs continues to contract (reflecting an increasingly selective market advance), the strategy should shift to selling laggards one at a time, and building cash reserves rather than reinvesting in a weakening trend. On a broader basis, each time the big-cap indices rise to new bull market highs, it is important to keep track of the percentage of individual stocks also rising to new bull market highs. Our original 2006 study titled An Exploration of the Nature of Bull Market Tops (updated to include the 2007 bull market peak) examined every bull market top since 1929 and found that, on the final top day of the Dow Jones Industrial

53


54

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Average, less than 11% of common stocks listed for trading on the New York Stock Exchange were also making new highs. In the 1929 case, when the DJIA made its final high on September 3rd, the percentage of NYSE-listed stocks at new highs was just 2.3% evidence of extreme selectivity.

2. Stocks that are down by 20 per cent or more As a corollary to monitoring new highs, the Lowry Analysis also tracks the percentage of stocks that are already down from their highs by 20% or more on the same day that the Dow Jones Industrial Average is making its final bull market high. A decline of 20% or more is generally viewed as signifying a bear market. As Table 1 oppposite shows, if an investor had the unique ability to sell her portfolio on the exact final top day of the DJIA, she would discover that a substantial portion of her portfolio had already suffered significant losses. Just as the leaves in autumn fall from the trees one at a time, stocks drop out of bull markets in the same slow, gradual manner.

3. Advance-Decline Lines The gradual process of individual stocks dropping out of the ageing bull market - like the leaves on a tree as winter approaches - can be monitored each day through various Advance-Decline lines. Usually, the gradual reduction in the number of stocks advancing versus those declining on days of rallies first appears in the small-cap Advance-Decline line. Investors often turn away from small-cap stocks first, as they are often viewed as “one-trick ponies” with less financial strength, and generally more vulnerable to protracted market declines. The weakening of the smallcap Advance-Decline Line indicates that fewer small-cap stocks are

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

still participating in the remaining bull market, thus encouraging investors to reduce exposure to small-caps on a stock by stock basis. Progressive weakness soon begins to spread - either simultaneously or a short time later - to the mid-cap stocks, as reflected in persistent weakness in the Advance-Decline line for the mid-cap components. Soon after (usually about four to six months before the final bull market high in the big-cap indices), the spreading weakness can be observed in a steady downturn in our Operating-Companies-Only (OCO) Advance Decline Line1, despite new highs in the big-cap indices. At this point, new highs in the major big-cap price indices are not unlike a sand castle with its foundations gradually washing away, leaving the turrets progressively more vulnerable. It should be emphasised that the weakness in the various segmented Advance-Decline lines does not ever call for a single all-encompassing sell signal, but simply encourages investors to reduce risk exposure on a case-bycase basis, as more and more individual stocks demonstrate they are rolling over into their own bear markets.

The New York Stock Exchange accepts issues for trading that are not common stocks. Closed-end bond funds and Preferred shares (which are not convertible into common shares) trade like bonds (since they are bonds) and thus can distort the patterns of the common stocks. Whilst the patterns of foreign stocks and ADRs are closely related to their home markets and can therefore distort the patterns of U.S. common stocks when the trends of their home markets do not match the trends in the U.S. As equity investors have in the past been badly misled by these distortions, Lowry have created a series of statistics that exclude closed-end bond funds, preferred stocks, and foreign issues, leaving us with “clean” data including just domestic common stocks. We describe these as Operating-Companies-Only. I

Chart 2: The phased decline of large, mid-cap and small cap stocks. Germany 2007


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

Chart 3: Market divergence between different indices. NYSE 2007

4. Market Index Divergences During the early stages of a bull market, virtually all equity indices tend to rise together, albeit at somewhat different rates. During the latter stages of a bull market, some price indices stop rising and begin to turn down as the remaining bull market becomes increasingly selective. Since most investor publications and websites tend to primarily display charts of the popular big-cap indices, it could be relatively easy for an investor to completely miss these important warning signs of major trend weakness. One of the earliest studies of the importance of price index divergences was the still highly regarded Dow Theory that addressed divergences between the Dow Jones Industrial Average and the Dow Jones Transportation Average near major market tops. Today, that original concept should be expanded to contrast the uptrends in the big-cap DJIA and S&P 500 Index to downtrends in a variety of sector or segment price indices - such as the S&P 100, 400 and 600 Indices, the Lowry Unweighted S&P 500 Components Index, the NYSE Composite, the Russell 2000 Index, or the Value Line Composite Index - during ageing bull markets. Each time another of the less popular price indices fails to confirm the continued strength in the big-cap indices, the warning signs become increasingly important. And, as the market advance becomes more selective, investors’ portfolios should become more selective. In this way, when the final highs of the bull market are registered, investors should be still holding only a small number of the strongest stocks.

5. Percentage of Stocks Above their 30-Week Moving Averages Another helpful way to be alerted to the increasing selectivity

typical of old bull markets is to observe regularly the percentage of NYSE-listed stocks above their 30-week moving averages each time the DJIA and S&P 500 Index rise to new bull market highs. During healthy bull markets, new highs in the popular price indices should be accompanied by 75% to as much as 90% of stocks above their 30-week moving averages. At the final bull market highs for the DJIA and S&P 500 Index, the number of stocks still in uptrend patterns has typically been dropping steadily to 60% or less. Remember, there is strength in numbers. The importance of the warning signs given off by any one indicator is multiplied when similar warning signs emerge in a variety of indicators, each of which measures the forces of supply and demand from a slightly different angle.

6. Buying Power vs. Selling Pressure: The Lowry analysis has become particularly well-known over the past 75 years for its composite measurements of supply (the Selling Pressure Index) and demand (the Buying Power Index). During the early stages of a healthy bull market, Buying Power typically rises sharply, reflecting expanding investor buying enthusiasm. At the same time, Selling Pressure usually drops steadily, showing that the supply of stocks being offered for sale is shrinking. However, during the latter stages of an old and increasingly fragile bull market, Buying Power commonly weakens as buyers become more cautious, while Selling Pressure rises steadily as initial profit-taking evolves into consistent distribution. A major reassessment of equity strategy is generally called for whenever the internal condition of the stock market deteriorates enough to cause Selling Pressure to rise to the dominant position

55


56

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Chart 4: Percentage of NYSE issues above their 30-week moving average. NYSE 2007

Moving Averages and Trends

during the months prior to a major market top. Therefore, for the purposes of this analysis, we will focus on the patterns of the NYSE and NASDAQ markets. One of the earliest initial warning signs of major market tops is a contraction in the number of stocks rising to New 52-week Highs. That is, as the major benchmark price index makes a series of new highs, the percentage of stocks also rising to new highs becomes progressively smaller.

Chart 5: Examples of the Selling and Buying Pressure Indices. NYSE 2007-8

above Buying Power, which represents the critical point when supply exceeds demand. There are many experienced investors who will argue strongly that there are no effective warning signs of major market tops. To those investors, bear markets just suddenly emerge without warning (somewhat like a sudden plague), and must simply be endured by investors until, hopefully, a new bull market eventually makes up their losses. The reason so many scholars and professional money managers were unable to see the warning signs reviewed in these pages is simply because their attention has been focused

exclusively on corporate earnings, and other macro-economic factors, rather than on the observance of the Law of Supply and Demand applied to the flows of money into and out of common stocks - the foundation, the starting point of all economic analysis and all stock market analysis.

Is the current bull market approaching a top? During 2015, we have been seeing a number of the early warning signs in the U.S. markets that typically emerge

This warning sign typically emerges during the last six to twelve months of an old bull market, but has persisted even longer in particularly extended bull markets. In the present case, the percentage of U.S. stocks rising to new 52-week highs has been steadily contracting, from 24.0% in October 2013 to just 4.9% on August 10, 2015. During that time, as Chart 6 above shows, each new high in the S&P 500 Index has been accompanied by a smaller and smaller number of stocks listed on the NYSE Exchange, rising to new 52-week highs. The important point here is, history shows that persistent, multi-month periods of a shrinking number of stocks making new 52-week highs can only be found during the final stages of old bull markets. Thus, as the original study noted, the persistent... “contraction in new highs serves as a gentle reminder that all bull markets eventually come to an end”. As a bull market continues to age, the number of stocks simply participating in the bull market (but not necessarily making new highs) begins to weaken. Chart 7 shows the percentage of the Lowry Operating-Companies-Only (OCO) stocks that are either at New 52- week Highs or are within 2% of New Highs. The idea behind this chart is to see if perhaps a large number of stocks might be lingering just below the surface, and ready to move to New Highs with just a few days of rally. However, the chart opposite shows that there were 35.14% of OCO stocks at, or within 2% of, New Highs as of Dec 29, 2014. But, that percentage has dropped to just 15.03% as of August 10, 2015, showing that more and more stocks have been dropping out of this old bull market in recent months. The lower indicator on chart 7 shows the percentage of OCO stocks that have already dropped by 20% or more from their 52-week highs. A 20% decline is generally considered to be the dividing line for identifying a bear market. Many investors take the position that the S&P 500 Index is an effective proxy for the broad market. Thus, as long as the S&P 500 Index is at, or near to, its bull market high, those investors will remain fully invested in stocks. However, this indicator clearly demonstrates that, in an old bull market with many stocks dropping out of


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Chart 6: NYSE New 52-week highs

Systematic Trading

buying interest fades, and those stocks eventually drop out of the bull market. At the top of Chart 8 is an Unweighted Index of the S&P 500 Components, which differs slightly from the standard weighted version. The unweighted index was actually at a new bull market high as of August 10, 2015. In the centre of the chart is Lowry’s unique Operating- CompaniesOnly (OCO) Advance-Decline Line. This indicator excludes all closed-end bond funds, preferred stocks, and ADRs from the roster of NYSE-listed stocks, leaving a “clean” list of domestic common stocks. The third indicator shown below is the NASDAQ Advance- Decline Line, which is dominated by small-cap and micro-cap stocks.

Chart 7: Percentage of stocks near their 52-week highs

As chart 8 shows, the OCO AdvanceDecline has been diverging from the Unweighted Index of S&P 500 Components for almost three months, since mid-May 2015. The divergence in the NASDAQ Advance-Decline Line is far more extensive due to substantial weakness in NASDAQ small-caps and micro-caps. It is important to recognise that, throughout the 88 year history of the Lowry Analysis, divergences between the S&P 500 Index and the Lowry OCO Advance-Decline Line have typically marked at least the last four to six months of bull markets. Investors hoping to minimise the effects of the next bear market on their portfolios must closely monitor the patterns of the Lowry OCO Advance-Decline Line in the period immediately ahead. Each time the S&P 500 Index increases to a new bull market high, while the OCO AdvanceDecline moves deeper into a pattern of lower tops, the final high in the S&P 500 Index will be just that much closer. The last chart in this analysis (Chart 9) provides just one more way of seeing the increasing selectivity of this old and fragile bull market. At the top of this weekly chart is the NYSE Composite Index.

the bull market one by one, the S&P 500 Index can be highly deceptive as a broad market indicator. In fact, it may represent only a very small number of heavily weighted big-cap stocks. In the current old bull market, as of August 10, 2015, 37.04% of our OCO universe of stocks have already lost 20%, or more, from their bull market highs - and thus, are already in a bear market. On a segmented basis, 48.16% of small-caps, 21.91% of mid-caps, and 16.36% of largecaps are already down by 20% or more from their bull market highs. Chart 8 shows two Advance-Decline Lines - simple but essential cumulative measurements of the number of stocks rising versus those falling each day. During the healthy stages of a bull market, the number of stocks being accumulated expands along with the major price indices. But, as stocks progressively rise to the point that they are viewed as over-valued,

The middle indicator shows the percentage of NYSE stocks that are still above their 10- week moving averages of price. Note in particular the sharp drop in the 10-week indicator between November 26, 2014, and August 6, 2015 - a period in which the NYSE Composite Index was moving essentially sideways. The most recent reading shows that only 42.0% of NYSE stocks are still above their 10-week moving averages - while 58.0% are below the 10- week moving average. The third indicator in chart 9 shows the percentage of NYSE-listed stocks that are still above their 30-week moving averages. It also clearly shows the increasing selectivity that began in earnest around mid-March, 2013, and has continued to its

57


58

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Chart 8: Advance-Decline lines

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

current reading of just 44.3% on August 6, 2015. Based on the various indicators of increasing selectivity shown throughout this article, it could be correctly argued that the bear market has already been underway for some time. The old approach of looking for a single, all encompassing sell-signal to announce the official start of a bear market simply is not realistic. Bear markets occur on a stock by stock basis and, if investors were to be able to sell all their stocks on the exact top day of the S&P 500, they would find that their broadly diversified portfolio had already been bleeding red ink for many months. While most proponents of Modern Portfolio Theory argue strongly that there are no warning signs occurring in advance of bear markets, the various indicators contained in this analysis speak loudly to the contrary.

Chart 9: Percentage of stocks above their 10- and 30-week moving averages

Paul F. Desmond is the President of Lowry Research Corporation. Founded by L. M. Lowry in 1938, it is the oldest continuously published advisory firm in the US. The company analyses global markets using its long history of various measurements of money flow.


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

59


60

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

CHAPTER THREE

Charting Types, Point & Figure and Candlesticks Articles in this chapter 62

Point and Figure Charts Jeremy du Plessis

64

An Introduction to Japanese Candle Charts Steve Nison

72

Japanese Candlestick Analysis Daniel Gramza

74

Measuring an Extended Market with Japanese Candlesticks Daniel Gramza

76

Trending with Heikin-Ashi Dan Valcu

78

Market Profile Peter Steidlmayer

Moving Averages and Trends


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

CHAPTER THREE INTRODUCTION

By Adam Sorab FSTA

Introduction After graduating in Economics, I started trading in the early ‘80s and quickly realised that my degree was fairly useless when it came to trading success in financial markets. However, I was lucky enough to be introduced to technical analysis charts and, despite having to draw them by hand in those days, I found they had significant advantages through the unique insight each different market visualisation provided. This Chapter is all about the different types of charts that technicians have invented to view and understand price action and many of the papers have been written by the leading proponents in their field. Jeremy Du Plessis is one of the leading Point & Figure experts in the world today. His paper, Point & Figure Charts, provides an introduction to the technique and highlights the importance of using intraday data whenever possible. It also hints at the synergies that can be generated by combining P&F with other methods. Steve Nison is regarded as being one of the first people to introduce the ancient skill of Japanese Candle Stick charts to the western world. His 1990 article, An Introduction to Japanese Candle Charts, was written nearly 400 years after they were first developed and it provides an excellent explanation of their principles and construction. The article goes on to highlight their power as both trend and reversal indicators; before culminating with a wide range of common candlestick patterns that remain as useful to traders today as they were for the 17th century Japanese rice traders who originally invested them. Dan Gramza is a professional trader who has spent decades sharing his skills through education and publication. In his first article, Japanese Candlestick Analysis, Dan revisits their candlestick chart construction to highlight the subtle but essential underlying changes in market psychology that candlesticks are so effective at illustrating. In his second article, Measuring an Extended Market with Japanese Candlesticks, Dan also highlights their qualities as reversal indicators; an uncommon feature with many other western approaches. Trending with Heikin Ashi by Swedish investor and educator, Dan Valcu, explores another Japanese chart type that is particularly renowned for its ability to remove noise and improve signal. In his article, Dan explains in detail how Heikin Ashi charts are created and also highlights a range of trading strategies that can be employed using this second generation candlestick charting approach. The final paper in this chapter, Market Profile, is by Peter Steidlmayer and I don’t mind admitting that Market Profile completely blew my mind when I was first

Psychology and Markets

Systematic Trading

introduced to it and the concepts behind it. As the son of a farmer, trading in Chicago’s commodity exchange, it’s perhaps not surprising that Market Profile is so different to other, more established, charting approaches. While the paper only touches on the subject in its two short pages, it nonetheless opens the reader’s eyes to this unique way of viewing markets and price action. Very importantly, it introduces the concept of a “point of market efficiency” and highlights how technicians can profit greatly from price stability, rather than having to rely on only trends. While these papers are all too short and only scratch the surface of their subject matter; they do demonstrate the very different ways in which technical analysts can and do look at markets. Like their authors, the different approaches vary in age and origin. However, they all share the same quality; that of improving our insight and chances of investment success within the financial markets in which we operate.

61


62

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Point and Figure Charts Jeremy du Plessis

Article originally featured in Market Technician 16 (February 1993)

The person who invented the Point and Figure chart was a genius. How they came to be invented is anyone’s guess, but it is not difficult to speculate on the reason.

Figure 1

The problem with all normal graphs (charts) is that one must have an evenly spaced X axis to display time, which is not practical unless you have a desk to lay out the chart and certainly no good to the turn-of-the-century floor-trader with only a small note pad to write on. So why not have a method of charting which takes into account price movement, but ignores time - quite a lateral thought. However, in order to fully see what is happening, the chartist must be able to see, not only the price level, but more important the direction of the trend. The Point and Figure (P&F) chart fulfils both these requirements and more. Point and Figure charts, as we all know, are constructed with X’s and O’s. The X’s indicate a rising trend and the O’s indicate a falling trend. The level of the X or 0 is the price level. The P&F inventor however, went one step further in devising his technique, in that he (or she) created the ability to incorporate a smoothing factor into the chart - a bit like a moving average. This, he achieved with a combination of the box size and the reversal. The box size is the value of each X and 0, the reversal is the number of X’s or O’s by which the price must reverse in order for a trend change to take place. Anyone who has plotted P&F charts by hand will know that no plot is necessary, unless a new box is filled or a reversal undertaken. So, the P&F chart filters out much of the market noise. The inventor therefore has given us a method of displaying price levels, tracking price movements, indicating the direction of the trend, smoothing out any trend-distorting fluctuations and finally, a chart which can be drawn without rulers or graph paper on the back of a cigarette box. And what is more, every single price change can be plotted, no matter how much time has passed between trades or how unequally spaced the time intervals between each trade are - what more could one ask for?

Figure 1 is a P&F of the FTSE-100 index. It’s a 5 x 3 P&F chart plotted from the hourly data which appears in the FT every day. This is probably the closest most Technical Analysts will come to a true P&F chart. It doesn’t reflect every price move, but it is better than a P&F chart based on the end of day close. Figure 2

It is this last point that so many people are not aware of. The age of computers and price download services has taken much away from the subtle art of Technical Analysis. At the end of the day, we take the last traded price and use it to plot our P&F charts, not realising what we are missing. P&F was designed for the trader who has access to and plots every single price change during the day. Of course it is not practical for most of us to plot every trade, so we have to make do with end of day prices. In his book on Point & Figure charts, A W Cohen attempted to overcome this inadequacy by proposing that the day’s high and low be incorporated into the Point and Figure chart as a proxy for plotting all the intraday movement. However, if we really want an accurate P&F chart, we should plot every trade. With the introduction of Financial Futures and the requirement that Stock Market indices are updated in real-time, the average trader can now take the published hourly changes in various indices and use these to obtain a virtually 100% P&F chart. If you are a short-term trader this is highly recommended.

Compare it with the chart in figure 2, which shows the same time period, but is a P&F chart plotted in the traditional way, based on the close at the end of the day. Notice how much trading detail is lost. Compare the horizontal count on the intraday P&F chart in figure 1 with the count in figure 2. The same period of data is being used, but the intraday chart gave a far more accurate target of 2725 from the base, whereas the close only chart gave a count of 2465.


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

If you are following the FTSE-100 closely for short-term trading, you really cannot rely on the close P&F to give you early signals. Notice how the intraday chart picks up a 45 degree channel which helped to identify support at 2730 and again at 2795. It would have been very difficult to have picked up a similar channel on the close only chart, although it is possible to draw a set of steep channel lines. The intraday P&F short-term vertical counts are more accurate. It predicted a top in the region of 2910 and then predicted the support at 2790 from the top. Figure 3

Psychology and Markets

Systematic Trading

particular levels, but this is not a true log scale. Log scale charts show percentage movements in price rather than absolute movements. A log scale P&F chart, therefore, has a constantly varying box size (value). As the price rises, so the chart becomes less sensitive to absolute price movements, but has the same sensitivity to percentage movements. The problem is that it becomes very difficult to determine the exact value of any box and so it is hard to conduct a vertical or horizontal count by hand. Figure 4 shows a Log scale chart of the FTSE-100. The box size is a percentage, in this case 1%. The reversal is still 3 boxes, but depending on the price level, 3 boxes equate to a different number of points. The chart shows the FTSE-100 on a daily basis (close only) since 1984. Notice the 45 degree bullish support line from the low has not been breached.

Point and Figure charts of Technical Indicators Most Technical Analysts use P&F charts for studying price movements, but they can be used for other indicators as well. Cohen showed that a number of Technical Indicators can be plotted using P&F charts - On-Balance Volume, Momentum, even Relative Strength.

The chart in figure 3 is drawn based on High & Low (the Cohen method). Whilst it is not an exact replica of the intraday chart in figure 1, it certainly bears a closer resemblance and does show a lot more detail than the close only chart in figure 2. The trend is easier to spot and the counts appear to be more accurate.

Figure 5 shows ICI relative to the FTSE-100 Index as a line chart and then as a P&F chart. Notice how well the P&F vertical counts apply to the RS chart. A downside count of 217 was predicted and achieved, as was a further downside of 195. The final upside of 206 has almost been achieved. Remember the X’s denote that ICI is performing better than the FTSE-100 index and O’s that it is performing worse. Space limitations prevent the inclusion of other examples, but the sky is the limit. Figure 5

If intraday figures are not published (like the FTSE-indices), then you should seriously consider a High/Low P&F chart - it may be showing you detail you would otherwise be missing on your close only chart.

Log Scale Point and Figure Most P&F charts are arithmetic, where each box value is the same no matter where you are on the chart. Some manually drawn P&F charts have their scale altered, so that the box value changes at Figure 4

Point and Figure charts are unique to Technical Analysis. It is therefore unlikely that you will encounter them in any other field, especially not in a presentation given by your local Fundamental Analyst (although line charts often are). This being the case, perhaps we need to spend more time studying and understanding them.

Cohen, A W, How to use the Three-Point Reversal Method of Point & Figure Stock Market Trading, Chartcraft Inc., NY,1968. Jeremy du Plessis is Managing Director of Indexia Research.

63


64

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

An Introduction to Japanese Candle Charts Steve Nison, CMT

Article originally featured in Market Technician 8 (July 1990)

Traditional methods of chart analysis use bar or point and figure charts. Yet over 100 years before these methods originated, the Japanese were using their own style of technical analysis in the rice market. This style evolved into the candlestick techniques currently used in Japan. In the 1600s the Japanese were trading forward contracts in rice - called “empty rice”, because they were trading rice that wasn’t there. Rice was not only Japan’s main staple, but it underpinned the economy. Samurai warriors’ salaries were rice stipends. These stipends were derived by a rice tax levied by the government. From the peasant on up, the price of rice was crucial. Although they were not the first to trade forward contracts (the Dutch were in the 1500s), the Japanese were the first to use technical analysis. It’s believed Japanese price charts originated around 1750. Some patterns used by the Japanese are similar to our own, but were discovered much earlier. For example, the equivalent of our head and shoulders formation was pre-dated by the Japanese three Buddha pattern. This was related to the Buddhist temples in which there is a large central Buddha with saints on both sides - a perfect analogy to a head and shoulders. With their extended historical foundation, Japanese candlestick patterns can offer us new insights. Their picturesque names also make them fun to use. Expressions such as “dark cloud covers,” “morning stars,” and “windows” abound. These techniques offer a wide spectrum of applications: • Candlestick charts are a useful stand alone tool. They can also be merged with other technical tools to create a synergy of techniques. • Knowing how the Japanese analyze markets provides valuable information given the extent of their participation in the U.S. financial markets. •

Certain Japanese candlestick combinations may imply a period of consolidation (therefore a decline in volatility), others hint of a forceful price move (thus a rise in volatility). They give deeper insight into market conditions that could provide benefits to option traders.

The Japanese method of plotting is called candlesticks because the daily lines resemble candles. The candlestick lines, alone and in combination, provide valuable assistance in trading. This report uses American names for the indicators. The “tsutsumi” line is a line that “engulfs” the prior day’s price action. So this is called an engulfing line. The Japanese names are in parenthesis. Indicators are illustrated and shown in chart examples. All charts are courtesy of Bloomberg Ltd. The patterns illustrated are representative examples. The lines don’t have to look exactly as they do in the illustrations to provide a valid signal. And, as with all charting methods, formations are somewhat subject to the interpretation of the user.

Reading the Candlestick Lines The daily candlestick line illustrates the market high, low, open and close. The thick part of the candlestick is called the “real body”. It represents the range between the open and close. When the real body is black (i.e., filled in), as in the Long Black Body below, it means the close was lower than the open. If, as in the Long White Body below, the real body is white (i.e., empty), it means the close was higher than the open. The thin lines above and below the real body are called the “shadows” and they represent the high and low of the day. The Long Black Body shows a close near the low of the day. The Long White Body shows a close near the high of the day. The relationship between the day’s open, high, low and close change the look of the daily candlestick.

LONG BLACK BODY

LONG WHITE BODY

high

high

open

close

close

open

low

low

Long Black Body This represents a bearish period in the market. Prices experienced a wide range, and the market opened near the high and closed near the low of the period. Long White Body This is the opposite of a long black body, and represents a bullish period in the market. Again, prices experienced a wide range, however, the market opened near the low and closed near the high of the trading period.


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

SPINNING TOPS high

high

open close

close open

low

low

DOJI LINES high

high

open & close low, open & close

low

Psychology and Markets

Systematic Trading

Umbrellas can be either bullish or bearish depending on where they appear in a trend. If they occur during a downtrend, they are called hammers and are bullish, as in “the market is ‘hammering out’ a base.” If an umbrella appears in an uptrend it’s bearish, and is referred to as a hanging man. Umbrella lines can be recognized by two criteria: 1) a real body at the upper end of the entire trading range, with little or no upper shadow, and 2) a lower shadow that is at least twice the length of the real body. The colour of the real body isn’t important. Hammers are excellent signals of a market at or near its bottom. At times, they are also important at the lower end of a congestion band. The ominous name of the hanging man line (derived because it looks like a hanging man with dangling legs) hints at its bearish nature. If with bearish confirmation one appears at the top of a prolonged uptrend, it’s usually time to vacate long positions.

Spinning Tops These are small real bodies, and can be either black or white. The small body represents a relatively tight range between the open and close for the period. In a trading range environment, spinning tops are neutral, but they may become important as parts of other chart patterns. (See stars and harami later in this report.) Doji Lines These illustrate periods where the opening and closing prices for the period are the same. The length of the shadows can vary. As you’ll see throughout this report, doji lines are important in a variety of patterns. In early June, a hanging man line signalled a reversal. Almost two weeks later, a hammer indicated the market was bottoming. The hammer’s lower shadow was a low in the sell-off.

Reversal Indicators Hammers and Hanging Man Lines (Takuri and Kubitsuri)

open & high close low

UMBRELLA LINES

at least 2 times body length

Bullish & Bearish Engulfing Patterns (Tsutsumi)

close & high

ENGULFING PATTERN

open low

HAMMER

Umbrella in Downtrend

BULLISH

ENGULFING PATTERN

BULLISH HANGING MAN

Umbrella in Uptrend

BEARISH

BEARISH

The engulfing pattern is a strong reversal signal, especially after a prolonged trend. It’s similar to the Western reversal pattern. Only the real body is important in this formation; shadows are virtually ignored. The bearish engulfing pattern has a black real body that engulfs the prior day’s white real body. This pattern is bearish during an

65


66

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

uptrend. Conversely, a white body at the bottom of a downtrend that engulfs the prior day’s black body is a potentially bullish signal.

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

The dark cloud cover has more significance as a top formation than the piercing line does as a bottoming formation. This is because tops are often formed faster than bottoms. Lows are made by indifference. Investors may be attracted to other markets, and in no rush to enter. This explains why bases often build over a relatively long period of time. At tops, however, there is more of a rush to get out. Thus, tops are often made more swiftly than are bottoms.

The engulfing line in early July called the high for the uptrend. Notice how this would not have been a reversal signal using traditional Western chart techniques - but was using candles.

Piercing Lines & Dark Cloud Covers

Near the end of May the piercing line hinted the week long downturn was over. The brief rally that followed was confirmed as over when the dark cloud cover appeared three days later.

(Kirikomi and Kabusi)

Upside Gap Two Crows

PIERCING LINE

(Narabi Kuro) UPSIDE GAP TWO CROWS low gap BULLISH

DARK CLOUD COVER BEARISH open high

Crows being an ominous bird, this is a bearish pattern. After a long white body we see a series of two black bodies. There is an upside gap between the white body and the first black body. Shadows are ignored. The second black body closes lower than the first. Although an upside gap is usually bullish in contemporary western analysis, this pattern is bearish in candles.

BEARISH The piercing line is a bullish pattern. This combination is composed of a long black body followed by a white body. The white body should open lower and then close above the center of the black body. Basically, the market gaps lower on the opening and then retraces to close above the midpoint of the previous period’s black body. If the white body does not “pierce” this halfway point, more weakness can be expected in the market. We have found this pattern to be most significant in a downtrend, or at the lower end of a congestion band. As its name implies, the dark cloud cover is a bearish pattern. This is the opposite of a piercing line. A strong white body is immediately followed by a black body. To qualify as a dark cloud cover, the black body must open above the high - and close below the center of - the previous white body.

In early February the upside gap two crows called an end to the upmove which began a month earlier.


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Tweezer Tops and Bottoms

Stars

(Kenuki)

(Hoshi) TWEEZER TOP

Psychology and Markets

Systematic Trading

There are various star combinations. All are reversal indicators and are more important after prolonged trends or large moves. A star is a small real body or a doji made on a gap that follows a long real body. Even if the shadows overlap, the formation is still considered a star, since only the real bodies are important. Morning and Evening Stars MORNING STAR

BEARISH

TWEEZER TOP

gap

BULLISH

BULLISH

EVENING STAR

A tweezer formation is simply two lines with matching highs or lows. The tweezer could be composed of candle lines with real bodies and/or dojis, and could occur on consecutive or nearby periods. The pattern is similar to the double top or bottom in traditional Western technical analysis. In a rising market, a tweezer top is formed when the highs, including any upper shadow, match on consecutive or nearby periods. This would be a bearish reversal indicator. In a falling market, a tweezer bottom is formed when the lows are the same (including any lower shadows). We have found that the tweezer is more important when it confirms either another bearish line (for a top) or bullish line (for a bottom).

BEARISH The morning star pattern is a signal of a potential bottom in the market. It is aptly called a morning star because it appears just before the sun rises (in the form of higher prices). After a long black body, we see a downside gap to a small real body. This is followed by a white body that closes above the midpoint of the black body made just before the star. The morning star is similar to a piercing line with a “star” in the middle. The evening star formation is the reverse of the morning star. Aptly named because it appears just before darkness set in, the evening star is a bearish signal. Basically, the evening star is similar to a dark cloud cover with a “star” in the middle.

In early July, a harami cross (described later in this report) showed the downward momentum was running out of steam. The bottom completed by this cross was essentially a tweezer bottom. This gave another indication of a market reversal. Two examples of the evening star within a period of a month. The second evening star confirmed that there was major resistance near $412 . This level was important support for most of March. Once broken, this support converted to significant resistance.

67


68

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Harami Lines

Doji & Shooting Stars The doji star appears after a prolonged move, and is composed of a gap and a doji line (remember a doji is when the open and the close are the same price). This is often the sign of an impending top or bottom. Doji stars often mark imminent turning points in the market, but more conservative traders should wait for the next day’s body to confirm a change in price trend.

HARAMI

DOJI STAR UPTREND

DOWNTREND

HARAMI CROSS

gap

BULLISH

DOJI STAR UPTREND

DOWNTREND

gap This is similar to an inside day in contemporary Western analysis. But while an inside day is usually considered neutral, the harami, line or cross is an indication of a waning of momentum. BEARISH

SHOOTING STAR

gap

The small body of the harami line is contained within the long body directly preceding it. (Harami appropriately means pregnant in Japanese). If the harami line is also a doji, it is referred to as a harami cross. These patterns indicate that the market is at a point of indecision and a trend change, or a reversal, is possible. We have found the harami cross pattern is useful in forecasting trend changes - especially after a long white body in an uptrend.

BEARISH

The shooting star pattern appears at short-term tops in the market, and is a bearish signal. As its name suggests, the shooting star is a small real body at the lower end of the price range with a long upper shadow.

Harami crosses a and c were a minor and major top. Harami cross b called the end of the downtrend.

Bearish doji and shooting stars appear from mid-September to mid-October.


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Meeting Lines (Deai) MEETING LINES

UPTREND

Psychology and Markets

Systematic Trading

In a rising market, a white body gaps higher. This is then followed by a black body which opens within the white body, and closes lower, but does not fill in the gap. It is important that any lower shadows also do not fill in the gap. The market should be bought on the close of the black body (as long as the gap was not closed). Normally, the black body would be considered bearish. But in this pattern, it is viewed as a temporary setback.

DOWNTREND

Meeting Lines occur when the market gaps higher or lower on the open, and then closes unchanged from the previous period. For example, in an uptrend, a white body is followed by a black body, and the closing prices meet. The reverse is true in a downtrend. A meeting line formation indicates that the prior direction of the market is uncertain. An upside tasuki gap appeared on June 22 and 23. Buying on the 23rd may have been recommended.

Windows (ku) DOWNTREND

WINDOW

window

Each set of meeting lines signalled that the short-term trend was uncertain. The first set demonstrated there would be no continuation to the downside. The second set showed there was no continuation to the minor rally that began a few days before.

Continuation Indicators The following candle formations are indications that price trends should continue. Price gaps within the patterns occur in each of these formations.

Upside Tasuki Gap

BULLISH

WINDOW

window

UPSIDE TASUKI GAP BEARISH gap

BULLISH

A window is the same as a gap in contemporary western analysis. While we say “filling in the gap”, the Japanese expression is “closing the window.” Our experience is that gaps often become support or resistance areas. And windows (i.e. gaps) are viewed in the same context as support or resistance. Shadows are also considered in closing the window. Unclosed windows signal continuation of the trend.

69


70

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Side-by-Side White Lines (narabiaka)

SIDE-BY-SIDE WHITE LINES

gap

BULLISH In mid-August, the window became support on a pullback a few days later. SIDE-BY-SIDE WHITE LINES

Rising and Falling Three Methods RISING THREE METHODS

gap

new high close BEARISH

BULLISH

In a rising market, two white bodies with the same opening prices form on an upside gap. This pattern is bullish and the gap should represent a strong support area. We have found it especially important as a break from a low price congestion area. However, side-by-side white lines following a downtrend are bearish, and viewed as temporary short covering.

FALLING THREE METHODS

Miscellaneous Doji Indicators

new low close BEARISH This is a rare, but important, candle pattern. The rising method is bullish in an uptrend. A long white body is followed by a pullback via a series of three or so small white or black bodies. Finally, another strong white body, which makes a new high close for the move, completes the formation. The falling method is bearish in a declining market.

During the latter part of October, an almost perfect example of the rising three method pattern. Augmenting this bullish signal was the fact that volume picked up on the two white body days.

Doji lines reflect indecision. If you see two or more doji lines within a short time in a market where this normally doesn’t occur then a forceful move is possible. Double dojis may foretell an increase in market volatility. Option traders who are confident of price direction could use this signal to buy options (assuming volatility levels are attractive). They could benefit from premium expansion based on increased volatility and a significant price move. For those not confident about the direction of the break, a long straddle or strangle (or similar long volatility plays) may be considered. Doji days can become support or resistance, usually on a shortterm basis. And a series of three doji lines after a prolonged move could signal a rare and important top or bottom.

The double doji in mid-September reflected a large degree of uncertainty that was resolved a week later in a sharp move to the upside.


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

Putting it together

A final note from the author

We’ve gone over many lines and patterns. Let’s put them into context and have some fun. Included is a chart of December 1989 Copper Futures prices with numbered lines and patterns. All have been discussed in this report. How would you interpret them? Our opinion of what these indicators mean follows. Try to figure them out before looking at our interpretation.

The subject of candlesticks offers an interesting contrast - they’re old and they’re new. They are based on the oldest form of technical analysis, yet they offer new avenues of research in the U.S.

Remember the interpretations are subjective. You may see something different or see indicators where we haven’t. As with any charting technique, different experiences will give different perspectives. There are no concrete rules, just guidelines. For example, what if a hanging man had a small upper shadow? A purist may say that was not a hanging man. But we would liquidate long positions with such a line. Others may prefer to wait a few days for confirmation.

Bloomberg and Commodity Quote Graphics offer candle charts. Since my introductory article appeared in the December 1989, other vendors plan to come on-line with candle charts. These include FutureSource, Trade Center, Market Vision, Bonneville Telecommunications and CompuTrac. In addition, Knight Ridder has software to convert prices to candlesticks. CRB and Knight Ridder can provide individual candle charts. With numerous services supplying candle charts, they should be easier than ever to follow.

Copper Candle Chart Indicators: 1. Bullish engulfing line. Although more important after a downtrend, this line also should be respected during a lateral trading band 2. A doji line at new highs could mean a top. However, the next day’s bullish line didn’t confirm this view. Notice how this doji line then became support on the sell-off the following week. 3. Bearish engulfing line. 4. Double doji - expect a big price move. 5. A doji line that became a support area. 6. A hanging man signals a top. 7. The harami pattern shows the large down day was followed by a day of uncertainty. The downtrend lacked follow-through. Prognosis: For now a downleg is unlikely. 8. Dark cloud cover giving another hint of a top. 9. A bullish engulfing line. 10. A bearish shooting star. 11. A bearish engulfing line called the top of this market.

Candle charts are exciting - where else can you talk about a hanging man line? They are powerful - they can be used alone or in conjunction with other techniques. And they are timely - virtually unknown in the U.S., yet underpinned by centuries of use. We can use the unique insights provided by the candlesticks to elevate our market understanding. “If you have knowledge, let others light their candles at it” Margaret Fuller.

Steve Nison CMT was Vice President Technical Analyst/Option Strategist at Merrill Lynch Futures Research, and an officer of the MTA. He had a book published early in 1991 on Candlestick Analysis, joining Candlestick Technicals with Western technical tools. This article is reprinted by permission. Copyright © 1990 Merrill Lynch, Pierce, Fenner & Smith Incorporated. It was written in conjuction with the MTA’s Chartered market Technician Program.

71


72

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Japanese Candlestick Analysis An Analytical Approach to Japanese Candlestick Charting

Dan Gramza

Article originally featured in Market Technician 12 (October 1991)

I was introduced to Japanese Candlestick charting while in Tokyo in 1988, and I have researched and traded with the approach since that time. In using Japanese Candlestick charting, I have found that the traditional trend reversal and continuation pattern recognition can be enhanced by exploring the dynamics of pattern development.

Figure 1

Japanese Candlestick Analysis introduces the element of interaction between buyers and sellers to the typical Japanese Candlestick application of pattern recognition, providing the benefits of pattern confirmation, and additional insights into market balance and imbalance. The elements of consideration are: 1. 2. 3. 4. 5. 6. 7.

Candlestick Colour Candlestick Range Candlestick Body Size Shadow Location Shadow Size Candlestick Development Candlestick Location

The first five elements focus on the traditional characteristics of the candlestick itself. The last two elements examine the market dynamics and the market environment in which the Japanese Candlestick occurs. It is the information that is gathered from these final key elements that helps to determine the validity of the pattern that is forming; and it is this information that is lacking in a totally mechanical pattern recognition approach.

1. Candlestick Colour While colour is not a consideration in the harami, hammer, hanging man or star patterns, the colour of a candlestick generally provides a visual indication of buying (white body) or selling (black body) activity for the period. Likewise, the colour of a series of candlesticks can indicate whether buying or selling pressure is dominant or balanced in the longer-term view. Typically, a white candlestick is the result of buying pressure pushing prices higher to close above the opening price. If buyers are not present, and sellers dominate, prices will move lower and close below the opening price, resulting in a black candlestick. For example, in the weekly chart shown in Figure 1, the white candlestick occurring in the first week in May implies that buying activity prevailed on the week. The series of black candlesticks in Section A of Figure 1 implies that selling pressure is maintained, causing the downtrending market. Similarly, the series of white candlesticks in Section B implies that buying pressure is being maintained, causing the uptrending market.

Source: Aspen Graphics

The alternating black and white candlesticks of Section C indicate a struggle between the buyers and the sellers, resulting in a balanced, sideways market with neither side dominating. A candlestick’s colour can also imply profit taking. In Section D of Figure 1, a white harami candlestick occurs after six weeks of a downtrending market, raising the question: Are buyers entering the market or are sellers taking profits? In this case, lack of followthrough, as represented by the doji formation the next period, suggests sellers taking profits, not new buyers entering the market. The resumption of the downtrend is expected; and this theory is confirmed by the return of sellers during the following period, represented by the black candlestick.

2. Candlestick Range The candlestick range can be a general indicator of the level of volatility and activity for the period when compared to previous periods. The length of the candlesticks in Sections A and B in Figure 1 are relatively long, as the market trends downward and upward. We see short candlesticks at the beginning of Section A and at the end of Section B, as the previously trending market slows down. This change from long to short candlesticks provides a general indication of decreasing market activity.

3. Candlestick Body Size Candlestick body size can serve as a measure of confidence and conviction on the part of the buyers or the sellers.


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

A large white candlestick body suggests that buyers are in control. A large black candlestick body suggests that sellers are in control. A candlestick with a small body or without a body, as in a doji formation, indicates uncertainty and indecision on the part of the buyers and the sellers, with neither dominating the period’s activity.

to rally by gapping higher, sellers drive prices lower. Finally, as in Section A, a large white daily candlestick results; however, upward continuation is in doubt. The shorter timeframe examination reveals that the daily session, dominated by selling, ends with an evening star pattern which is a potential reversal pattern.

4. Shadow Location

In both cases, a white daily candlestick is the result of the market’s activity; but it is only upon closer examination of how the daily candlestick was formed, using intraday candlesticks of generally 60, 30, or 15 minute duration, that its strength or weakness is revealed.

A shadow represents the difference, if any, between the body and the high and low, and can be another indicator of buyer or seller strength. In Figure 1, the upper shadows during the second and third weeks of July, the first two weeks of March, and the third week in April show a lack of buyer strength, and the buyers’ inability to close the week above the 3000 level.

5. Shadow Size The length of a shadow can indicate how a market reacts to certain price levels. The longer the shadow, the greater the potential market rejection of the price levels tested, suggesting that the market may not be able to continue in that direction. In Figure 1, the long upper shadows for the second and third weeks in July, the first two weeks of March, and the third week of April indicate market rejection and a lack of buyer confidence as the Dow moves above the 3000 level. A show of buyer confidence would have resulted in a large white candlestick with possible small shadows closing above the 3000 level, as seen in the first week of May.

6. Shadow Development An examination of a candlestick’s development by using shorter timeframe candlesticks can offer insight into the strength or weakness of the candlestick, as well as illuminate the buying or selling activity that created it. Figure 2: Candlestick Development

7. Candlestick Location The significance of a reversal or continuation pattern can often be measured by examining the market environment in which it occurs. This is a two-part consideration. The first consideration is candlestick location in relation to previous candlesticks. Its location helps to determine whether the candlestick is the beginning, an element of, or the end of a reversal or continuation pattern. The second consideration is candlestick location with respect to the overall market environment - contract highs or lows, seasonal reference points, previous high volume prices, etc. The more mature the market, the more important the forming reversal pattern becomes. Past levels of support and resistance may also provide clues into possible outcomes as the pattern unfolds. This information can assist in trade management decisions, such as whether the trade should be initiated or when an existing position should be reduced or exited.

Source: Gramza Capital Management, Inc.

Figure 2 presents two very different, yet viable, breakdowns of a large white daily candlestick’s development. In Example A, the buyers dominate the trading session, with the intraday white candlesticks trending up with higher highs and higher lows. The market opens on its low and closes a little down from its high, leaving a small shadow - which indicates possible profit-taking at the end of the session. Upward continuation would be expected. Example B, however, reveals a market with a strong seller presence, as represented by the large black intraday candlesticks which are longer than the white intraday ones. Each time the market attempts

In Figure 1, the reverse hanging man on the third week of July, is a potential reversal pattern, and requires a closer look. It occurs after the previous week has tested, and failed, at what is considered to be a significant high level (the Dow above 3000). The reverse hanging man formation is confirmed the following week by a black candlestick that also breaks the identified support trendline. With Western technical analysis, the conclusion for this week could be favourable because the market has made a higher high and a higher low. Japanese Candlestick Analysis, however, reveals possible market weakness. In this case, the reverse hanging man may have alerted a trader who was long to reduce or exit the position, or provided an opportunity to short the stock market. The underlying concept of Japanese Candlestick Analysis is the creation of a thought process to evaluate Japanese Candlesticks. Typically, the identification of formations and patterns is mechanical; however, the application of Japanese Candlestick Analysis provides a revealing and informative process of market evaluation.

73


74

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Measuring an Extended Market with Japanese Candlesticks Daniel Gramza

Article originally featured in Market Technician 18 (November 1993)

What clues does a market reveal when it is becoming overextended and nearing a correction or a change in direction? To answer this question, many traders would begin to think of oversold or over-bought oscillators; but long before these mathematical studies were created, the Japanese had a simple, direct way of determining when a market was becoming over-extended and due for a correction. It is a candlestick pattern called 8 new records or 8 new price lines. This pattern is based on the number of new highs or new lows made by the market, and it is formed when the market has made 8, 10 or 13 new highs or new lows from the trend’s beginning. Each higher high is considered a new record high, and each lower low is considered a new record low. Experience in applying this approach indicates that eight to nine and 12 to 13 new record highs or lows occur most frequently. The expectations of a correction or reversal are increased when these higher highs or lower lows occur consecutively; but consecutive highs or lows are not necessary for the formation of the pattern. This is shown in Figure 1. The candlesticks without numbers failed to make new record highs or lows.

man pattern. The eighth new record low on the same chart results in a possible bullish morning star, which suggests that a short position should take some profits. Figure 2

Figure 1 Source: Knight-Ridder ProfitCenter

The significance of exiting a trade with the new records pattern increases when the new record highs or lows culminate in a reversal candlestick pattern such as the evening star pattern on the daily February Pork Belly chart in Figure 3, or the unusually large, modified bearish evening stars occurring at the 11th and 12th new record highs of the daily November Soyabean chart in Figure 4. Figure 3

Source: Gramza Capital Management, Inc.

The intention of this candlestick pattern is not to reveal the absolute highs or lows of a market move, but rather to assist the trader in reducing or exiting a position before the market corrects, and then to step aside and rest, or possibly reverse their position. This philosophy is illustrated in the Japanese sayings, “One should rest after eight new price lines”, and “The stomach is 80 per cent full” which means after eight new records profits should be taken. The suggestion is to exit part of the position (one-half or up to 80 per cent) after the eighth new record high or low, and to exit the remaining portion of the position as the market reaches 10 to 13 new record highs or lows. The weekly October Sugar chart in Figure 2 provides an example of how this philosophy is applied. In Figure 2, part of a long position would be taken off after the eight new record high, and the balance of the position liquidated after the 11th new record high, which happens to be a bearish hanging

Source: Knight-Ridder ProfitCenter


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Figure 4

Psychology and Markets

Systematic Trading

Figure 6

Source: Knight-Ridder ProfitCenter

Source: Knight-Ridder ProfitCenter

The new records concept not only provides insight into how a market is currently unfolding, but also into how it may reoccur in the future. The June 1992 US Treasury Bond contract chart in Figure 5 and the September 1993 US Treasury Bond contract chart in Figure 6 illustrate this point. The first rising leg in both of these figures occurred over approximately a four-month period and had 14 new record highs before retracing. The second rising leg in Figure 5 had nine new highs before reversing and the second leg of Figure 6 has had eight new record highs before a small correction.

new record count is the continuation of the new record count of the first leg of a market move. In Figures 5 and 6, this extended count results in 21 new record highs, as noted by the numbers in parentheses. The 21 new record highs in Figure 5 suggest a similar market correction in Figure 6, and this occurred on the third week in July with a bearish black candlestick. Figure 7

Figure 5

Source: Knight-Ridder ProfitCenter

The extended new record count of 13 in the August Gold chart in Figure 7 also demonstrates a market that may be reaching its limit, with a bearish black candlestick as the 13th high. This chart also provides a good example of a market taking a rest after the eighth new record high and the bearish evening star on the ninth new record high.

Source: Knight-Ridder ProfitCenter

Another interesting aspect of these two charts is what I refer to as an extended new record count. This is not part of the traditional eight new record theory, but it can be used to provide a broader view of a market that may be stretching to its limit. This extended

The new records pattern should not be the only criterion, but it does provide a simple, direct method of measuring an extended market.

Daniel Gramza is President of Gramza Capital Management, Inc. 2227 Foreswiew Road, Evanston, III 60201

75


76

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Trending with Heikin-Ashi Dan Valcu

Article originally featured in Market Technician 49 (April 2004)

Abstract

Figure 1: FTSE-100 - Candle chart vs. Heikin-ashi chart.

The article introduces an almost unknown Japanese trend technique, Heikin-ashi, which is based on modified open-high-lowclose values and is displayed using candlesticks.

Uptrends and downtrends are easier to identify on Heikin-ashi chart.

It is a visual instrument which assesses the current trend and its strength. The Heikin-ashi technique, also known as ‘modified candles’, takes the noise out from the price and displays a smoother evolution of the stock price or index value. This technique is mainly used as a visual way to assess the trend, but it can also be quantified with indicators. It can be applied successfully to all instruments with open-high-low-close values (stocks, options, commodities, futures). Heikin-ashi charts and related indicators are easy to generate by most of the technical analysis software packages available on the market. They can also be produced using common spreadsheet software available on all personal computers. The article will also discuss how this technique may apply to UK stock market.

What is Heikin-Ashi? Last summer, while researching for Ichimoku charts material, I came across a very rare way to look at trends: Heikinashi or modified candlesticks. At that time there was only one English language site which displayed charts built using this technique. Now there are a number of sites and discussion forums which focus on this form of analysis. However, it is still very difficult to trace its origin due to the lack of available literature in English. I could only obtain the way the modified OHLC values are computed from a Japanese trader. From there, interpretation and quantification as technical indicators came as logical steps. Let’s start with the definitions of modified values and continue with discussions about use of this technique: Table 1: Computation of Heikin-Ashi Values Modified OHLC

Mnemonic

Definition

Comments

Modified Close

HaClose

(O+H+L+C)/4

O-H-L-C values are for the current bar

Modified Open

HaOpen

HaOpen (prev.bar) + HaClose (prev.bar))/2

Modified High

HaHigh

Max (H, HaOpen, HaClose)

H value is for current bar

Modified Low

HaLow

Min (L, HaOpen, HaClose)

L value is for current bar

Case Studies The best way to understand Heikin-ashi is to apply it to familiar instruments. For example, let’s apply the code to display Heikinashi / modified candlesticks (see box on next page) to the FTSE100 index: The periods associated with ascending trends are shown as sequences of continuous white candles, while descending trends are displayed as black filled candles. A trend may be strong, weak, or ready to change. A strong uptrend appears as a sequence of long white candles with no lower shadows. A strong downtrend is associated with a sequence of long black candles with no higher shadows. The reduction in size for the candles during a trend may signal a weakening of the current trend. Periods of consolidations are associated with Heikin-ashi candles having small bodies, but longer upper and lower shadows. And finally, possible trend changes are identified by a candle with a small body having both high upper and lower shadow. The table below summarizes these basic rules to identify and assess the strength of any trend using the Heikin-ashi technique:


Indicators and Momentum

Rule #

Rule Description

Elliott Wave and Fibonacci

Area on Chart

Gann Analysis, Cycles and Forecasting

Comments

1

Uptrend

A-B, C-D, E-F, G-H, K-L, M-N, O-P, R-S

White candles

2

Downtrend

B-C, D-E, F-G, J-K, L-M, P-R, S-T

Black candles

3

Strong Uptrend

A-B, G-H, O-P, R-S

white candles with no lower shadows

4

Strong Downtrend

F-G, J-K

Black candles with no high shadows

5

Consolidation

C, E, I between N-O, T

Candles with small bodies (white or black) and with upper and lower shadows

6

Change of Trend

C, I, F, first white candle after R

Similar to #5, but only one occurrence. Not always reliable as it may also be part of a consolidation sequence

Taking into account the findings in the previous table and looking at the second subchart, we can see some obvious trends. The main advantage of Heikin-ashi technique is that the modified OHLC values are displayed as Japanese candlesticks. This powerful representation makes the trend clear for everybody. The original Japanese candlesticks theory is very complex and it requires very good skills for an accurate translation of patterns. Heikin-ashi charts eliminate the need for a sophisticated understanding of Japanese candlesticks theory and their related patterns. Since this technique is simple and easy to understand, let’s use and discuss it for another example from the UK stock market, LogicaCMG. Figure 2 shows three different subcharts: regular candles, modified candles, and Heikin-ashi indicator HaDiffCO together with its 3-day moving average. As mentioned before, Heikin-ashi is a visual technique and as a result, there is a certain degree of subjectivism when we try to translate what we see. To avoid this and make the process more mechanical, I have created an indicator, HaDiffCO, which is simply the difference between HaClose and HaOpen:

HaDiffCO = HaClose - HaOpen HaDiffCO is nothing more than the value of the modified candle’s body. When this indicator is positive, the Heikin-ashi candle is white. Negative values for HaDiffCO are equivalent with black-filled Heikin-ashi candles. If we look simultaneously at subcharts #2 and #3, we see the equivalence I just mentioned. There are two immediate advantages of having this simple indicator: (a) for those who use software with no built-in capability to display modified candles and (b) possibility to apply moving averages or other indicators. The indicator is smoothed with its 3-bar moving average and the resulting crossovers suggest entry and exit points. Going one step further (not shown in this article), we may apply another moving average to the moving average and the intersections suggest another set of entries and exits (the lag introduced by moving averages will delay these signals).

Psychology and Markets

Systematic Trading

Conclusions Heikin-ashi trend charting technique is simple to implement either as candlesticks or as traditional technical analysis indicator(s). This can be done with most of the software packages available on the market. It is a visual way to look at the trend and assess its strength. It works with all timeframes and there are a small number of simple rules required to understand in order to make an easy and correct judgment of any trend. It fits well in the category of visual Japanese techniques, such as Kagi, Renko, and three-line break charts. Heikin-ashi charts are not the solution for instant success in the markets, but they help to enter and remain on the right side of a trend for a great deal of its duration. It is highly recommended to use modified candles together with other independent indicators, patterns, or techniques.

Further reference (1) Valcu, Dan “Using the Heikin-Ashi Technique” Technical Analysis of Stocks and Commodities: February 2004. (2) “Trader’s Tips” Technical Analysis of Stocks and Commodities: February 2004. (3) http://www.educofin.com (4) http://plaza4.mbn.or.jp/~skoike/heikin.htm

How to build and display Heikin-ashi charts using technical analysis software For this article, all formulae and charts have been generated using AmiBroker v4.50 and are described below as implemented for the article (1). Code to display Heikin-ashi (modified) candlesticks: HaClose = (O+H+L+C)/4; HaOpen = AMA( Ref( HaClose, -1 ), 0.5 ); HaHigh = Max( H,Max( HaClose, HaOpen ) ); HaLow = Min( L,Min( HaClose, HaOpen ) ); PlotOHLC( HaOpen, HaHigh, HaLow, HaClose, “Modified” + Name(), colorBlack, styleCandle ); Code to generate Heikin-ashi indicators HaOpen, HaClose, and HADiffCO with its average: HaClose = (O+H+L+C)/4; HaOpen = AMA( Ref( HaClose, -1 ), 0.5 ); HaHigh = Max( H,Max( HaClose, HaOpen ) ); HaLow = Min( L,Min( HaClose, HaOpen ) ); HaDiffCO = HaClose - HaOpen; per = Param(“MA Periods”, 3, 3, 50, 1 ); Plot( HaDiffCo,“HaDiffCO”, colorRed ); Plot( MA( HaDiffCo, per ),“MA(“+per+”)”, colorBlue ); Daily data from Yahoo Finance UK&Ireland: http://uk.finance.yahoo.com/

Dan Valcu is a Swedish private investor and IT consultant. His website, www.educofin.com focuses on trends and trading using Heikin-ashi and trend-based proprietary systems. He may be reached via email at ta@educofin.com

77


78

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Figure 2: LogicaCMG - Quantifying Heikin-ashi candles

Market Profile Peter Steidlmayer

Article originally featured in Market Technician 29

The market is a dynamic self-organizing system whose purpose is the elimination of the episodic volatility that punctuates trading activity. This purpose is achieved through a method that brings these periodic flare-ups of inefficiency back to an efficient or fair price. The force that accomplishes this end is always present, always active, and always trying to get control. It is the essence of the market. Periods of volatility undergo an underlying course that can be seen and quantified. The lack of awareness of this process and the inability to plumb its dimensions result in the less than adequate performance of the majority of market participants. Most successful traders have an intuitive grasp of this dynamic.

Market Profile Capital Flows is a system that objectifies this organic progression. It affords the market the opportunity to express itself clearly by use of a system of data arrangement that mirrors the cycle of elimination of inefficiency and allows eventual establishment of a point of market efficiency. An example will illustrate this point. A firm releases earnings that exceed market expectations. A flood of buy orders push prices dramatically higher. An imbalance has occurred. Sellers enter the market and push it in the opposite direction. This can be a short or long term phenomena, but eventually inefficiency (vertical movement) is eliminated and an efficient price is established and accepted (horizontal movement) over time. Equilibrium has been found. This becomes the base for the next break out of market inefficiency. The market begins the process of reining in this new inefficiency. This is the cycle of action of the market establishment of efficiency (horizontal development over time): inefficiency enters the market (vertical movement), the market bridles in and eliminates the inefficiency over time, and efficiency is re-established. This is the


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

market expressing itself in market time. Horizontal development is the set-up for eventual change. Change (vertical movement) will occur and be factored out over time. Efficiency will then reestablish itself. Market Profile is a data base/charting system that permits the market to measure these relative periods of vertical movement and the corresponding periods of horizontal development by the market’s internal yardstick - market time. A chart only has value if it communicates market condition. The bar chart has a distinct limitation. It communicates freely only in the vertical dimension. The horizontal is controlled by chronological time. It automatically moves to the right when the clock or calendar changes Chronological time, not market time, controls the horizontal dimension. Bar charts are thus a limited vehicle of expression. The following example of Capital Flow data organization will illustrate how this limitation is over-come and the market freed to express vertical and horizontal movement in market time. Each 30 minute period of the day is assigned a letter of the alphabet, either upper or lower case. As a price trades during a period the letter is placed on the chart at its corresponding price. Each unit or bar contains a minimum of two 30 minute periods. The unit splits off to form a new group when the current letter exceeds the range of the previous period. The June Japanese Yen contract, which trades at The Chicago Mercantile Exchange is used to illustrate this.

Psychology and Markets

Systematic Trading

The market’s transition from horizontal to vertical and back is thus communicated in its own natural time through proper data base arrangement. The benefits are profound. Studies can be applied to the data base that are measured in units of market time (relative vertical and horizontal development) as opposed to studies based on chronological criteria.

Here is another example using the Japanese Yen contract. This example comes from Page 2 of our database. Page 2 represents a longer term study of our database. This study has the advantage of measuring market activity over a larger sample of time periods. Each bar contains a minimum of sixty 30 minute periods. When a price exceeds the vertical range of the last thirty 30 minute periods, a new bar begins. A control price for the previous bar is established and highlighted. The date the split occurred is marked on the new bar. Once again change on this page is dictated by market activity and not chronological time. The number of thirty minute blocks in each bar is dictated by market condition. It can be as few as sixty in a volatile, vertical market or over one hundred in a horizontal, price control market.

D period on the 19th trades beyond the range of C period so a new bar begins. Z period on the 19th trades within the range of Y period so no new bar begins. This bar contains three 30 minute periods. The Z period did not split off because the market did not indicate a change in its condition. As any given day moves along the time line from left to right, it could form as few as six to eight bars or as many as twenty bars. It is contingent upon relative horizontal and vertical activity. The next bar is not determined arbitrarily by chronological time. It is dictated by market time. A proprietary algorithm assigns a control price to the previous Market Profile unit as each new unit begins. Starting with the N-P period on the 19th the market remains very horizontal during the Q-R and S-T periods. We look to these tight horizontal ranges as the vertical starting points of vertical price activity. Relative inefficiency manifested itself in the v-w and w-y-z periods. The market became almost one-dimensional (vertical) in the a-b period. The market tried to rein in this vertical move in the c-d-e and f-g periods but the inefficiency re-asserted itself as the day progressed. This example illustrates the proper use of the horizontal, and vertical axis in developing an objective representation of market activity. Chronological time is avoided as it is arbitrary. Market time is freed to express its inherent dimensionality in its own time. Development which is manifested in the horizontal is the real market time. Development begins with the initiation of a vertical move that must be brought to efficiency. When the vertical range has been established, the market drifts into more of a horizontal phase. This development of efficiency represents the time clock of the market. The cycle of market time is thus completed.

The benefit of utilizing a Page 2 Study in conjunction with the shorter term study is that it allows one to view market condition from a longer term perspective by incorporating more time periods in each bar. The current bar begins on May 21. The numbers on the bottom indicate the date that each new bar begins. It is evident that most of April was spent in horizontal development between 7900 and 8000. It had moved vertically to enter this area and was then locked in horizontal development waiting for a new beginning. This strong vertical move began in early May and lasted approximately two weeks. Attention should be paid to the length of these bars and the location of the control prices. The bar length indicates the dominance of the directional move. The control price placement on the extreme of the first two bars underscores this dominance. The best use of Page 2 is keeping one in a trade that has directional integrity or to alert one to avoid going against a long term trend. Note that in this example there were eight successive higher control prices once the vertical move began. Although the ranges of the most recent five bars are much more vertical than April’s bars, they are beginning to overlap each other. This indicates that the market has become more horizontal than earlier in the month. This longer term perspective is an invaluable complement to the shorter term study shown in example one. Numerous other studies exist in the database. All are market based in that they are derived from the horizontal and vertical language in which our database allows the market to express itself. A limited data base affords only limited opportunity. A data base that expresses itself freely in both dimensions opens wide the doors of opportunity.

79


80

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

CHAPTER FOUR

Pattern Recognition and Pattern Analysis Articles in this chapter 82

Who Says Patterns Never Repeat Themselves? Andrew Rose

84

Complex Patterns Josh Bruzzi

87

Statistical Evaluation of Classic Chart Patterns Gregor Bauer

90

Some thoughts on the Mathematics of Turning Points Tony Plummer

96

Some Reflections on Stock Market Interrelationships Anne Whitby

Moving Averages and Trends


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

CHAPTER FOUR INTRODUCTION

By Nicole Elliott FSTA

Introduction One’s maternal tongue is picked up via osmosis; foreign languages are taught (usually badly) at school. Some languages are considered difficult, like quantum physics and algebra - though not arithmetic, which seems to be part of our DNA. Other languages require serious input from the teacher and the student, but can be so very rewarding, like the ability to read music, and with it the joy of hearing a whole orchestra play in one’s head. I would put technical analysis in this last category. I’m not talking about dabbling with a histogram and a couple of averages; that’s the stuff of economical economists. I’m talking about reading the nuances behind price moves in financial markets, coming up with decent reasons for this, and extrapolate where we might go next - profitably. In a way, bar and candlestick charts are a type of shorthand which, you might like to know, is still taught at the Financial Times today. I grew up in a very visually-orientated household, my mother especially, who was a fashion magazine illustrator and tailor, making her own paper patterns. As toddlers my sister and I were dragged around museums, churches and archaeological sites, noting the angle of the leaning Tower of Pisa, the scale of the dome on St Peter’s, Rome, and the colours - and fat ladies of Rubens paintings. I think this helped my enforced push into technical analysis (my first City boss told me I would be their chartist; full stop). My eureka moment arrived when I went on a two-day course and conference organised by the late David Fuller. We were a big audience so I sat at the front to hear and see clearly; and I ‘got it’. I recently learned that he amassed a fantastic collection of watercolours (which he sold for £1 million) - which didn’t surprise me. You see, he too, like me, was about line, design, pattern, nuance, and colour. So you can understand why I gravitated towards the more ‘arty’ side of the profession; I can equally understand why more rigorous analysts would steer towards the ‘scientific’ end of the scale, using oscillators, back-testing and algorithms; it’s horses for courses. I’m light bulb, inspiration, flash-in-the-pan, some might say; but I could retort with plodding, boring and slow. And, by the way, in 1986 in the bank’s treasury dealing room we were experimenting with automated pattern recognition using one of Microsoft Windows’ earlier versions - a beast. The most common drawback was that the machine was not sensitive enough to detect subtleties and variations on a theme. Then again, I could see why you might want to reduce the number of humans analysing patterns - and them arguing about them!

Psychology and Markets

Systematic Trading

Now, to the articles in this chapter, 5 of them, which I think reflects how many people are drawn to this subject, core to technical analysis and often an introductory topic as it’s easier than some others. The paper by Andrew Rose, first published in 1986, must be up there in the pantheon of early STA work. Rather than repeating patterns, I’d say his focus is on trends making higher highs, lower lows and vice versa, to which he adds the concept of a 1 x 1 line - very interesting. The article’s real strength lies in its summary of the outlook for investors, and tactics for traders; top class, as one would expect from a leading stockbroking firm. The complex patterns tackled by Josh Bruzzi (in 1995 judging by the dates on his charts which, incidentally, look rather like those in Edwards & McGee’s book first published in 1948) are triangles, wedges and more especially broadening patterns which he, like me, find hard to work with; Elliott Wave counts feature heavily. To my surprise he says, “few text books mention broadening patterns’’; maybe this was the case, but I don’t remember. Here I’d like to interject, suggesting the Market Technician should date all published papers - for future reference. I’d also suggest more space is allocated to detailed charts. In Gregor Bauer’s ‘Statistical evaluation of classic chart patterns’ he contends these are a “cornerstone of technical analysis”, with which I think we’ll all agree. His paper’s based on Thomas N Bulkowski’s 2002 book in which he looks for excess returns (alpha) on the S&P 500 index versus the index’s rally (beta). His tables are worth looking at, but of more interest are the ‘failure’ rates where we’ve seen a generalised increase between 1991 and 2008; pity we don’t have subsequent data. Tony Plummer’s view that “price movements [of financial markets] appear to be bounded by mathematics’’ focuses on the Golden Number, Fibonacci relationships, and a little Elliott Wave. Analysing the 1987 stock market crash, he feels the event was “an assault on the nature of capitalism’’. Hardly the stuff of simple, classic chart patterns. As a multiple auntie and godmother I’m really not allowed to say this, but I have my favourites. Hence, so looking forward to Anne Whitby’s piece on stock market interrelationships. Nothing about patterns, but she’s clearly at the forefront of globalisation, looking at developed and developing market stock indices since 1980. Loved her comment about business news: “what did Japan do today?’’ Reminds me of Edo and the floating world, Hokusai and Mount Fuji; another time and place long, long ago.

81


82

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Who Says Patterns Never Repeat Themselves? Andrew Rose

Article originally featured in Market Technician 2 (May 1988)

The following article is a combination of extracts from a piece written in September ‘86 for clients of James Capel, and an update as at May 1988. The initial piece was entitled “Methods of Chart Analysis”.

Chart 2: Cable & Wireless

Part One - Then The first step in our analytical process is the establishment of the 3 week trend using a “swing chart” of the 5 year weekly range graph. This involves looking for 3 consecutive higher weekly range tops and bottoms to instigate an upswing, and 3 consecutive weekly ranges with lower bottoms to start a downswing. We see a pattern of higher tops and bottoms on this swing chart as showing that the long term buying cycles are beginning and ending at progressively higher levels and that therefore the outlook is bullish. Investment Data Services

Chart 1: Cable & Wireless

Investment Data Services

Chart 1 shows the 5 year weekly range chart for C&W with the 3 week trend superimposed on the left hand side and corresponding highs and lows duly marked. The higher tops and bottoms pattern is in place and will remain so unless the current downswing breaks fully beneath the last major low (point H) at 272p. On this basis C&W still rates a core-holding. The next step is to establish what support there is above 272p which would delay a breakdown of the long term uptrend, or indeed whether there is anything in this area likely to reverse the trading trend back up and thereby confirm the long term uptrend with another major low on the chart. It could be a fundamental reason causing such a rally, on the other hand it could be a technical reason; most likely, especially if it is to be a reversal worth following, it would be both. Technical indicators to focus on when trying to establish such support levels include:- momentum angles from major trend change points, and percentages up or down from them.

Chart 2 shows a line, marked 1 x 1, drawn from the all-time low on the stock price (which adjusted for - issues etc. stands at 61p). This line is the graphical representation of a one point rise per week. What is interesting and important about this 1 x 1 momentum angle is how it makes the whole chart seem somehow clearer. For the first 3 years of the stock’s existence one point per week, proved to be the maximum rate of rise. There were times when the chart showed above the 1 x 1 but crucially there were no confirmatory lows above the angle and in fact each move above the line proved a useful overbought indicator. In early ‘85 there was another move above the line and this time there was a confirmatory low, and with the bullish higher tops and bottoms pattern continuing suddenly we had a different scenario. The 1 x 1 had become a supportive not a resistance factor and subsequent dips below the line proved to be a useful oversold indicator. This is why we have marked the break point as a gear shift. The trends were up and suddenly the rate of rise was not a maximum, but a minimum, of one point per week. On this basis the current dip beneath the angle would seem to be just another oversold indicator, but there is a problem, namely a breakdown in the basic bullish chart pattern of ascending highs and lows. This is a summary of the pieces we have assembled so far:1. The long term indicators are bullish and will continue to be above 272p. 2.

The trading indicators have however broken down, leaving the stock vulnerable to further weakness and signalling the buildup of defensive positions in the move back to test the longterm trend support at around 272p.

3. A major top below 313p will signal the end of C&W’s advance at a rate of more than 1 point per week. This will make it, at best, a stock in which to sell call options. 4. The top at 360p is looking particularly significant,


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

especially since the bearish angles from this trend change point are now having a direct effect on the chart’s progress. 5.

The chart pattern after the top at 360p is bearish. Unsurprisingly therefore we maintain a cautious stance on the stock prior to what looks like being the final test of whether or not the chart has broken down on a long-term basis i.e. whether or not it goes on to form a top below 272p.

Part Two - Now Chart 3: Cable & Wireless - Investment Data Weekly High / Low

Psychology and Markets

Systematic Trading

Mark’s Musings The main thrust of the argument in the last musings was that it is still vital to distinguish between information and judgement. The former is only a means to an end - the latter. Investors, certainly the professional ones, are grossly over-informed by the plethora of information thrown at them. Not only is this confusing, but time consuming. I, for one, feel like the hapless Judge Willis, whose place in history was to be worsted frequently by the great F.E. Smith. “Mr Smith, I am none the wiser,” “No, m’lud, but vastly better informed!” Confucius said “A picture is worth a thousand words” and was making the point that, when communicating, it is essential to effect an immediate, yet lasting impact. I find it far easier to assimilate graphics, rather than columns of figures or pages of text. “The big profits come from the big pictures”. Mark Tennyson - d’Eyncourt

Investment Data Services

It is rather interesting running through the chart’s development since September ‘86. The story continued as follows:Another low spot on 272p (ONE). Then no top below 313p. The pattern became mixed with lower tops, picked by the - 1 x 1 angle drawn from the 360p peak, and higher lows. The squeeze was on between the + 1 x 1 and the - 1 x 1 angles. This reached the apex (TWO), and duly resolved on the upside. The bullish higher tops and bottoms pattern was reestablished and, with the new lows (THREE) on the bullish 1 x 1 angle, so was a minimum upward momentum. A subsequent low formed above 360p giving free rein to C&W to participate in the general market explosion. Interestingly enough this low above 360p, on June 29th 1987, was the spur Cable & Wireless needed to have its first major run of outperformance (vs. the FTSE 100) since November 1984. The crash broke most trends, patterns etc. but Cable & Wireless bottomed out (FOUR) on the combination of the old 272p horizontal support level, and the - 1x1 angle drawn from the 360p high. Thus what looked to have become a meaningless line into space after the late ‘86 up-break proved it still had some relevance. The subsequent rally topped out (FIVE) on the same pivotal + 1 x 1 angle which continues to dominate trading in the stock. The resistance at 387p was compounded by the fact that a halfretracement of the collapse had then occurred, and this level is 25% down from the 517p high. Currently a 2x1 angle from the post-crash low is attempting to create the momentum necessary to move the stock back above 360p, 387p, and the old 1 x 1, but you would have to rate its chances as only reasonable. The fundamentals are fine, if too closely linked to the fortunes of the US dollar, and the chart is not at a critical break-up point. Our recommendation therefore is to hold while the trading trend remains intact, but we can see little point in institutions or private investors being heavily overweight in the stock - its outperformance potential lies above the pivotal 1 x 1.

Andrew Rose

83


84

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Complex Patterns Josh Bruzzi BSc

Article originally featured in Market Technician 26 (July 1996)

How to deal with broadening formations (using the FTSE-100 as a case study) Triangular patterns are an easier complex pattern to deal with for market technicians than broadening formations. Triangles show the process of indecision in the market whereby the two trends (identifiable as two converging trendlines) complete the “battle” between buyers and sellers. An upside break of the resistance trendline is a clear indication that the bull camp is more powerful. This breaks the reassuring trend of successively lower highs for the bears and vindicates the trend of higher lows for the bulls, giving the latter more to go for. The bears are left looking confused and are likely to step aside from the market. Hence price action is likely to push higher with the triangle pattern itself giving clues to the likely validity, size and possibly timing of the resultant move. Should a break-out occur more than two-thirds of the way through the pattern (from likely inception to crossover of the converging trendlines), the break-out takes on less significance, possibly because the indecision is too compelling to create an earlier, and thus strong, signal of dominance either way. This is to say, if you see a triangle pattern converge close to its intersection point, be wary of the signal either way. The triangle also provides a target: the size of the initial discrepancy of opinion, i.e, the price differential from the two trendlines added to the level at the breakout (just like a head and shoulders pattern, which could be a form of triangle pattern anyway - see figure 1).

higher highs and lower lows, again suggesting confusion but in a less disciplined way. They are often symptomatic of a drifting market that is easy to push to fresh highs and lows but is then pulled back into the market’s perceived equilibrium price in the middle of the range if it strays too far either way. Finally one side (that of the wider trend) galvanises a new swing in the same direction. A good starting point is to look at the direction from which these patterns were entered and seek opportunities to buy towards the bottom of broadening patterns if the major trend is up and vice versa. The number of times a broadening pattern is likely to make new highs and lows is difficult to estimate because of variation and interpretration but of the ones I have seen, most turn after between four and six fresh highs and lows. Figure 2 shows stylised versions of some of the many possible broadening patterns that conform to Elliott Wave theory. Figure 2a shows a standard a-b-c correction in a market that is strong enough for the minor b-wave to reach a higher level than the upward impulse wave. The second is similar but shows a longer term corrective phase with identifiable minor waves. The last is a stylised wave four complex pattern with an irregular b-wave forming the core of the broadening formation with two (a and c) 5-wave corrective swings either side.

Figure 2 Figure 1

Simple

Implied Target

Simple

It also provides a time target for some kind of turning point, this being the time of the intersection of the two trendlines. Broadening patterns provide none of these clues and are thus much harder to trade. But the common theme is that they are complex patterns and are thus likely to be an indecisive phase in the main trend with continuation in the same direction as that prior to the formation of the pattern. The theme of indecision is the same, but whereas triangles could be seen as a “battle” between two forces to determine precedence broadening patterns reach

Irregular (Wave 4)


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

Figure 3

An initial examination of recent action in the FTSE-100 shows the two recent broadening patterns are consolidation/continuation patterns and comply with possible Elliott Wave four patterns, the wider implications of which will be covered later. A look at figure 5 shows another type of broadening formation, a possible wave two in the bigger picture of the bull market. The broadening ab-c corrective phase from 1991-1992 predicted a strong market environment with the corrective b-wave able to make fresh all-time highs above the likely wave one of the move from the 1987-91 triangle pattern breakout (note how the intersection of the big triangle pattern was almost a perfect turning point signal). Thus we now have two rules Figure 4 of thumb; 1) that broadening formations are continuation patterns and should be treated as such. They present opportunities for buying/selling into a correction of the main direction (the trend is your friend - however unlikely it may seem at the time). 2) that broadening patterns formed by a-b-c corrections, are a very bullish sign, if they can be recognised, while complex Elliott Wave broadening formations are continuation patterns (wave four) that signal the top is close. Complex continuation patterns are not necessarily Elliott Wave four patterns, but if correctly identified as such, warn of an impending top for that move. The FTSE-100 broadening pattern of late autumn 1995 adheres almost exactly to the stylised pattern in figure 2c, except that the last 5-wave sequence showed signs of strength by failing to break

lower than the preceeding a and complex b waves. Indeed, together with the start of the next impulsive up-move, it formed an inverse head and shoulders to create a buy signal that predicted the next rally. The second, most recently, was even trickier, generating a possible sell signal on good volume by breaking an established support trendline. Here, the last leg of the broadening phase did break lower, warning us, that although still bullish, the bull market is beginning to weaken, but has not totally run out of steam yet. The double bottom shortly afterwards confirmed the market’s willingness to push higher. So two additional rules of thumb can be added: 3) Elliott wave analysts should watch closely once the corrective wave c has run its course for clues/buy signals if this phase is identifiable. Those who do not use Elliott Wave analysis, should note that, after four to six fresh highs and lows, the move is likely to have run its course. 4) be very wary of signals generated by these patterns that give contrarian signals, as being a false sell signal. (J.J. Murphy warns in “Technical Analysis of the Futures Markets” that broadening patterns are indeed likely to generate false signals before proving to be continuation patterns.) The above notwithstanding, few text books mention broadening patterns, those that do say they are continuation patterns as long as they take longer than a month to develop. The longer they take to develop, the stronger the signal. Here again it is useful to look at

85


86

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Figure 5

the FTSE-100 from 1991-92. For those with really good databases, see the broadening pattern formed by the the Dow Jones Industrial average from 1966 to 1982 for possibly one of the best clear-cut equity market buy signal of all time. This signal was easier with the psychologically important 1,000 level proving very strong resistance for 15 years or so. Visual clues clearly help - for both triangles/wedges and broadening patterns, a break of horizontal, big number resistance is an easier buy signal and suggests strength. A year-long corresponding triangle/wedge characterised action in the FTSE-100 and DJIA during 1994/95. The health of the US market was seen in the complex pattern formed before the upward impulsive move. The DJIA surged after getting past the 4,000 level, this was seen as clear support underpinning the market whereby the FTSE-100 triangle pattern exhibited lower highs during construction, revealing its relative weakness. The FTSE all time high of almost 3,500 acted as a possible strong resistance in the minds of investors, which duly underperformed against the US market even more during 1995/6. This is an example not only of the visual clues in the construction of complex patterns, but it also shows how the construction of complex continuation patterns can predict relative outperformance/ underperformance in related markets (i.e the above would have represented a buy the DJIA, sell the FTSE-100 spread play signal - a very profitable one to date and also possibly contrarian to perceived wisdom at the time). Back to the FTSE-100: although the two recent broadening patterns signal continuation, the visual clue is the ability of the most recent move (Feb - March 1996) to make aggressive fresh lows on good volume towards its completion. Unlike the first, this is

a possible sign of weakness to come. As for the Elliott Wave count, if wave fours are always complex patterns, how can we have seen two recently? The possible answer is that while during 1993 we saw an impulsive wave five, (see figure 5) we have recently seen an impulsive wave three with the two broadening patterns representing wave four of this move and another wave four once it had finished. (See figure 3). Again, if my wave count is to be accepted, and only time will tell, we are very close to a major top with a likely non-impulsive wave five of the big move since late 1990 soon to be underway (for a possible size reference see the wave five of the first impulsive move during the summer of 1991 of roughly 270 points). Also, if the above analysis proves correct, by the time this article is published we should have seen a triangle pattern formed, broken out of on the upside and the FTSE-100 hitting new all-time highs, though probably on unexciting volume. If the possible Elliott Wave count is put aside, the visual clues for any market and inter-market strength/weakness) and other points made above should remain valid. Should it prove correct, the next series of complex continuation patterns in this market are likely to be within an extensive retracement of the bull market since 1991, perhaps the beginning of something bigger should the upward channel in figure 5 fail to provide sufficient support. Thus the buying clues above will be reversed to find selling clues.


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

Statistical Evaluation of Classic Chart Patterns Gregor Bauer

Article originally featured in Market Technician 69 (April 2011)

Introduction Classical chart patterns (such as triangles and head-shoulders) are a cornerstone of technical analysis, but do they really show the expected performance? Do stock prices, when forming rising triangles as a consolidation within an uptrend, really break out upwards - thereby confirming the trend - or do they more often act as reversal patterns? And what about the performance after the breakout? The former US software engineer and nowadays professional trader Thomas N Bulkowski has conducted the most comprehensive statistical evaluation of classic chart patterns to date. This article is based on three of his books and focuses on some of the key facts which every trader must know when trading classic chart patterns.

is a downward breakout; Head-and-shoulders bottom: 95% of these patterns result in an upward breakout. So both at the top and bottom of markets, this pattern is significant in identifying a reversing trend.

Evaluation of Classic Chart Patterns

The most important question for traders and portfolio managers is, does buying stocks that breakout from a chart pattern generate excess returns compared to the S&P 500? If the answer is negative, clearly this kind of pattern-based stock picking would be a useless exercise.

Methodology: In his book Trading Classic Chart Patterns (published by Wiley & Sons, 2002), Bulkowski measured the performance of stocks from the S&P500 Index after the breakout of specific chart patterns. The breakout price is defined as the highest price of a pattern or the break of a trend line. The performance is measured from the breakout price to the “ultimate high”, which is defined as the highest peak before the price declines by at least 20%. (For more detailed information about the methodology, see the website www.thepatternsite.com.) Here are the findings: 1. Triangles A rising triangle breaks out upwards about 70% of the time - and this is the surprise - regardless of trend direction leading up to the pattern. That means, even if the trend prior to the pattern is downwards, 70% of the time the price breaks out higher, thus resulting in a reversal pattern. In the case of a falling triangle, in about 55% of cases the price breaks out downwards, and for a symmetrical triangle it is closer to 50% (again, the price can be moving in any direction leading up to the chart pattern). 2. Rectangles Rectangles should, in theory, be a trend-confirming pattern. Rectangle in an uptrend: 68% of the time there is an upward breakout (i.e. trend-following after the consolidation). Rectangle in a downtrend: in 56% of cases the price breaks downwards - again trend-following. But, vice versa, 44% of the time rectangles in a downtrend act as reversal patterns. 3. Head and shoulders Head-and-shoulders tops and bottoms are supposed to be reversal patterns, and the evidence shows that that is exactly what they are. Head-and-shoulders top: 93% of the time there

Can trading chart pattern generate Alpha?

But, as can be seen in Table 1, the results are positive. It is possible to generate alpha from this method of stock selection. The table shows the findings in detail. Listed are the average performance results of the single stocks of the S&P 500 after the breakout of various patterns. Table 1: Best performing patterns for an upward breakout Pattern

Average Rise

Failure rate: Rise less than 5%

Failure rate: Rise less than 10%

Falling Triangle

42%

3%

9%

Double Bottom

27-37%

4%

18%

Rectangle

37%

10%

19%

Symmetrial Triangle

35%

5%

15%

Broadening Top/Bottom

34%

8%

19%

Head-Shoulder Bottom

34%

3%

9%

Rising Triangle Dreieck

34%

6%

17%

Source: Thomas N Bulkowski, Trading Classic Chart Patterns, Wiley, 2002

The comparison with the S&P500 index The average performance of the S&P 500 in the same period of time was only 3-6%.

87


88

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Example: the average breakout performance from a falling triangle with an upward breakout (several thousand examples were investigated) for single stocks is a 42% rise in price. This compares very favourably with the performance of the S&P 500 (an average rise of only 3-6%). The difference highlights the minor influence of the overall market (beta) and the real performance of the pattern (alpha).

Other interesting data are the failure rates Besides analysing the average rise after a breakout, measuring failure rates is important, because it gives investors valuable information about the minimum breakout performance they can expect. Table 1 shows the 5% and the 10% failure rates related to each pattern. Taking the example of the falling triangle, the columns: “Failure rate: Rise less than 5%” and “Failure Rate: Rise less than 10%” show, that in only 3% of the time, the breakout performance was less than 5%, and in only 9% of the time, the performance after the breakout was less than 10%. Vice versa, in 91% of the time investors can expect a performance of 10% and more after the upward breakout from a falling triangle. Table 2 shows the best performing pattern for a downward breakout. Table 2: Best performing pattern, downward breakout Pattern

Average Drop

Perform S&P 500

Failure rate: Drop less than 5%

Failure rate: Drop less than 10%

Head-Shoulder-Top

42%

+3%

6%

18%

Symmetrical Triangle

20%

-2%

5%

24%

Rising Triangle

19%

0%

6%

24%

Double-Top

15-18%

+1%

13%

32%

Falling Triangle

18%

-1%

7%

27%

Broadening Top/Bottom

17%

-2%

13%

33%

Source: Thomas N Bulkowski, Trading Classic Chart Patterns, Wiley, 2002

Here, the average performance of the S&P 500 is given with every pattern to show that, even at a time when the overall market climate is bullish, downward breakouts from a pattern can be traded successfully against the market direction. This is particularly true of the Head-and-Shoulders top. One might, of course, argue that markets have changed in recent years and, consequently, pattern analysis may not perform so well nowadays. A more recent evaluation was therefore carried out.

Did failure rates change in recent years? Methodology: Database: 1991-2008; 14,000 patterns evaluated in total.

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Patterns evaluated were: • • • • • • •

Diamond tops/bottoms Double tops/bottoms Head-and-shoulder tops/bottoms Rectangle tops/bottoms Rising/falling wedges Triangles: rising, falling, symmetrical Triple tops/bottoms

Figure 1 shows the results for upward breakout failure rates. The graph in fact shows that failure rates increased during the bull market period from 2003 to 2008, compared to the bull market period in the 1990s. Of course, during bear market periods, the failure rates for upward breakouts increase by definition. Breaking this down into more detail: • • • •

1991: 11% of all patterns of upward breakouts failed to show at least a 10% performance after the breakout (up to the ultimate high); 22% failed to show a 20% performance; 64% failed to show a 40% rise. From 1991-2008 there is an overall increase in failure rates. Comparing the bull markets: 1990s vs. 2003-2008: The average 10% failure rate climbed from 14% to 28%. Comparison of the relative highs of the failure rates In 2000 (beginning of the bear market): 60% of the breakouts failed to show 40% performance In 2008 (beginning of the bear market): 88% failed to show 40% performance.

Figure 2 shows the results for downward breakout failure rates. The graph shows that failure rates for downward breakouts during bull markets are in general higher than the failure rates for upward breakouts (as depicted in Figure 1). That is, of course, to be expected, but Figure 2 also shows the overall increase in failure rates when comparing the bullmarket periods, which is especially true for the 10% and the 20% failure rates. Breaking this down into more detail: Results: downward breakouts - failure rates • •

The average 10% failure rate in the 1990s bull market was 26%, but increased dramatically during the bull-market period from 2003-2007 (i.e. to 49%) The average 20% failure rate in the 1990s was 59%, but rose to 75% from 2003-2007.

At the changeover from bear to bull market in 2002/2003, the failure rate spiked, as expected.

Conclusion Trading classic charts pattern still work and traders can generate excess return over the market index by applying this strategy for stock picking. But we also have to be aware of the fact that, due to the worldwide increase in volatility, throwbacks and pullbacks (Bulkowski conducted several studies on this subject as well) will increase, as well as failure rates (as seen in figures 1 and 2), thus resulting in the necessity for even stricter money - and riskmanagement.

Performance measured for: • • •

Upward breakout - from closing price of the day before the breakout to the ultimate high. Downward breakout - from the closing price the day before the breakout to the ultimate low. Failure rates calculated for 10%, 20%, 40% performance after breakout.

Dr. Gregor Bauer works as independent portfolio manager, as lecturer for technical analysis and portfolio management at several universities and serves as president of the German Association of Technical Analysts (VTAD) and as a member of the board of the InternationalFederation of Technical Analysts (IFTA).


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

Figure 1: Failure rates for upward breakouts change during the period 1991-2008

Source: Thomas Bulkowski, www.thepatternsite.com

Figure 2: Failure rates for downward breakouts change during the years

Source: Thomas Bulkowski, www.thepatternsite.com

89


90

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Some thoughts on the Mathematics of Turning Points Tony Plummer FSTA

Article originally featured in Market Technician 34 (March 1999)

Introduction One of the most surprising, and fascinating, aspects of financial markets is that price movements appear to be bound by mathematics. In particular, they seem to be constrained by the influence of the Golden Number (0.618) and its derivatives. These numbers are now quite well known within the investment community and are widely used. First, the relationship between Unity, the Golden Number, and the various contractions of the Golden Number (i.e., 1, 0.764, 0.618, 0.382, 0.236 etc.) are used to calculate retracement support and resistance levels. Second, the relationship between Unity, the Golden Number, and expansions of the Golden Number (i.e., 1, 1.618, 2.618, 4.236 etc.) are used to calculate the potential end of trend highs or lows. However, apart from the problem of explaining why it is that these numbers work in the first place, there is a great deal of difficulty in determining when they are likely to work. First, there is often some difficulty in determining the specific price swing that should be used for calculating a level that might operate as a reversal. Second, a calculated number is usually used as only one criterion for a reversal point, from amongst a large number of other criteria for reversal points. The appropriate reversal levels are accordingly much easier to spot after the event than before the event. This Figure 1: FTSE100 in run-up to the 1987 crash

uncertainty tends to undermine the value of using the Golden Number. The purpose of this article is to explore some of the ways in which combinations of the Golden Number can be used to calculate potential turning points.

The 1987 Equity ‘Crash’ A useful starting point for the analysis is the traumatic equity event of the summer and autumn of 1987. The FTSE 100 share index peaked on 16th July 1987 at 2455.2. It then fell by 12.1%, in a clear (Elliott-type) five-wave pattern, to 2157.2 on 20th August, before rallying by 11.2% to 2399.9 on 5th October. The subsequent move is now part of stock market legend. The index fell by 36.9%, finally reaching an intra-day low of 1515.0 on 10th November. By themselves, of course, the various price swings in the 1987 trauma appear to be no more than the accidental outcome of unusual market conditions. However, once knowledge of the Golden Ratio is applied to the movements, an extraordinary set of non-random relationships emerges. The price swings actually become very meaningful. Some of the relationships are shown in Figure 1 below. The first important one is the relationship between the last downwave (the


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

5th wave) of the July-August sell-off and the subsequent rally. Between the 14th and 18th August, the FTSE100 index fell by 150.1 points. The subsequent upmove was 242.7 points. The rally was therefore almost exactly 1.618 times the 5th wave fall (i.e., 242.7/150.1 = 1.617). In my experience this is not unusual. Indeed, there is strong evidence that any impulsive wave will tend to have a relationship, which is based on the Golden Number, with the base pattern that preceded it.1 Second, it can be seen that the August - October re-test rally peaked at exactly the same level as the peak of wave 2 of the July sell-off. Further, at the re-test high on October 5th, the market registered a ‘one-day reversal’, closing at a low of 2385.8. Such reversals usually imply that the market has, in some sense, overshot natural levels and has returned to a more balanced level (albeit temporarily) at the close. Under these circumstances, the closing price becomes a valid one for conducting calculations. Importantly, the closing level on 5th October was almost exactly 0.764 of the July-August fall (i.e., 228.6/298 = 0.767). Third, the whole of the fall - from the 16th July peak to the 10th November low - amounted to 38.2% (i.e., [2455.2 - 1515.0]/2455.2 = 0.383). Since 0.382 = 1 - 0.618, this was a stunning confirmation of the influence of the Golden Number. This, in itself, could have been enough. However, the operation of the number 0.382 was a direct hint at the influence of one of Nature’s laws. This is that there are always constraints on the ability of a system - any system - to regress. In markets, if the evolving economic environment is consistent with a bull run, then a correction simply cannot start to discount a nonexistent deterioration in fundamentals. In financial markets, the constraint is 38.2% of the impulse move that preceded it.

The number 0.382 and the definition of corrections Let us just look briefly at the implications of these movements. Those who experienced the traumas of the 1987 equity fall may well remember that, at that time, economic activity in the UK was basically strong. Indeed, the authorities had felt obliged during the summer to tighten monetary policy in order to dampen spreading consumer enthusiasm. The important point, however, was that prior to the October ‘crash’ - there were no independent indications anywhere that a major economic calamity was unfolding. Indeed, quite the reverse. In this sense, the crash event was an anomaly2 and the biggest mistake that commentators made was to assume that the market was ‘forecasting’ a previously hidden crisis. In fact, the authorities eased monetary policy aggressively in the UK (as they did in the USA), and the markets subsequently recovered. However, the point is that, at the low of the fall in October 1987, the market came up against its own internal constraints. At a level of 1515.0, the market had retraced 38.2% of the rally since it was theoretically valued at zero. It could not fall beyond this level unless the underlying economic situation was destined to involve an assault on the nature of capitalism. There are two important implications of this conclusion. The first is that the number 0.382 provides the basis for distinguishing between a correction and reversal. The market might oscillate around the 0.382 retracement, while it ‘waits’ for fundamentals, but a definitive break has definite implications. Feedback develops between fundamentals and prices, so that news and prices move together for the main part of the subsequent impulse wave. This is, for example, exactly what occurred in Hong Kong when the Hang Seng Index broke 10,400 on the downside. In this way, 0.382 provides the ‘stop-loss’ for a trend. Up to this point, the contra-trend movement is a technical correction; after that point, it becomes a fundamental reversal.

Psychology and Markets

Systematic Trading

The second implication is that market movements cannot be divorced from the passage of time. Obviously, a huge price movement in a short period of time has a different meaning to investors than a similar price movement over a long period of time. Fundamentals are more likely to be involved in the latter than with the former. The important point, however, is that there is always a hierarchy of fundamentals, equivalent to the difference (for example) between stock building and innovation, or between cycles and evolution. Stock building and simple cycles relate to quantitative change, while innovation and evolution involve qualitative change. Often it is impossible for observers to distinguish between the two concepts - partly because the qualitative changes are so subtle, and partly because quantitative and qualitative changes occur simultaneously. But the market itself will know. It is the market that will decide the validity of support and resistance levels from within its own context. The basic rule is that the longer it takes to penetrate a 0.382 level, the greater will be the likely subsequent impact on prices.

The Golden Number 0.618 and its implication for reversal patterns In 1935, H.M. Gartley in the USA, drew attention to the potential profitability of trading against re-test rallies or re-test falls.3 His specific rule was that a trader should sell a rally after it had retraced 0.618 of the initial fall (or buy a fall after it had retraced 0.618 of an initial rally), and place a stop at the high (or low).4 If the rule were applied rigorously, then the trader would be on the right side of almost all of the major impulse waves and would only suffer small drawdowns on the trades that went wrong. Following this rule on the FTSE100 index in late August/early October 1987, for example, would have reaped dividends almost beyond the dreams of avarice. The relevant point here is that Mr. Gartley had stumbled on (or deduced) a fundamental principle concerning market behaviour. This is that a market will tend to retrace about 62% of an initial price reversal before developing into an impulse wave. I do not know very much about Mr. Gartley, but he seems to have spotted this phenomenon at the same time as - and possibly independently of - R. N. Elliott. It is important, however, to note an element of pliability in Mr. Gartley’s rule. Observation will confirm that, although the 61.8% retracement is important, it is not always absolutely precise. To some extent, since the market is going through a transition phase, it will be more vulnerable to short-term factors. These may cause either an overshoot (and have the trader reaching for a stiff brandy) or an undershoot (in which case, the trade may be missed completely). Hence in 1987, the actual retracement level that was achieved by the FTSE was 76.4% of the initial price fall. However, as already indicated, this level did not just arise in isolation. It also combined with a 1.618 expansion target and the top of the second wave of the initial sell-off. The more calculations that coincide with a particular level, the more likely it is that this will be the actual objective. So there were important independent reasons for assuming that an index level of around 2400 would form the centre of gravity. The flexibility of the 0.618 retracement target does not therefore invalidate the essence of the Gartley rule. However, it is possible to fine tune it by looking at other near-term influences.

The three-wave blueprint The 0.618 rule - which I would definitely recommend to the committed trader - depends for its validity on a specific concept about market behaviour. This is that financial price movements5 conform to a three-wave blueprint. In my opinion, this three-

91


92

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

wave blueprint underlies all market behaviour. It consists of: an information shock (which may be an unexpected price reversal), a partial retracement to form a base pattern, and then an impulse wave. See Figure 2. For the greater part of the impulse wave, prices travel along the path of least resistance. The market has ‘learnt’ that it is in trend (up or down) and the flow of news will largely be consistent with (and confirm) that trend. However, there comes a point where all this news is ‘in the market’ and the trend will be overextended. An energy gap develops within the market, such that it becomes acutely sensitive to qualitative information shocks. Hence, the trend will either undergo a polarity switch and reverse, or it will experience a pro-trend shock and extend.

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

new three-wave structure concatenated onto the initial (or old) three-wave structure. This results in the appearance of a fivewave movement. See Figure 3. This five-wave movement - first discovered by Mr. Elliott - is the hallmark of evolution within financial markets. Importantly, however, the mathematics of the Golden Number imply that the retracement after the pro-trend information shock (i.e., the 4th wave, in Elliott’s terminology) should be consistent with the 0.618 principle of retracements. That is, the 4th wave should retrace 0.618 of the market move since the moment of the information shock. Figure 3: The five-wave pattern

As it happens, market crowds are not very good at picking up information that relates to qualitative changes in the forces that influence market trends. This is because, by definition, revolutionary changes are not consistent with the current paradigm (or belief system) embedded in the market.6 If these qualitative changes are in fact occurring in the background of a trend, then there will inevitably be a stage when the relevant information about these changes suddenly impacts as a shock to the majority of market participants. On the one hand, this information could be consistent with a market reversal. On the other hand, however, it could be consistent with the trend. It is the latter that we are concerned with here. Figure 2: The three-wave blueprint

This is shown in Figure 4 below. In the diagram (which relates to a bull market) SP is the shock point, a is the top of the resulting bear squeeze, and is the low of the subsequent setback. Then,

b

b = a - 0.618(a - SP) So that

SP =

b - 0.382 a)/0.618

This implies that, once a correction has occurred, and the trend appears to have re-asserted itself, then it should be possible to identify the exact moment in price and time when the shock occurred.9

Bridging energy gaps with pro-trend shocks Hence, when a pro-trend shock occurs, the energy gap will be bridged. The trend will extend itself on the back of increased enthusiasm by market participants. This extension may arise either when the market is fully stretched, so that more good news is literally needed; or it may occur after a small temporary setback, so that some free financial resources are already available to be utilised.7 Another way of describing this is to say that more financial energy (which was not readily available under the old matrix of market knowledge) is made available, because extra investments can now be justified. Therefore the trend extends in accordance with the new matrix of knowledge. In any case, the identifying feature of the shock point will be some combination of accelerating price changes, intra-period price gaps and higher trading volumes.8

Bridging energy gaps with pro-trend shocks In theory, the pro-trend information shock should create a

Figure 4: The 0.618 retracement


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

The importance of the shock point The assumption here, of course, is that it is not possible to isolate the shock point until after both the ensuing rally and associated retracement have occurred. At first glance, therefore, the formula may seem to be of no more than academic interest. However, if it is possible to identify the shock point, two conclusions follow. First, it should be possible to analyse the information that occurred at the shock point to assess and confirm its importance. Second, if the information is seen as constituting a genuine shock, then this implies that the subsequent retracement relates directly to that shock. This, in turn, suggests that it is possible to confirm the moment in real time when the 0.618 retracement is probably over.10 Such information gives extra confidence when committing to a trade. If we apply this principle to the recent correction in the Dow, which lasted from 20th July 1998 to 1st September 1998, it is possible to identify the shock point level as being 6177.0.11, 12 This level was last seen on 7th/8th November 1996. See Figure 5. In fact, looking at an appropriate daily chart will show that the actual day of the shock was almost certainly Wednesday 6th November, when the market opened at just over 6074 and then jumped by almost 1-3/4%, to close a fraction below the day’s high of 6178.08. The reason? The re-election on the previous day of a Democratic President (Bill Clinton) together with a Republican-led Congress. The resulting stand-off was taken to mean no radical change in economic or social policy and therefore no significant increases in government spending. Off such news was an extension to the bull market born. Figure 5: The shock point in the Dow

Psychology and Markets

Systematic Trading

the high. This places the shock point between 3670 and 3790. In fact, the technical break away from the correction occurred on the 2nd August close, at 3770.7. See Figure 6 again. This means that once the market started to correct from the 2nd October high at 5367.3, it would be aiming to correct 0.618 of 5367.3 - 3770.7 (i.e., 1596.6) points. The calculated likely fall was therefore 986.7 points, to 4380.6. The actual low on 28th October was 4382.8. Figure 6: The 0.618 target on the FTSE100 index

These findings are extraordinary. They mean (at least) four things. First, they confirm absolutely that the 0.618 retracement should be calculated by reference to the shock point. Second, as a corollary, they suggest that this calculation is more accurate than the traditional one using the base of the wave that is being corrected, unless this base actually qualifies as a shock point in its own right.13 Third, the findings highlight the fact that the shock point may well be found quite early on in an impulse wave, rather than in the middle. Finally, of course, this analysis clearly explains why significant setbacks can occur apparently out of nowhere.

The importance of the shock point

(Stock market aficionados might also like to register the fact that the shock point relevant to the 1987 crash in US equities was 924.0 on the Dow. That is, (1616.21 - (0.382 x 2736.60))/0.618= 923.67. Unfortunately, I cannot now remember the reason for the shock, but it seems to have occurred on 6th/7th October 1982, right at the beginning of the acceleration that took the Dow through 1000.0. This latter level had effectively been the ceiling on the Dow since January 1966.) It may be, of course, that we already know the precise location of the shock point. In this case, we can immediately calculate the likely extent of the 0.618 retracement once the market has begun to reverse. Let us look at an example. See Figure 6 below. On 16th July 1996, the UK FTSE 100 index ended a correction and began an advance that effectively lasted until 2nd October 1997. The overall rally lasted 1754.7 index points, or 48.6%. The market then turned down into the mini-crash of October 1997. Looking back to the beginning of the bull move, it is apparent that the information shock - as confirmed by accelerating prices, after the 16th July correction low - occurred somewhere between 31st July and 5th August. During those 4 days the market rallied by 3.35%, opening each day at or near the low and closing at or near

The retracement that constitutes the 4th wave provides the platform for the 5th - and final - wave in any particular impulse movement. Very often it is possible to refine the analysis of 4th wave targets because of the presence of clear objectives for the whole trend move. These objectives may, for example, be previous turning points (see Figure 7a), or they may be longer-term targets established by using Golden Number expansions (see Figure 7b). Where this is the case then the base of the retracement (= ) is related to the beginning of the correction/end of the impulse wave (= ) and the probable terminal juncture (= TJ) by the formula:14

b

a

b = (1.618a - TJ)/0.618 This formula may provide a target that amends or confirms the 0.618 target calculated from information shocks. Figure 7a:

Figure 7b:

93


94

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Calculating the last wave

Fine-tuning the Gartley rule

After the end of the 5th wave, there will always be a correction that is at least as large in time and price as the two corrections that preceded it. This follows because, as a result of the earlier pro-trend shock, the correction is emerging at a higher level in the price-time hierarchy. Refer back to Figure 3 again.

The idea of incorporating known objectives for the terminal juncture (= TJ) into the calculations gives us another method of fine-tuning the Gartley rule. This calculation assumes that the impulse wave following the retracement will move to the base of the correction of lower degree (i.e., to the base of Elliott’s 4th wave). See Chart 10.

Because of the mathematics of price movements, it is highly likely that the 5th wave will be related to the 4th wave correction either by the appropriate version of the Golden Number (i.e., 1.618) or by one of its expansions (2.618, 4.236, etc). However, because the 5th wave is usually somewhat speculative in nature, it tends to materialise on weak volumes, falling momentum and even (sometimes) falling open interest in the futures markets. It is not unrealistic therefore to expect the 5th wave to be shorter than the 3rd wave. This suggests that, very often, the appropriate expansion to use will in fact be 1.618. See Figure 8. Figure 8: Targeting the 5th wave

If we assume that the bottom of this impulse wave (= TJ) will be a Golden Ratio expansion of the retracement, then the top of that retracement (= ) will be defined as:16

b

b = (1.618a - TJ)/0.618 Or

b = (2.618a - TJ)/1.618 The former assumes that the Golden Ratio expansion is 1.618:1, the latter that it is 2.618:1. It should be easy to see, just by looking at the chart, which of the two is applicable. Figure 10: Targeting the retracement

However, the foregoing analysis suggests another way of calculating the level for a potential top, once the 5th wave has begun. This is to calculate a level for the 5th wave at which the following criteria are met. First, the subsequent correction would return to the base of the 4th wave. And, second, this correction constitutes 0.382 of the whole rally, measured from the base of the 1st wave to the peak of the (as yet not completed) 5th wave. This is shown in Chart 9 below, where the start of the move (or shock point) is SP, the terminal juncture is TJ, and is the base of wave 4. If lies at a point that is 0.382 of the distance between SP to TJ, then15

b

b

TJ = ( - 0.382SP)/0.618

b

Figure 9: Targeting the 5th wave


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

Conclusion It is my opinion that financial markets are subject to an underlying order that is reflected in the operation of mathematical laws. These laws work because the integrative influence of large numbers of people focusing on one thing - market price movements - creates order out of chaos.17 The fact that we are often unable to deduce the relevant laws, and/or are unable to forecast market trends with sufficient accuracy, does not alter the underlying reality. It is certainly my personal experience that the 0.382/0.618 retracements and the 1.618/2.618 expansions work with a high degree of precision. Moreover, they work so often that it is virtually inconceivable that the phenomenon is purely accidental. The use of the 38.2% retracement calculation to isolate the boundary between a technical correction and a fundamental reversal may help to give more rigour to an otherwise never-ending debate. It is by no means the only criterion, but it is an important one. Furthermore, it determines the moment in the analysis when it becomes appropriate to pay more attention to so-called ‘fundamentals’. The use of the 61.8% retracement calculation to isolate the end of a counter-trend correction is intimately related to an earlier shock point, where market psychology changed in some important way. It can be argued either that the shock point can be deduced from the correction, or that the correction can be calculated from the shock point. In either case, our knowledge of the market is enhanced in a relevant way, and investment decisions can thereby be improved. The use of 1.618 and 2.618 expansion targets remains, as ever, an important tool in the decision-making process. Their use in conjunction with the numbers 0.382 and 0.618, as described herewith, greatly enhances their power. Trading in financial markets has never been easy and requires a great deal of work. The mathematical formulae outlined in this article may be seen as additions to this work; but the results are well worth it.

I

Tony Plummer, Forecasting Financial Markets. London, Kogan Page, 1998. The same comment also applies to down moves away from a top pattern.

2

Such anomalies are important items of information and can be the basis of very profitable trading in financial markets. See Tony Plummer, op.cit.

3

H.M. Gartley, Profits in the Stock Market. Reprint: Lambert-Gann, 1981.

4

If the market returns to the old high, it is either already in a 5th wave, or in the B-wave of a complex correction. In either case, the longer-term trend is intact.

5

Actually, the psychology of the financial market crowd.

6

This includes forecasting points of inflexion in the business cycle. No qualitative changes are necessarily involved in these forecasts.

7

This comment amends my analysis in earlier articles. See, for example, Information Shocks and Energy Gaps in Market Technician, March 1998. There the only case considered was the situation where a market is already fully stretched.

8

Note a sharp change in price that is not ‘confirmed’ by the closing price does not necessarily qualify. A closing high confirms a rally and a closing low confirms a fall, but a close mid-way in the trading range of the day is inconclusive.

9

The formula for the shock point is exactly the same for a bear market, provided that a is taken as being the end of the impulse wave down and b is the end of the counter-trend rally.

10

In other words, it is possible to identify the point where the Elliott 4th wave correction is probably complete.

11

The calculation is: (7400.3 – (0.382 x 9380.2))/0.618 = 6176.5.

12

The accuracy of this calculation is confirmed by the fall between 7th August 1997 and 28th October 1997. The shock point for this fall is: (6971.32 – (0.382 x 8299.03))/0.618 = 6150.63. The surprising implication is that, even though this particular correction was associated with the crash in Asian equities, the extent of the fall in the US was, in some sense, predetermined.

13

Which may be the case if there is a one-day reversal, or an outside day, at the terminal juncture of the previous move.

14

Again, the formula is exactly the same for a bear move, provided that a is taken as the end of the impulse wave and b is taken as the top of the countertrend correction.

15

The same formula may be used for a bear phase, provided that a is taken as the end of the impulse wave and b is taken as the end of a countertrend rally.

16

These formulae apply to rallies following a five wave fall, provided that a is taken as being the end of the first impulse wave upwards and b is taken as being the end of the base pattern.

17

There is an important corollary to the existence of mathematical relationships between price swings in financial markets. This is that people do not make their decisions independently of one another. Quite apart from the contentious issue of free will, this phenomenon directly confronts one of the traditional assumptions of economic theory.

95


96

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Some Reflections on Stock Market Interrelationships Anne Whitby

Article originally featured in Market Technician 15 (November 1992)

Much has been written on the interrelationship between stocks, bonds and commodities, notably by John Murphy whose studies are of immense value to technicians. This piece is not intended to retread such ably covered ground, but rather presents a series of thoughts on the way global stock markets interrelate, or sometimes diverge. Bound up inextricably with this is the movement of currencies. Over the last twenty years stock market cycles around the world have generally moved closer together. This can be explained by many factors, the relative importance of which would be a long, and probably inconclusive, discussion. A list of these would include the rapid growth in the speed of communications, enabling traders and investors around the world to know instantly what is happening in other markets. The growth in the economies of continental European and Far Eastern countries, filtering through into their stock markets, meant a challenge to the previously dominant US and UK exchanges, with a particularly strong effect on the latter. Capital movements round the world have become easier because of the growth of international banking groups, the dismantling of exchange controls and fixed parities and again improved communications. The increased pressure on fund managers to ‘perform’ has meant that lagging markets have been sought out. Such markets have tended to be smaller ones, so they have often caught up quickly and risen even further than the leaders in the end. In addition, and tied up with all the other reasons, world economies, interest rates, inflation levels and growth rates have been pulled closer together, not necessarily in terms of absolute numbers but in their direction of movement. This list may well be incomplete, but it offers some reasons for the increasing commonality of stock markets. It was during the 1980s, when the world industrial economies all experienced a strong growth phase, that this convergence became particularly marked. This led to the 1987 crash, when everyone except South Korea, and possibly a few other emerging markets, imploded in a way few of us are likely to forget. In the memorable words of Tom Lehrer, “we’ll all burn together when we burn.” Some interesting short-term divergences emerged after the crash, notably the subsequent recovery being led by the Far Eastern markets and, in Europe, by Belgium, Holland, Spain and Sweden. With the underlying economic picture intact the rest eventually followed upwards, but the most mature markets of the US and UK were slow to overcome their distrust of the leadership being shown. Probably as a result of its leadership, people watched Japan avidly for some time after that, the first question of the morning being “what did Japan do today?” Then the news items would start rolling - “Following Japan’s 500 point rise (fall), the market opened strong (weak)”. This seems to have faded in recent times, which leads on to the way in which market interrelationships have changed since the late 1980s, or not. Not surprisingly the cohesion of the western European markets, excluding Finland (a major disaster), Norway and Sweden, has held fairly well. The influence of the EC and the associated ERM have been important, but the similarity becomes more striking

when you consider the relative economic position of, say, the UK and Germany. While the actual degree of movement has varied, the direction of the swings has been remarkably close. It will be enlightening to see over the coming months whether this interrelationship continues to hold, as a number of these markets begin a weaker phase. It should. Interestingly, many of the smaller Far Eastern markets, plus Australia and Canada, are following the same basic shape, reinforcing the overall commonality theory. However the riveting divergence between the US market, Hong Kong to an even greater extent, and the Tokyo exchange may be seen as an alarming threat to the whole argument. To take Hong Kong first, it would appear that for the time being, given the Chinese factor, we are dealing with a wild card. In fact, apart from the Tiananmen Square decline and the present dizzy rise it has not deviated greatly from the overall pattern of swings. Another 5000 points on its own would mark a more serious divergence! Even the US is not too far adrift from other markets in terms of swing direction. Most saw troughs in 1987/88, peaks in 1989/90 and troughs in 1990, a pattern also apparent in Wall Street. Even though it is at present relatively higher than most other markets, the timing of its directional movements has not been dissimilar. Japan, however, remains the glaring anomaly, having failed to participate in any sort of real upswing since it saw a good, but short-lived, rally from the October 1990 low. (But note the convergence of that low with other markets). A number of reasons can be propounded to explain this, not least the particular situation of the Tokkin funds, but also a horrendous property crash and the financial and political scandals, plus the country’s non-involvement in the Gulf War, which have led to a certain introspection. Any of these problems can be seen in other markets, but not all together. Given the factors discussed earlier, which are of longer-term import, it seems reasonable to suppose that, one way or the other, Japan will eventually get back into line. Currencies were mentioned at the beginning of this article as being inextricably bound up with the markets, but their interrelationship with markets is on a completely different plane. The most important point to bear in mind is that a rising currency does not mean a rising market, but nor does a falling currency mean a rising market. Any market at any time may, or may not, fall or rise in line with its currency. In this situation, we cannot find any useful, consistent interrelationship. This means that whenever we make a decision to invest in a non-domestic market it is essential to take a view on the currency. There is very little point investing a large sum from, say, the US into Germany if the Mark drops significantly against the US Dollar over the period of the investment. That is, unless this risk has been recognised and suitable defensive action taken in the form of a currency hedge.

Anne Whitby is the managing director of Chart Analysis Ltd.


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

97


98

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

CHAPTER FIVE

Moving Averages and Trends Articles in this chapter 100 Crossover Methods in Moving Averages Steven Kwan

102 Improved Moving Average (IMA) Strategies Fotis Papailias and Dimitrios D. Thomakos

108 The Trend Intensity Indicator Richard Lie

110

Least Squares Regression Jeremy du Plessis

113

Volstall: Using volatility to identify early signs of Trend Exhaustion Oliver Woolf


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

CHAPTER FIVE INTRODUCTION

By Axel Rudolph FSTA

Introduction A few months after finishing my Masters degree in Economics, I started trading Fixed Income and FX in a major French bank in Paris in September 1992, the month in which George Soros and other speculators forced the British government to pull the pound from the European Exchange Mechanism (ERM). Needless to say that in this high volatility environment my recently acquired theoretical, fundamental knowledge was of little use to me. Luckily a colleague showed me his green and black monitor on which I could see charts with trends and moving averages. The moving averages smoothed out the price action and helped me get a clearer picture of what trends I was actually trading, from the mostly short-term, intraday, to the mediumterm, in the case of FX trading back then, perhaps two to three days. Moving averages and trends have probably been looked at as far back as the 18th century when traders analysed the corn markets of England and the rice markets of Japan. In the US, back in 1882 Charles Henry Dow went into partnership with Edward Jones to form Dow, Jones & Company, providing financial news and stock prices. A year later they started a regular daily financial newsletter which by 1885 became “The Wall Street Journal” in which charts and trends were mentioned alongside fundamental news. My dear friend Akira Homma from the Nippon Technical Analysts Association (NTAA), a founder member in 1986 of the International Federation of Technical Analysts (IFTA), found references to moving averages as far back as 1913 in a book by Rakusen Iwatani entitled “Kimai Kansoku Rakusen Hiroku” which literally means “Rice Futures Observations - Rakusen’s Secret Records.” Of course the methods by which moving averages and trends have been traded since then have come a long way, with these being calculated and used in a multitude of ways. Trend analysis and moving averages, as a topic, have attracted some insightful research, as some of our chosen articles will show. We now take a brief look at the contributions to this chapter, beginning with Crossover Methods in Moving Averages, written by Steven Kwan during his time as head of technical research at MMS Standard & Poor’s. He explains both construction and application of the three most commonly used moving average types, simple, exponential and linearly weighted, and the reader comes away armed with a clear understanding of ‘the basics.’ Steven’s contribution is followed by Improved Moving Average (IMA) Strategies. The authors, Fotis Papailias and Dimitrios Thomakos, are experts in econometrics,

Psychology and Markets

Systematic Trading

hold senior positions at universities, and are also responsible for research at www.quantf.com. In their May 2012 contribution to the STA journal they shared with us the core ideas of two of their research papers, giving us a comprehensive explanation of a system designed for the ‘moving average dedicated investor.’ This detailed contribution is well worth careful study. Academic research plays an important part in the advancement of technical analysis. The role of sentiment in trend development and trend strength is examined by Richard Lie, who offers us The Trend Intensity Indicator, ahead of a mathematical perspective of trend provided by Jeremy du Plessis. In his article Least Squares Regression - Fitting Data to a straight Line, Jeremy reminds us that a regression line is all well and good, but that technical analysts like to know if ‘a trend is still intact.’ So he suggests that placing parallel lines a certain number of standard deviations away from the least squares regression line will be useful in determining trend as we like to see it. This article, dating back to the 1990s, seems to have been quite prescient: the use of regression lines in technical analysis looks to have grown, no doubt helped by increasingly sophisticated charting software since that time. The issues of the trend eventually ceasing to be our friend, and even the MACD not always being our friend, also have a place in this section. This is examined in Oliver Woolf’s ‘Volstall’ (volatility stall), which builds on the Bollinger Band technique. We would ideally conclude with suggesting some specialist book titles for those wishing to read further. Surprisingly, perhaps, there seems to be a dearth of authoritative titles dealing exclusively with moving averages. Perhaps because the tool is, in essence, quite a straightforward one? However, the generalist books on our STA Diploma reading list cover the subject extremely well. These are: Technical Analysis, The complete Resource for Market Technicians by Kirkpatrick & Dahlquist, Technical Analysis explained by Martin Pring and, of course, Technical Analysis of the Financial Markets by John Murphy.

99


100

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Crossover Methods in Moving Averages Steven Kwan MSTA

Article originally featured in Market Technician 33 (October 1998)

There are already a lot of articles and research papers that have been published over the years about the subject of Moving Averages (MA). Given that it is one of the less ’subjective‘ tools in modern technical analysis, it is therefore not surprising to see researchers using MA as a ’synonym‘ of technical analysis when carrying out research on the profitability or the usefulness of this ’non-scientific‘ approach to the markets. We have to remember MA is only a small part of technical analysis. The idea behind MA is not too complicated. The Simple Moving Average (SMA) is simply the average of prices in a given time window e.g. a 21-day SMA shows the average price for the past 21-days. The time period chosen would indirectly imply the sensitivity of this average number to the present market condition. For example, a 5-day SMA is much more responsive to the current trading environment than a 10-day SMA and the 4-day SMA response is even quicker than the 5-day, and so on. Therefore, the longer the period, the more ‘lagging’ it is. The data that are being averaged can be any of the four key elements in charting, the open, high, low or close although the latter is the most popular one. Given the differing sensitivities of different MA periods, a crossover would be potentially one of the most useful ways to gauge the underlying strength or weakness of prices i.e. to check whether current trading prices are above or below an ‘average’ price for a specific period of time. The convergence and divergence of different MA periods are helpful in gauging trend change as well. We have already seen technicians mentioning this on several occasions in financial newspapers over the past few months. In additional to the simple calculation of an average price, there are two other types of MA which are commonly applied in the market as well: Exponential and Weighted. The Exponential Moving Average (EMA) assigns greater weight to the latest data and one single weight is placed on the historical average. For the linearlyweighted Moving Average (WMA), a heavier weight is placed on the latest data but it ignores information beyond the observed period e.g. in the case of a 10-day average, the tenth day would be multiplied by ten and the ninth day by nine, and so on. But no weighting is given to days further back than 10. Although both EMA and WMA are ‘weighted’ moving averages, the implications are very different. One obvious point here is the ‘placement of value’ on the data. The EMA imposes one ‘single’ weight (much more valuable) i.e. K on the latest data and another ‘single’ weight for the rest of the data with the relationship of 1-K. See below:

Whereas, the weighting for the WMA is spread across the observed period of data points, see below: WMA = (P1*1 + P2*2 +P3*3 +P4*4 +…PN*N)/(1+2+3+4+…+N) Where P is the price being averaged N is the numbers of days in the Moving Average (selected by the trader) From these formulae, there are three useful observations: a) The EMA allows us to track the ’turning’ points much more quickly than the WMA, see note 1. b) WMA is more useful as a ’confirming‘ signal due to the ‘group’ of heavy weightings placed on the latest set of data. c) Both EMA and WMA are much more sensitive than the SMA, see note 2.

Note 1: The weighting placed on the latest data is the same for both the EMA and WMA e.g. in a 10-day observed period, the weight is 18.18% (2/11) on the latest data for the EMA and the same on the WMA (10/55). In a 20-day period, the EMA has 9.52% (2/21) and the calculation for the WMA is 20/210, equivalent to 9.52% as well. But, the crux is the preceding numbers. The WMA has a stronger set of weights assigned and in fact the second last number has only 5% less in a 20-day average. But, the EMA is a single weight for the rest of the average period,

Note 2: SMA = (P1 +P2 +…+PN)/N Where P is the prices being averaged N is the number of days in the moving averages (chosen by traders)

Where K = 2/(N+1)

This variation can be used to generate potential crossover signals. In this case, it is not between different time periods using one type of MA but a crossover with different MA methods i.e. the combination of SMA, EMA and WMA. Since we know these averages are designed for different purposes, it is reasonable to expect some leading signals to come out from this Moving Average Crossover (MAC) mix.

N = the number of days in the EMA (chosen by the trader) P(t) = today’s price. EMA (y) = the EMA of yesterday

Putting all these on the same bar chart, it is not surprising to see the MA lines have the same overall direction and sometimes they can hug closely together. However, taking a closer look, there are a

EMA = P(t) * K + EMA (y) * (1-K)


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

few MAC signals which could be quite valuable. Points to look for in a bullish trend are: 1) The MAC from the EMA/WMA of SMA as the initial signal of trending, subsequently confirmed by the turn from SMA. 2) The MAC from the WMA of EMA (above SMA) as the signal of trend confirmation. 3) The firm triple MAC of WMA, EMA and SMA is indicative of strong trending i.e. EMA is above SMA and WMA is above EMA. 4) The MAC from EMA of SMA after point 3) as the signal of trend change i.e. from bullish to consolidative/choppy. Below are examples derived from DM/$. This method is applicable to other financial instruments as well.

Chart 1:

Chart 1 is the weekly data of DM/$ over the past three years with SMA, EMA and WMA of nine days overlaid on the bar chart. The triple cross mentioned in (3) worked very well in 1997. But, the move from trending to consolidative/ choppy stated in (4) was captured nicely in the period of March and August of 97, see chart 2.

Chart 2:

Chart 2 shows a more detailed picture of the period in August when the high of 1.8910 was hit. The corrective trend was subsequently confirmed by the crossover from the EMA of SMA and then later the WMA reacted to the negative move.

Steven Kwan is head of technical research at Standard & Poor’s MMS, London.

101


102

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Improved Moving Average (IMA) Strategies Fotis Papailias (l) and Dimitrios D. Thomakos (r)

Article originally featured in Market Technician 72 (May 2012)

A modified version of, perhaps, the most widely used technical trading strategy - moving averages - is discussed in this article. The suggested approach combines cross-over “buy” signals and a dynamic threshold value which acts as a trailing stop. The trading behaviour and performance achieved using this modified strategy is different to the standard approach with results showing that, on average, the proposed modification increases the cumulative return and the Sharpe ratio of the investor while exhibiting smaller maximum drawdown and less of a drawdown duration than that obtained by using the standard moving average strategy. This article explains how a “moving average dedicated investor” can obtain higher profits (possibly combined with smaller risk characteristics, such as lower standard deviation and lower drawdown) using an improved moving average methodology (henceforth IMA). We define a “moving average dedicated investor” as an individual who forms his or her strategy based on trading signals of price or moving average cross-overs by carefully selecting the moving average (henceforth MA) length (or lengths) via back-testing. The modification we propose is simple, intuitive, has a probabilistic explanation (based on the notion of “return to the origin” in random walk parlance) and can be easily implemented. It consists of a rule that relates the current price of an asset with the price of the last “buy” signal issued by a moving average strategy (making this latter price a dynamic threshold) which then acts as a dynamic trailing stop (either stop-loss or stop-profit depending on the current position). For reasons of simplicity, we focus on the longonly approach. However, for short-sales analysis and/or a full mathematical discussion, we refer readers to our academic papers detailed at the end of this article.

Improved Moving Average Methodology Consider the standard MA(k), where k denotes the length (or “lookback” period), and the trading signals it produces. In the simple case of a price cross-over, the standard strategy is to open a long position if the current price is greater than the current value of the MA(k) (“buy” signal) and, consequently, close a long position if the current price is below the value of the MA(k) (“exit” signal). Our IMA(k) first opens a position using the entry signal from the standard MA(k), and marks this as the (first) “entry” price. If, in the meantime, the standard MA exits a trade and then reenters while the current price is higher than the entry price, the IMA stays put but its entry price is renewed; hence the new entry price acts as a “trailing stop” which has been dynamically updated. In order for the IMA to close a long position there are two conditions: 1) The IMA exits a trade if the current price is below the last updated entry price, or In order for the IMA to close a long

2)

IMA exits a trade if the current price is below a percentage threshold of the last updated entry price. The threshold parameter must first be defined in percentage terms, i.e. t%, and then the threshold price is given by (1-t%) times the last updated entry price. For example, if the last updated entry price is $75 and our threshold parameter is t=0.03 or 3%, the relevant threshold of the last updated entry price is (1-0.03)*$75=$72.75. This has proved to be useful in practice since it reduces the sensitivity to short-term price changes and therefore to over-active trading during choppy markets (this concept is further discussed in the second example).

If the IMA exits the long trade (in one of the above ways), it then enters a trade again: a) If and only if the standard MA provides a new buy signal, or b) If the current price rises above the last updated entry price (while standard MA is still in the trade). For illustration purposes we will discuss a step-by-step example using (1) and (a) of the above conditions when the market is bullish, i.e. the IMA exits a trade if and only if the current price is below the last updated entry price and, once out, IMA opens again a long position if the standard MA provides a new buy signal. Figure 1 (opposite) presents the MA and IMA strategies based on a 200-day look back period. The bottom red line depicts the periods that an investor who follows the standard MA strategy is in a long position. Similarly the top green line depicts the relevant period that an investor who follows the IMA strategy is in a long position. The horizontal green dashed lines mark the updates of the entry price. We are using a one day delay so that our results mimic real life situations, i.e. if we observe a buy signal at time “t”, we open a long position at time “t+1”. In all our analysis we are using the daily adjusted closing prices1. • On 17-04-2003 the current closing price is $89.56, above the 200-day moving average, which is $88.39. Hence, the standard MA gives a buy signal and on 21-04-2003 the investor opens a long position at $89.65. The IMA uses this as its initial entry signal and the investor who follows the IMA strategy opens a long position as well (the green dotted line depicts the entry price of the IMA). • On 16-07-2004 the current price is $110.71 and the current standard MA equals $110.80. The standard MA provides a “sell” signal and the investor who follows the standard MA exits the trade on 19-07-2004 at $110.24. The IMA stays put as the current price of $110.80 is greater than its entry price of $89.65. • On 02-09-2004 the current price is $112.58 and the standard MA value is $111.67. The standard MA provides a buy signal and the investor opens a long position on 03-09-2004 at $112.12.


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

The IMA entry price is, therefore, updated to $112.12. Now, the investor who follows the IMA strategy (who is still long) will exit the trade if and only if the current price falls below the new updated entry price of $112.12.

Figure 1: MA example using a 200-day Moving Average Improved Moving Average Trading Rule

• The above procedure is repeated through the years 2004 - 2007. During this period of rising prices the standard MA enters and exits trades and the IMA entry price is updated. In 2007 prices fall sharply. After staying long during the period of rising prices, the IMA captures the change in trend as a result of the dynamically updated trailing stop. • On 11-09-2007 the investor who follows the standard MA is not in a position and the current price is $147.49. The standard MA is equal to $146.13, so the standard MA signal gives a “buy”. The investor opens a long position on 12-09-2007 at $147.87. As analysed above, the entry price of the IMA is updated to $147.87 and the investor who follows the IMA strategy will close the position if and only if the price falls below the level of $147.87.

Long

• Indeed, on 07-11-2007 the current price is $147.91 and the standard MA is $148.37. Hence the standard MA strategy provides a “sell” signal and the investor exits the trade on 08-11-2007 at $147.16. On 11-08-2007-11 with the current price being equal to $147.16, IMA provides an exit signal due to the fact that the current price fell below the last updated entry price of $147.87, hence the investor who follows the IMA strategy closes the position on 09-112007 at $145.14.

Long

By the above analysis we obtain the following trades from Table 1.

Source: Spy Daily Data

Table 1: Cumulative Return of the example described in Figure 1 Date

Price $

MA Position

Return

21-04-2003

89.65

Enter Long

19-07-2004

110.24

Exit Long

22.97%

03-09-2004

112.12

Enter Long

Long

14-10-2004

110.64

Exit Long

-1.32%

28-10-2004

113.22

Enter Long

Long

18-04-2005

114.5

Exit Long

1.13%

03-05-2005

116.6

Enter Long

Long

06-10-2005

119.2

Exit Long

2.23%

01-11-2005

120.49

Enter Long

Long

08-06-2006

125.75

Exit Long

4.37%

30-06-2006

127.28

Enter Long

Long

13-07-2006

124

Exit Long

-2.58%

26-07-2006

126.83

Enter Long

Long

13-08-2007

145.23

Exit Long

14.51%

12-09-2007

147.87

Enter Long

Long

08-11-2007

147.16

Exit Long

-0.48%

09-11-2007

145.14

Cumulative Return

MA Position

Return

Enter Long

Long

Long

Long

Figure 2 presents an example of IMA exit signals using MA cross-overs.

Long

Assume that an investor trades using signals of short and long moving average cross-overs with lengths of 40- and 200days. Initially, the investor is out of the market.

Long

Long Exit Long

45.36%

In this example the IMA might look like it behaves as a buy and hold strategy, however this is due to (i) the strong trending behaviour of the market and (ii) the large MA length being used. Another aspect that is illustrated in Figure 1 is the IMA’s ability to capture cyclical price behaviour and wait until the right moment when the cycle pattern is broken to exit the position.

61.90% 61.90%

• On 17-06-2009 the 40-day SMA crosses above the 200-day SMA providing a “buy” signal and the investor opens a long position. The IMA

103


104

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

• On 15-07-2009 the current price rises once again above the entry price at $93.26 and the investor enters in a long position on 16-07-2009.

Figure 2: IMA Example using a 20-day Moving Average Improved Moving Average Trading Rule

Using the IMA without incorporating some additional threshold to trigger exit decisions might lead to unsuccessful trades like those described above especially in times when the market is bearish and volatile. This can be avoided using a threshold percentage of the last updated entry price as described in the beginning of this article. Indeed, using a 3% threshold so that the IMA will trigger an exit signal if the current price is below 1-0.03 of the entry price. In Table 2 we see that the 3% threshold price is never below the entry price of $92.22 and hence the sensitivity to sudden short-term changes in prices is elimatinated.

Performance results and combined usage

Source: Spy Daily Data

Table 2: IMA using a threshold price exit signal Date

Price $

18-06-2009

92.22

19-06-2009

92.04

22-06-2009

Price $ IMA 3% threshold Position

IMA - a Position

IMA - b Position

Open Long

Open Long

Open Long

89.28

Long

Long

Long

89.28

86.6

Long

Close Long

Long

29-06-2009

92.7

89.92

Long

30-06-2009

91.95

89.19

Long

Open Long

Long

01-07-2009

92.33

89.56

Long

Close Long

Long

15-07-2009

93.26

90.46

Long

16-07-2009

93.11

90.32

Long

Long

Long Open Long

Long

Notes: IMA- a denotes the IMA using entry condition(a), IMA-b C denotes the IMA using entry condition(b).

uses this as its entry price and provides a “buy” signal as well. The investor opens the position on 18-06-2009 at $92.22. • On 19-06-2009, the current price is $92.04 below the last updated entry price, hence the IMA gives an exit signal. The IMA will enter again using one of the ways described in the previous section, i.e. IMA will either stay out of the market for a large period waiting for the standard MA to exit and re-enter in order to update

it entry price (entry condition [a]), or, given that the MA is still in the market will enter again when the current price rises above the entry price (entry condition [b]). This indeed happens on 29-06-2009 when the current price is $92.7, hence the investor who follows the IMA strategy opens a long position again on 30-06-2009. • However, on 30-6-2009 the current price drops to $91.95 so the IMA gives another exit signal.

In the last part of this article we discuss some back-testing results on “SPY”, which tracks the S&P 500 index. We do so since a frequent criticism of the IMA approach is that it does not always improve the relevant standard MA approach. This is true - but no strategy can be totally fool-proof! We are looking for an improvement that performs well on average, across combinations of MA methods and look-back periods. However, as noted earlier, we are proponents of fully back-testing different MAs (simple, weighted, exponential etc.) using different lengths and varying periods in order to find the combination that provides the most desired result on average. Our claim then is that one will find an IMA approach that, if followed, will give the investor better returns on balance than a standard MA approach. We refer the interested reader to our website to see full details as to how one can avoid the curse of data mining. We stress that the example that follows is for illustrating the potential of the method. Let us then continue by focusing on a period where the market was going down, that is from 01-10-2007 to 01-02-2009 and a buy and hold or the standard MA suffered losses. We expect that the IMA will suffer losses too, but not always. If we look carefully at Table 3, we observe that the IMA-a (which is using condition (a) to determine an entry) and IMA-b (which is using way (b) ) provide lower cumulative losses compared to the standard approach and the buy and hold. If we try different MA lengths and different MA rules, we then see that if the IMA-a cross-over of the exponential MA (20, 50) is followed, the investor achieves a noloss position.


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

Table 3: IMA Evaluation: 01-10-2007 to 01-02-2009. Notes: Sharpe is calculated using the annualised average return and standard deviation of returns, MAC denotes the MA Cross-Over, IMA-a C denotes the IMA using entry condition (a) Cross-Over, IMA-b C denotes the IMA using entry condition (b) Cross-Over. Simple MA (20, 50)

Simple MA (20, 100)

Sharpe

Cumulative

Drawdown

Duration

Sharpe

Cumulative

Drawdown

Duration

Standard MA(20)

-1.384

-0.235

0.242

285

-1.384

-0.235

0.242

285

Standard MA(50)

-2.260

-0.203

0.211

276

-0.651

-0.046

0.064

177

Standard MAC

-1.207

-0.196

0.252

177

-1.020

-0.072

0.083

177

IMA-a(20)

-1.161

-0.0194

0.229

285

-0.726

-0.125

0.182

177

IMA-a(50)

-1.299

-0.139

0.147

285

-1.600

-0.016

0.016

197

IMA-a C

-0.549

-0.050

0.095

177

-0.982

-0.020

0.020

186

IMA-b(20)

-1.700

-0.244

0.252

285

-1.243

-0.080

0.189

177

IMA-b(50)

-1.301

-0.095

0.110

177

-1.338

-0.052

0.054

197

IMA-b C

-0.549

-0.050

0.095

177

-1.418

-0.042

0.042

186

Weighted MA (20, 50)

Weighted MA (20, 100)

Sharpe

Cumulative

Drawdown

Duration

Sharpe

Cumulative

Drawdown

Duration

Standard MA(20)

-1.405

-0.273

0.280

285

-1.405

-0.273

0.280

285

Standard MA(50)

-1.702

-0.213

0.220

284

-0.089

-0.008

0.047

177

Standard MAC

-0.744

-0.020

0.165

276

-0.271

-0.022

0.064

177

IMA-a(20)

-0.814

-0.165

0.202

285

-0.428

-0.095

0.152

177

IMA-a(50)

-1.284

-0.155

0.164

285

-0.089

-0.008

0.047

177

IMA-a C

-0.931

-0.030

0.045

276

-0.110

-0.006

0.006

197

IMA-b(20)

-1.459

-0.262

0.269

285

-0.079

-0.199

0.205

235

IMA-b(50)

-1.059

-0.128

0.137

284

-0.089

-0.008

0.047

177

IMA-b C

-0.649

-0.069

0.088

177

-0.419

-0.022

0.040

177

Exponential MA (20, 50)

Exponential MA (20, 100)

Sharpe

Cumulative

Drawdown

Duration

Sharpe

Cumulative

Drawdown

Duration

Standard MA(20)

-2.111

-0.344

0.351

285

-2.111

-0.344

0.351

285

Standard MA(50)

-1.040

-0.077

0.101

177

-1.248

-0.048

0.050

192

Standard MAC

-0.131

-0.011

0.064

177

-0.257

-0.011

0.038

177

IMA-a(20)

-1.404

-0.245

0.255

177

-1.349

-0.232

0.255

177

IMA-a(50)

-1.191

-0.094

0.106

285

-1.229

-0.046

0.057

177

IMA-a C

0.154

-0.007

0.038

177

-1.394

-0.039

0.042

177

IMA-b(20)

-1.842

-0.300

0.307

285

-1.662

-0.272

0.279

177

IMA-b(50)

-0.764

-0.055

0.100

177

-0.875

-0.029

0.041

177

IMA-b C

-0.159

-0.013

0.052

177

-0.901

-0.020

0.025

177

Buy and Hold

Buy & Hold

Sharpe

Cumulative

Drawdown

Duration

-1.041

-0.463

0.518

330

105


106

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Table 4: IMA Evaluation: 01-02-2009 to 04-03-2012. Notes: Sharpe is calculated using the annualised average return and standard deviation of returns, MAC denotes the MA Cross-Over, IMA-a C denotes the IMA using entry condition (a) Cross-Over, IMA-b C denotes the IMA using entry condition (b) Cross-Over. Simple MA (20, 50)

Simple MA (20, 100)

Sharpe

Cumulative

Drawdown

Duration

Sharpe

Cumulative

Drawdown

Duration

Standard MA(20)

0.626

0.249

0.160

260

0.551

0.193

0.160

260

Standard MA(50)

0.897

0.407

0.129

173

0.688

0.263

0.175

212

Standard MAC

0.280

0.099

0.254

248

0.688

0.285

0.225

212

IMA-a(20)

0.489

0.150

0.102

428

0.456

0.129

0.102

428

IMA-a(50)

1.021

0.440

0.088

170

0.706

0.171

0.105

205

IMA-a C

0.254

0.056

0.116

245

0.494

0.119

0.087

212

IMA-b(20)

0.522

0.191

0.139

260

0.487

0.158

0.139

260

IMA-b(50)

0.777

0.334

0.141

260

0.786

0.285

0.147

212

IMA-b C

0.394

0.140

0.156

270

0.934

0.338

0.116

212

Weighted MA (20, 50)

Weighted MA (20, 100)

Sharpe

Cumulative

Drawdown

Duration

Sharpe

Cumulative

Drawdown

Duration

Standard MA(20)

0.716

0.291

0.163

260

0.603

0.213

0.163

260

Standard MA(50)

0.769

0.335

0.133

260

0.774

0.303

0.150

260

Standard MAC

0.560

0.233

0.177

260

0.637

0.262

0.173

321

IMA-a(20)

1.080

0.481

0.104

188

0.985

0.378

0.104

188

IMA-a(50)

0.727

0.287

0.140

451

0.910

0.340

0.115

469

IMA-a C

-0.007

-0.006

0.098

469

-0.233

-0.041

0.121

469

IMA-b(20)

0.874

0.361

0.115

260

0.772

0.275

0.115

260

IMA-b(50)

0.864

0.372

0.106

240

0.981

0.397

0.150

253

IMA-b C

0.573

0.223

0.163

260

0.705

0.256

0.152

266

Exponential MA (20, 50)

Exponential MA (20, 100)

Sharpe

Cumulative

Drawdown

Duration

Sharpe

Cumulative

Drawdown

Duration

Standard MA(20)

0.658

0.267

0.191

260

0.544

0.187

0.191

260

Standard MA(50)

1.000

0.474

0.133

249

0.845

0.346

0.140

212

Standard MAC

0.821

0.392

0.151

212

0.471

0.175

0.271

212

IMA-a(20)

0.894

0.327

0.098

222

1.224

0.452

0.098

118

IMA-a(50)

0.801

0.303

0.104

464

0.908

0.237

0.082

180

IMA-a C

0.357

0.065

0.087

382

0.687

0.153

0.071

212

IMA-b(20)

0.861

0.365

0.129

258

0.859

0.308

0.129

258

IMA-b(50)

1.049

0.492

0.149

239

0.879

0.339

0.140

212

IMA-b C

0.922

0.432

0.110

212

0.893

0.345

0.128

212

Buy and Hold

Buy & Hold

Sharpe

Cumulative

Drawdown

Duration

0.868

0.663

0.218

203


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

The next best option is the IMA-b (50) of the exponential MA where the cumulative return is -5.5%. Both these qualitative results are extremely important as they stress the fact that there was an IMA that would have kept the investor in the market -alive- in a difficult period. Furthermore, if the investor kept on following the IMA-b(50) during the subsequent years after the crisis (i.e. 01-02-2009 to 04-03-2012), we see in Table 4, that it is the most profitable strategy after the buy and hold. This outcome is important as it stresses the fact that the IMA is not a result of datamining. By carefully selecting an IMA strategy and following it, the investor suffers minimum losses during the crisis and a large return when the market is bullish. Furthermore, IMA-b(50) provides a drawdown which is one of the lowest among all strategies and it is one and a half times less compared to the drawdown of the buy and hold. And not only that: among all the strategies considered, in all combinations presented in Table 4, we can see that the IMA approach provides, on average, lower drawdowns and that, out of the five cases with a Sharpe ratio greater than one, four are using the IMA approach while only one uses the standard approach.

Conclusions In this article we have presented a review of our academic research on an improved moving average methodology, which can be easily and we think profitably - used by market practitioners. Our modification is based on an updated threshold value which is defined by the time-varying “buy” signals of the standard cross-over strategy and acts as a dynamic trailing stop. This implies a different behaviour and performance for the modified strategy compared to the standard one and we find that, on average, the modification improves trading performance by a wide margin across a number of evaluation measures. In theory, the dynamic trailing stop as introduced here could be possibly applied to any technical rule that provides trading signals. However, we have not tested this yet and our underlying theory justifies its usefulness when used with moving averages. An investor interested in using this approach for trading would do well to perform some extensive back-testing in order to choose the IMA that has the best average performance across different evaluation periods (which should include bear markets). The IMA methodology will not, in fact it cannot, always provide better results against the standard MA, but there will always be one or more combinations where the IMA will be the most profitable: in our research we clearly document it is on average more profitable, the more you use it the higher the chances that it will be the best performer - and here lies its strength. More importantly, besides increasing the cumulative return, it does so without increasing the risk-reward ratio: the modified strategy exhibits, on average, smaller maximum drawdown and smaller drawdown duration and in many cases a higher Sharpe ratio. These quantities are important to the investor: large drawdowns are catastrophic since they wipe out a large part of the invested capital making it difficult, if not impossible, for someone to return to the markets. We hope that the IMA approach will help anyone that is interested in this type of technical trading. This article is a non-technical overview of the main ideas from two of our research papers “An Improved Moving Average Technical Trading Rule” and “An Improved Moving Average Technical Trading Rule II: Can we obtain performance improvements with short sales?” where additional results and discussion, as well as freely available R-Code can be found. We would like to thank Bob Fulks and Kent Russell for suggestions on the methodology. We are also indebted to the editor, Deborah Owen, for her help and patience in the preparation of this article.

Fotis Papailias is a PhD Candidate in Econometrics at the School of Economics and Finance, Queen Mary, University of London, UK and co-owner and co-maintainer of www.quantf.com

Dr. Dimitrios D. Thomakos is Professor of Applied Econometrics at the Department of Economics, University of Peloponnese, Greece and co-owner and co-maintainer of www.quantf.com

107


108

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

The Trend Intensity Indicator Richard Lie

Article originally featured in Market Technician 48 (November 2003)

Introduction Market sentiment plays a critical role in assessing share price movements. This article describes an indicator that I have developed which delivers a consistent measurement of market sentiment using a unique combination of indicators. I have called it the Trend Intensity Indicator. The Trend Intensity Indicator combines and weighs four simple tools: trend, volume, moving averages and price momentum. This generates an invaluable benchmark that highlights only those stocks with compelling trending qualities that offer the best prospects for sustained price movement.

Why sentiment?

The Trend Intensity Indicator calculates a single value from a “basket” of sentiment indicators. 1) 2) 3) 4)

Trend - for direction Price/volume - investor participation/non participation Moving average - for averaging probabilities Price momentum - to define the power of the crowd behaviour.

The calculation Indicator

Status

Price:

The motivation for designing an indicator evolved from a realisation that fundamental analysis was not necessarily providing all the answers, nor explaining many share price movements.

1a. Trend Up

Higher Highs and Higher Lows

1b. Trend Down

Lower Lows and Lower Highs

Price analysis is the examination of a company’s share price. At any one point in time buyers and sellers agree on a market price, which is a direct reflection of market sentiment and drives the share price.

Volume:

Fundamental and price analysis are two entirely separate moving targets that very often diverge, but it is movement in the share price which we are most interested in and from which we profit.

2a. Volume Bullish

Volume Expanding & Price Rising

2b. Corrective in up trend

Volume Contracting & Price Rising

2c. Volume Bearish

Volume Expanding & Price Falling

2d. Corrective in down trend

Volume Contracting & Price Falling

The “sentiment” factor can drive prices far from fundamental value. The unexpected news of the 2002 profit downgrades were dealt with so harshly by the market that it drove prices well below their fundamental value. By contrast, future expectations can build prices to unrealistic levels far beyond fundamental value. This is what made the Biotech sector such a burial ground of shattered dreams - so many expectations. And of course, the Internet bubble was built almost entirely on sentiment with little consideration for value at all.

Moving Average (XMA):

Sentiment is a powerful force, and an understanding is essential to successful stock market trading and investing. The difference between a stock’s fundamental valuation and its share price could be explained as the “sentiment” factor.

Moving Average Convergence / Divergence Indicator:

How do we measure sentiment? To initiate analysis using the Trend Intensity Indicator a definition of trend is first established. A stock, which moves in a sequence of higher highs and higher lows, is defined as having an uptrend. At the point that this sequence begins, i.e. when it changes from a downtrend of lower highs and lower lows,we consider the trend reversed. Only weekly reversals are employed in our approach. The basis for this rule is that a weekly trend change avoids daily market noise. It is a reliable medium term indicator and provides a clear and objective view of market sentiment. Once a stock fulfils a simple trend definition, our Trend Intensity Indicator then rates the power of that trend and establishes a clearview of market sentiment towards it, or against it.

3a. Moving Average Positive

Price < XMA

3b. Moving Average Negative

Price < XMA

3c. Moving Average Neutral

Prices closes once above/ below XMA

4a. MACD - Positive

MACD lines > 0, MACD Histogram Rising and > 0

4b

MACD lines > 0, MACD Histogram Rising and < 0

4c

MACD lines > 0, MACD Histogram Falling and > 0

4d

MACD lines > 0, MACD Histogram Falling and < 0

4e. MACD - Negative

MACD lines < 0, MACD Histogram Falling < 0

4bf

MACD lines < 0, MACD Histogram Falling > 0

4g

MACD lines < 0, MACD Histogram Rising > 0

4h

MACD lines < 0, MACD Histogram Rising < 0


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

By taking each indicator and breaking it down to the most basic signals, we provide a value for the state each indicator is in. For trend, it is either up or down which then receives a negative or positive value reflecting that state. This is then weighted into the end result, its Trend Intensity Rating. Of the other three indicators we ask the question: Does volume support the rise? Where is the price in relation to its moving average? And is price movement attracting positive crowd behaviour? The table on the left shows the indicators used for calculating the Trend Intensity Indicator and their different states. The values and calculations that generate the Trend Intensity Rating for each stock are proprietary to Stockradar, however, the objective rule-based approach becomes clear with only the most reliable signals being employed.

Trend Intensity Indicator The Trend Intensity Indicator calculation generates a stock rating between 10 and -10. The highest value of 10 reflects a consensus agreement by all indicators that all positive sentiment rules have been satisfied and the stock is, therefore, rated at a maximum on the Trend Intensity scale. The lowest value of -10 reflects consensus agreement from all indicators that no conditions have been met that suggest a stock has any positive sentiment towards it. A stock that reverses its trend to up and has a Trend Intensity Rating of 4 or greater will qualify as a Stockradar Stock Pick and, as such, will have compelling trending qualities and offer the best prospects of price movement. Alternatively, a stock that reverses its trend and has a Trend Intensity Rating of -4 or less will be disqualified from Stockradar’s Stock Picks on the grounds that it has lost its trending qualities. This breaks the market down into two distinct groups of stocks. One that is trending, or one that is not. Our focus is on up trending stocks only.

Trend Intensity Indicator Rating Scale Rating

Status

10 to 7

Trending up strongly

6 to 4

Trending up

3 to -3

Neutral

-3 to -6

Trending down

-6 to -10

Trending down strongly

The Result Stockradar’s coverage is of the ASX/300. Weekly results are presented each Monday with our recommended Stock Picks at Stockradar.com.au. Our weekly Stock Picks are supported by a Stock Alert facility that scans the market daily, targeting stocks that are moving in and out of their trends. Along with specific stock analysis, the Weekly Sector Update takes on a thorough review of a market sector. Published bi-monthly our free newsletter features a selection of market highlights.

Richard Lie is an independent research provider licensed by the ASIC (Australian Securities and Investments Commission).

Psychology and Markets

Systematic Trading

109


110

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Least Squares Regression: Fitting Data to a Straight Line Jeremy du Plessis

Article originally featured in Market Technician 21 (November 1994)

Ordinary Least Squares Regression (OLSR), to give it its correct title, is a fancy name for a rather simple, but effective mathematical technique for establishing the relationship between two sets of data. In Technical Analysis, the one set is usually price and the other is always time. Through a rather tedious calculation, OLSR determines the equation (y = bx + - c) for the straight line which best fits the data and thus establishes its trend. It is therefore another powerful tool in the Technical Analyst’s arsenal - powerful because it does not require subjective positioning.

the line become Irrelevant. Using the FT-SE 100 index, let us investigate to see whether or not OLSR lines are constant. Chart 1:

As the OLSR line is derived from the data under consideration, the price should not deviate too far away from the line for any length of time. Placing parallel lines either side of the OLSR line forms a, channel in which the price should remain. If it reaches the upper bound of the channel, it could be considered overbought, and if it reaches the lower bound, it could be considered oversold. There is some argument for placing parallel lines a certain number of standard deviations away from the OLSR line in order to contain a certain percentage of the price movement. For example, parallel lines drawn one standard deviation away from the OLSR line give a confidence level of 68%, meaning that 68% of price behaviour should be contained within the parallels. This is a rather low percentage, so it is more common to use 1.96 standard deviations, which gives a confidence level of 95%. These confidence levels may be obtained from Normal Probability Distribution tables. Many analysts however, including this writer, prefer simply to draw the parallel lines in such a way that they form trend lines picking up a series of tops or bottoms instead. This is more consistent with Technical Analysis, where much of the analyst’s time is spent determining the trend and whether it is still intact. There is also no justification for the bands being placed equidistant from the OLSR line. The reason being, that in rising markets, the price will tend to reach greater extremes above the OLSR line than below it, although in total it will, by definition, spend an equal amount of ‘time’ below as above.

Chart 2:

Chart 1 shows the FT-SE 100 on a weekly basis since inception. The OLSR line is shown running through the data, with a parallel line below, that picks up the trend very well and a parallel line above that contains much of the price movement. It is significant to note that on only two occasions has the price deviated by an excessive amount above the upper parallel - in October 1987 and January 1994! Both have led to fairly substantial corrections. It also deviated below the lower band on two occasions, which preceded substantial upmoves.

How Constant is the Gradient of the OLSR Over Time?

Chart 2 shows the FT-SE 100 index on a weekly basis, from 1984 with a number of OLSR lines grouped as follows:

For the OLSR line to be effective, its slope should remain fairly constant. If it does not, the interpretation of the deviations from

Group labelled (A) - One line drawn: from 1/1/84 to 1/1/88


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Group labelled (B) - Two lines drawn: from 1/1/84 to 1/1/89 from 1/1/84 to 1/1/90 Group labelled (C) - Four lines drawn: from 1/1/84 to 1/1/91 from 1/1/84 to 1/1/92 from 1/1/84 to 1/1/93 from 1/1/84 to 1/1/94 The four lines labelled (C) are the OLSR lines for the last four years. It is difficult to see each line, but that only goes to show that between January 1991 and January 1994 there was very little difference in the equation for the OLSR line drawn from the beginning of the time series in 1984. The two lines labelled (B) are for the two preceding years and the line (A) is up to January 1988, just after the ‘87 correction. So, what does this tell us? For a start it shows that the OLSR line has been constant since January 1991. In fact, the 4 lines are so close together that they are almost on top of one another. This means that anyone using the OLSR line and its parallels as an overbought oversold indicator since 1991 has had constant results. But if one goes back even further, the two lines labelled (B) for the two preceding years are very similar as well. In fact parallels would also have shown January 1994 as an extremely overbought time. Only the line drawn to the beginning of 1988 has a different slope, which is not unexpected. After all, the market had a major change of mood towards the end of 1987.

Psychology and Markets

Systematic Trading

Where does the Least Squares Detrend fit in? Allied to the OLSR line is the Least Squares Detrend. This is obtained by removing the slope of the OLSR line, making it the zero line and plotting the percentage difference between the price and the OLSR line. This produces an oscillator which shows not only overbought and oversold levels better than the OLSR line itself - but also shows up cycles in the data far better. It is slightly different from the OLSR line and parallels described earlier, because it deals with percentage differences, whereas the OLSR line and parallels show absolute differences. Chart 4 shows the FT-SE 100 on a daily basis with an OLSR line and parallels drawn. Below in Chart 5 is a Least Squares Detrend. The zero line is the OLSR line with its trend removed. Notice how the line, which is the FT-SE 100 index with its trend removed, oscillates above and below the zero line. The scaling is in percentages. Notice also that when the line gets to around 10% overbought or oversold from the OLSR line (zero line), it tends to make a top Charts 4 and 5:

The same exercise was carried out on a daily Chart 3 of the FTSE 100. The first line was drawn from 31st March 1992 to 31st March 1993 and then every three months thereafter up to 30th June 1994. Chart 3:

or bottom. The upper and lower charts are directly comparable. Where the price crosses the OLSR line in Chart 4, the oscillator crosses the zero line in Chart 5. In many cases, it is far better to use the Least Squares Detrend when looking for overbought and oversold conditions, because it shows them in percentage, rather than absolute’ terms. This is demonstrated in Charts 6 and 7 which are of the FT-SE index on a weekly basis. Charts 6 and 7:

Once again, the lines are fairly close together, although not quite as close as on the weekly chart. What is important however, is that any extreme deviations from the prevailing trend (the OLSR line) always show up as constant and can thus be defined as overbought or oversold. It must be concluded therefore that the trend of the OLSR line tends to change slowly unless there is a major change in the market, as witnessed by the 1987 downturn. OLSR lines on weekly time series charts are more constant than those on daily charts. OLSR lines do therefore provide a reliable way of fitting the data to a channel. If new OLSR lines are calculated and drawn on a regular basis, any change in the trend will be gradual. The analyst will therefore automatically adjust his analysis as these gradual changes take place.

111


112

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Although the index crosses the OLSR line in Chart 6 at exactly the same time as the oscillator crosses the zero line in Chart 7, the deviations are different. For example, the Least Squares Detrend in Chart 7 shows the index as 15% overbought in January 1994, but nearly 40% overbought in October 1987. Whereas in Chart 6, the index deviated by a similar amount above the upper band on those two dates. This discrepancy is only apparent when there is a big price difference, such as 2400 in 1987 compared to 3500 in 1994. This means that it will be not so apparent with daily charts which show less data and where the price difference from one year to next is smaller. This brings us to another aspect of OLSR analysis which is not generally explored - log scaled charts and OLSR lines.

Log Scaled Charts Some analysts believe that you cannot draw OLSR lines on log scaled charts because they would appear as curved lines and show much the same thing as straight lines on arithmetic charts. This is rather inflexible. We all draw straight trend lines on log scaled charts and many of us believe that these are better than those drawn on arithmetic scaled charts, so why can’t we draw straight OLSR lines on log scaled charts? It’s simple to do — instead of logging the data after calculating the OLSR line, the data is logged first and then the OLSR line is calculated on the logged data. The result is quite interesting as Chart 8 shows. Chart 8:

Compare the OLSR line with Chart 1. The OLSR line goes through different points on the log scaled chart, as do the parallels. However, what is interesting is that the upper and lower bands of the channel are exceeded by differing degrees. For example, in the arithmetic Chart 1, Jan ‘94 was just as overbought as Oct ‘87, but in log Chart 8, Jan ‘94 was nowhere near as overbought as Oct ‘87. In fact it was not as overbought as Oct ‘89 or Apr ‘86. This does cast some doubt over the arithmetic technique and begs the question as to which method is better. It does not however detract from the initial principle that the OLSR method establishes the trend of the data under consideration. Whether parallel lines on an arithmetic chart can be used to measure the degree with which the price is overbought or oversold, or whether a log scaled chart is better, is a matter for further discussion and debate. It may be as well to remember however, that the difference between log and arithmetic OLSR lines and parallels is only significant when there are large price differences between their beginning and end of the time series.

Jeremy du Plessis is a Director of INDEXIA Research Limited.

Pattern Recognition and Pattern Analysis

Moving Averages and Trends


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

Volstall: Using volatility to identify early signs of Trend Exhaustion Oliver Woolf

Article originally featured in Market Technician 76 (April 2014)

Woolf’s Volatility Stall, better known as Volstall, is designed to identify trend exhaustions by exploiting the trend factor that is characteristic of standard deviation, more commonly employed as a measure of volatility. The most familiar volatility-based study is John Bollinger’s Bollinger Bands, which draws on the statistical theorem that 95% of observations within a standard distribution should fall within two standard deviations of the mean. Bollinger Bands interpret this mean as a moving average, which by default is set at 20 periods, and subsequently two standard deviation bands are drawn around the average as seen in Figure 1 on the FTSE 100.

It must be noted that neither use of Bollinger Bands should be implemented alone, and that both can be extremely effective when combined with other studies to better detect the probability of mean reversion or breakout. There is, however, a ‘third way’ with standard deviation and this is Volstall. It is a hybrid philosophy that is mean reversion in nature, but achieves its goal by drawing on standard deviation’s sensitivity to trend strength. Figure 4 (overleaf) again displays Bollinger Figure 1:

In principle, there are two customary methods of trading Bollinger Bands: 1) As they represent two standard deviations on either side of the mean (and thus theoretically 95% of all observations), the natural and most common application is to employ them in a mean reversion strategy, buying a cross above the lower band and selling a cross below the upper band. Figure 2 highlights where this would have generated signals on the FTSE, the blue circles marking potential buys and the red sells. The flaw of this strategy is that price movement is not normally distributed. Therefore, as a contrarian strategy it produces reversal signals at exactly the wrong time - when a trend has strong momentum. In these cases, such as the 1st blue signal on the left hand side of the chart, the backward-looking nature of the strategy suggests that there is exhaustion when precisely the opposite is the case. In addition, when this approach is applied in such a systematic manner, it fails to spot a potential reversal if, despite a strong move, there is no actual cross above or below the bands, such as was the case at the recent high in late February, and the high of May last year, marked by the orange circles. 2)

The alternative strategy adopts completely the opposite approach, thus converting the flaw of the mean reversal principle into its favour. Its goal is to identify breakouts from ranges into new trends, the notion being that breakouts will be accompanied by strong, volatile moves. To do this it flips the mean reversion signals so that a cross above the upper band (blue circle in Figure 3 (overleaf) is a breakout into a new bull trend and a cross below the lower (red circle) a breakout into a new bear trend.

This second approach, highlighted in Figure 3, which is very much trend following as opposed to mean reversion, gave decent signals when there were strong impulses, such as that marked by the first red circle on the chart. However, there were numerous disastrous signals, most notably the red short at the June low when the FTSE subsequently rebounded strongly.

Figure 2: BB Mean Reversion

113


114

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Figure 3: BB Breakout

Figure 5:

Figure 4: Slope & Standard Dev 1

Figure 6:

Bands on the FTSE 100, but this time with a sub-panel which depicts the Bollinger Bandwidth, the amplitude between the 2 bands and hence a proxy for standard deviation. The fact that standard deviation of price is indicative of trend strength is clearly evident; when the slope of the trend becomes steep, highlighted by the thick, red regression lines on the price, the standard deviation increases, as marked by the corresponding red arrows in the subpanel. The statistical logic here is simple; the standard deviation is increasing as the slope steepens because the sample of close prices over a given time period is becoming more widely distributed. Conversely, as the trend begins to wane the distribution of the price sample becomes narrower and the repercussion on the standard deviation is that it starts to invert from its peak. These inversions are marked by the dotted red lines in Figure 5, whilst the loss of trend strength is highlighted by the flattening of the red regression lines on the price. There are two observations that are apparent in Figure 5: Firstly, the standard deviation inversions from peaks naturally occur when the trend is slowing and often at the point of price reversals.

Moving Averages and Trends

However, whilst they occur simultaneously with price reversal they do not provide any leading indication. Secondly, the slope of the price, as denoted by the regression line, flattens as the price and standard deviation approach their peaks. Importantly, this does precede the trend peak and therefore does act as a leading indicator. The most effective way to measure the rate of ascent/descent is to use a rate of change and hence, to calculate Volstall, a rate of change is applied to the standard deviation. In doing this it identifies when the acceleration in the standard deviation’s ascent starts to slow before it inverts and consequently it anticipates a trend change before it actually materialises. To be precise, the calculation takes a 19-period moving standard deviation (19 was chosen to be slightly quicker than Bollinger Bands) and applies to it a 5-period rate of change. The resultant Volstall indicator is the blue histogram in the subpanel of Figure 6. Overlaid on the histogram is a 1 period lag of the histogram (black line) and a 15% threshold level (dashed red line). The existence of the black lagged line is to identify the points at which the histogram turns downwards from a peak as it will cross below the line. The threshold level is added to filter out minor


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Figure 7:

Figure 9:

Figure 8:

Figure 10:

noise and thus isolate only the strong moves. Figure 7 highlights when the trend is particularly strong and still accelerating by painting the bars pink when the Volstall is above the threshold and rising. This serves as a warning that exhaustion may be looming. The subsequent Figure 8 goes one step further by marking the exhaustions when the slope of the histogram turns from positive to negative (crosses below the one day lag) with a yellow X on the price. It is apparent here that these signals are able to spot trend exhaustions rather effectively. The reason for classifying these signals as exhaustion points, rather than reversals, is that the mere fact that a trend has slowed does not guarantee a reversal, although it does increase the probability. To paraphrase my colleague, Eoghan Leahy, we can draw parallels with the speed and acceleration of a car. If we compare standard deviation which, as proved above, measures the strength of the trend, with the speed of a car, then Volstall, as the rate of change of the standard deviation, is akin to the level of acceleration. If you take the foot off the accelerator it does not mean that the car will reverse but, in order to reverse the car, you must first decelerate. Similarly, a loss of trend acceleration, measured as a fall in the Volstall level does not automatically harbinger a reversal but it is a likely precursor.

Systematic Trading

With this in mind, as Volstall does seem to generate extremely good reversals in wide ranging markets, as has been the case with the FTSE 100 over the last few months, it is necessary to combine the raw Volstall described above with another ingredient in order to convert the generic yellow reversal Xs into more refined bullish and bearish signals. This extra ingredient is the RSI. If RSI is greater than 50 the implication is that the prior trend must have been positive and therefore the yellow X becomes a downward red arrow. Conversely, if the Volstall signal is concurrent with an RSI reading below 50 then an upward blue arrow is marked on the chart. These bullish and bearish Volstall signals are exhibited in Figure 9. As discussed previously, Volstall is an indicator rather than a complete trading system. As its name suggests it identifies when volatility, measured as standard deviation, is beginning to stall, but it does not necessarily herald a complete trend change. For this reason it would be recommended that its use within a trading system be complemented with other timing techniques and adequate management of risk such as stop losses. However, a historical simulation of Volstall in a fully invested stop and reverse strategy does nevertheless prove its efficacy at timing reversals in wide range-bound markets. The example in Figure 10, backtested

115


116

Theory of Technical Analysis

Figure 11:

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

on GBP/USD using a daily frequency over the last three years, has a profit of close to 40% with a winning ratio of 85% from a total of 20 trades. For Volstall to succeed, the key is an appreciation of the behaviour of the security onto which it is applied. As stated previously, given that Volstall’s goal is to identify potential reversals, it is most effective on wide range bound markets - range bound, to increase the probability of mean reversion and wide because Volstall relies on some element of a trend being present to locate the exhaustion point. Having discussed GBP/USD, Figure 11 demonstrates how effective it is on other major USD crosses, such as, EUR, CHF, NZD, CAD and SEK. This is probably because, as ratios, there are inherent pressures from both directions which often constrain them within certain boundaries. However, their liquidity also means that they move very quickly and this facilitates the important trend element. By re-testing the strategy previously simulated on GBP/USD, but this time on a basket including the other five crosses from Figure 11, the result, displayed in Figure 12 would be a fairly decent return and, from late 2011, very minimal drawdowns.

Figure 12:


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

117


118

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

CHAPTER SIX

Indicators and Momentum Articles in this chapter 121 The Momentum Effect - Puzzling but True Deborah Owen

124 The MACD Reversal Indicator Gerald Ashley

125 Momentum, Trending and Sentiment Richard Adcock

130 Variable MACD: Adapting to Financial Market Dynamics G. Gandolfi, M. Rossolini and A. Sabatini

133 Bollinger Bands John Bollinger


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

CHAPTER SIX INTRODUCTION

By Professor Ronald Giles, former lecturer at Queen Mary University London

Introduction Momentum and Momentum Indicators Charts display a picture of financial price information activity over time. This is unhelpful for technical analysts who wish to manage a trading system. Normally incremental changes are measured as deviations from a forward moving average (fma) construct to the actual data. This construct is subjective and depends on the time frame and individual preference. However there are guidelines. The convergence/divergence (MACD) to the actual data is often shown as a supplement section under the price time series. Momentum measures how quickly prices are rising (falling) by changing the steepness of the trendline. Oversmoothing (undersmoothing) can produce too many (too little) signals. There is a series of momentum indicators running into double figures showing this is an active and expanding part of technical analysis. The following is a distinct selection of considered articles demonstrating the breadth of the subject. This topic will evolve as it is one of the main drivers for technicians and many use it as a core part to trading. It is useful to source out origin articles to have a fuller meaning of results rather than current readings interpretation. When momentum is confirming a price trend, a confirmation occurs. A sign of price change trend is when momentum fails to confirm the price trend slope then the resulting divergence is a signal that a technical analyst will note. Confirmation has another attribute in identifying overbought/oversold conditions. When prices are severely below a trend they are deemed oversold with prospects of returning to the central trend or beyond. Hence momentum indicators are the beginnings of core elements in technical analysis. There are other oscillators that use additional information besides price to confirm a position. Volume is a consideration as are option trading opening and closing positions. However too many momentum oscillators can give mixed signals if not false signals. Most academic studies attempt to demonstrate whether price action is random or whether the non-randomness violates the efficient market hypothesis. This is little use to practical technical analysis. However academic activity should not be dismissed outright. Diverse academic research from seemingly unrelated topics can have a future benefit. If current research by backtesting shows that buy and hold consistently outperforms a particular oscillator then this importance should be recorded. A selection of diverse momentum and momentum indicators are presented to illustrate the range of research undertaken in this area. They are recent advances in momentum and momentum indicators taken from past editions of the Society of Technical

Psychology and Markets

Systematic Trading

Analysts journal. Hence the subject continues to advance new ideas. Constructing momentum statistical tests present a challenge because of stop losses amongst other triggers. However, Deborah Owen has highlighted a backtest that demonstrates that momentum trading gives a higher return compared to standard buy and hold. Gerald Ashley presents a discussion on the MACD Reversal indicator as an extension of the Appel MACD by picking changes in trend. Richard Adcock uses information in Japanese candlestick groupings to understand sentiment in the market. G. Gandolfi, M. Rossolini and A. Sabatini construct a variable MACD that adapts to financial market dynamics. John Bollinger has developed the famous Bollinger bands to give range to moving average based on a quasi 95 per cent interval. Owen. The momentum effect is puzzling but true. Dimson et al in the ABN year book 2008 report there is a tendency for UK stock returns to trend in the same direction for the period 1900-2007. This is a major puzzle highlighted by an academic team. Other markets come to the same conclusion but with differing time frames. Deborah Owen reports that the momentum effect persisted throughout the 20th century and continues to hold good. A portfolio of the top 20 per cent is bought and the bottom 20 per cent sold. This is recalculated every month. The overall annualised difference was 10.8 percent. Hence this produces a better performance than using the overall equity market. Ashley develops a complex formula to determine future price levels at which the MACD line crosses the signal line. An example is useful in the FOREX market. One can consider this as a form of stoploss. The benefit is false signalling is reduced requiring less frequent trading. When viewed as a histogram it shows the first signs of a movement fading. MACD was developed by Gerald Appel. The formula is based upon 2 exponential moving averages (ewma) of 12.33 and 25.67 days. This equates to 0.15 and 0.075 smoothing constants. i.e.. (2-B)/B =( 2- 0.15)/0.15 = 12.33 days Conversely the number of days formula can be recalculated as 2/(n+1) = 2/(12.33 +1) = 0.15 The so called MACDREV figure proposed by Ashley is a combination of the 2 ewmas. It will give the exchange rate using the latest spot quote at which the macd lines will cross over. Ashley plots this as a line on a spot bar chart together with the macd signal lines and a histogram of their value. Ashley includes a histogram of the difference of the macdrev in the latest price. This virtually replicates the other histogram. Based on the macdrev an example of $/yen moves away from the direction of the market. As the pace slows the macdrev will move towards the spot price. The result is that macdrev is considered as a stop loss. Ashley concludes that knowing the actual level at

119


120

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

which this indicator will change allows the analyst on a potential macd cross over. Adcock concentrates on the pattern names of Japanese candles. Open and close price relationship is important in terms of sentiment. One problem with momentum based indicators are that they highlight overbought but prices continue to rise. Under a stochastic momentum prices rise within a trend towards the high range. For the longer period, prices trend until the current closing levels begin to move away from the extreme of the last x day range. This reflects a change in sentiment. To aid sentiment indicators trading position reports are available, however, there is a delay before reporting. Adcock offers a short run objective solution. The opening price is the first reference to market value. The closing price distance confirms the market sentiment. A Doji candlestick indicates indecision in the market. Perhaps a consolidation is due. A number of scenarios are considered. A black down candle ends a series of white up movements. The market has turned bearish. A sentiment change of consolidation. A white up candlestick is posted and the close is on the highest price. Previous periods have seen selling off at the close. Adcock suggests that a number of reversal patterns work well with candlesticks. A shooting star indicates a bearish reversal pattern. A strong directional up with limited correction then when reversal forms price initially gaps higher but rejection follows. An engulfing pattern occurs with a 2 day reversal. More extreme is a morning evening star pattern with a 3 day bullish bearish 3 day reversal. Gandolfi et al present a more scientific/engineering based article. One of the major limits in technical analysis is to supply an operational signal for all market conditions. A variable MACD should perform well in all markets. VIDYA (variable index dynamic average) is constructed. It reduces the moving average time period when volatility increases and vice versa. The Chaude Momentum indicator (CMO) has a range between +1 and -1. it is used for volatility measuring. CMO = (Su -Sd)/(Su +Sd) Where Su = sum of movements up in previous n periods Sd = sum of previous down in Previous n periods Moves up is the difference between 2 price bars with current price higher than the previous close. VIDYA(t) = A * K * Ct + (1-A *K) * VIDYA(t-1) Where Ct = current closing price K = CMO A = constant. VARIABLE MACD = (VIDYA12) - (VIDYA 26) The results contrast MACD with variable MACD in 2

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

markets (NIKKEI 225 - MIB30). For both markets the number of trades required is between ¹⁄ ³ to ¼ with the variable MACD. Performance profit is double on the Nikkei and a turnaround performance from loss to profit in MIB30. The maximum drawdown is reduced by ¹⁄ ³. Bollinger bands were devised to eliminate the determined proper band width for any given time series. They are based on 2 standard deviation from the moving average so they are a forward standard deviation. The upper band is a resistance and the lower band a support. The overbought region is beyond the upper band the oversold is below the lower band. This percentage envelope determines the band within in which prices tend to oscillate. Bollinger bands are the most often used in public chart services. Bands are self-adjusting becoming wider with increased volatility. Because price action is not stationary and non-random it cannot follow statistical properties precisely however it provides a good estimate.


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

The Momentum effect - Puzzling but True Deborah Owen

Article originally featured in Market Technician 61 (June 2008)

The debate as to whether stock markets move in trends or trace out a random walk has tended to divide along professional lines. Most academics subscribe to the efficient markets theory which is predicated on the belief that, at any given time, share prices incorporate all known information about a company. The next move in the share price is, therefore, as random as tossing a coin it could just as easily go down as up - and has nothing to do with how the price has performed recently. Almost all market professionals, on the other hand, tend to embrace some form of trend-following, or momentum technique when managing their portfolios. Anthony Bolton, the well-known contrarian investor, for example, finds momentum analysis is useful in determining both the timing and size of his trades. It is also noteworthy that the relative newcomers to the investment community - hedge and quant funds - rely almost exclusively on momentum-based models to give trading recommendations.

Some academics have done research into “persistent regularities” in the market but more with a view to explaining them away than to challenge the underlying assumption that share prices move irregularly. So the tendency for market professionals to lean towards the momentum style of investing has - until now - been based on empirical experience rather than the practical application of theoretical research. However, a new study by a distinguished triumvirate of authors based at the London Business School gives an academic underpinning to trend-following techniques. Elroy Dimson, Paul Marsh and Mike Staunton confess that “Momentum, or the tendency for stock returns to trend in the same direction, is a major puzzle. In well functioning markets, it should not be possible to make money from the naïve strategy of simply buying winners and selling losers.” Yet this is just what their research does show.

Chart 1: Value-weighted cumulative returns for the UK equity market, 1956-2007

Source: Dimson, Marsh and Staunton (ABN AMRO).

This chart shows value-weighted returns for winner and loser portfolios from the main market (ex-investment companies), defined with breakpoints at the 20th and 80th percentiles. The shaded area is the cumulative difference between winners and losers, and measures the value of a long-short WML portfolio. The momentum process followed here is a 12/1/1 strategy.

121


122

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Chart 2: Annual value-weighted momentum portfolio returns for the Top 100 UK equities 1900-2007

Source: Dimson, Marsh and Staunton (ABN AMRO).

This chart shows value-weighted returns for winner and loser portfolios among the Top 100 equities, defined with breakpoints at the 20th and 80th percentiles. The shaded area is the cumulative difference between winners and losers, and measures the value of a longshort WML portfolio. The momentum process followed here is a 12/1/1 strategy.

The study, published in the ABN Amro’s latest Global Investment Returns Yearbook 2008, is the most wide-reaching of its kind. The main focus is on the UK market and uses data from the start of 1900 to the end of 2007. But the authors also investigated the efficacy of systematic momentum investing in the other major global stock markets. For most of the markets the data spans from 1975 to 2007 but, in the case of the US, it goes as far back as 1926. Shares were ranked according to their performance over the past 12 months and, from this ranking, two portfolios were created. One consisted of the top 20 per cent of companies over the period and the other the bottom 20 per cent. The top performing stocks were then bought while the bottom 20 per cent were sold, giving a zero investment portfolio. The difference between the winner and loser portfolios represents the alpha of the strategy. The portfolios were recalculated every month allowing a one-month skip period.

Returns from 1955-2007 Initially, the authors analysed the performance from 1955-2007 for the overall UK market on a value-weighted basis. As Chart 1 on the opposite page shows, the portfolio of top performing stocks yielded annual returns of 18.29 per cent over the period while that of the losers generated only 6.79 per cent (the middle 60 per cent rose by 13.21 per cent). These figures compare with an overall market return of 13.52 per cent and a 7.81 per cent annualised return on UK treasury bills. The cumulative difference between the winners and losers was 10.8 per cent. Other research by the authors shows that since 1955 smaller companies have produced higher returns than the larger

capitalisation stocks. It is not, therefore, surprising that the returns on an equal-weighted portfolio (which gives as much weighting to smaller portfolio constituents as to the larger ones) is higher than those on a value-weighted portfolio. The returns were as follows: Top performing stocks: Middle performing stocks: Bottom performing stocks: Winners minus losers:

25.5% 17.9% 12.2% 12%

Returns also vary according to the universe of stocks from which the winner and loser portfolios are drawn. Although still large, picking stocks from the top 100 companies each year produced a lower return than that obtained by using the overall equity market.

Returns from 1900 to 2007 Analysis of the top 100 stocks going back to 1900 confirms the momentum phenomenon. The portfolio of winners produced an annualised gain of 15.2 per cent (on a value-weighted basis) while the investment ‘dogs’ produced an annual return of just 4.5 per cent. The cumulative difference between the winners and losers is 10.3 per cent.


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

Chart 3: Monthly momentum returns in 17 stock markets up to 2000 and 2001-2007

Source: Dimson, Marsh and Staunton (ABN AMRO), Griffin, Ji and Martin (2003), and Thomson Financial Datastream.

This chart shows the winner-minus-loser (WML) return from a 6/1/6 momentum strategy, following the methodology described in Griffin, Ji and Martin (2003). The breakpoints are the 20th and 80th percentiles. The Griffin, Ji and Martin sample period begins in 1975 (or, for a few countries, a different year) and ends in 2000. The subsequent period runs from start-2001 to end-2007. Data is from the LSPD (for the UK) and Datastream (other countries).

A global phenomenon The authors extended their analysis to 17 other countries. In the period up to the end of 2000 in all but one country - Sweden - the winners-minus-losers return was positive. In the period post 2000, the US was the only country where the returns were negative due to the sharp reversal after 2002.

Implementing a momentum strategy The high turnover of the momentum investment strategy involves considerable transaction costs and these will obviously chisel away at overall returns. To mitigate transactions costs, the holding period can be extended but, even with doing this, it may still be difficult for investors who face high dealing charges (such as individual investors), to achieve a net profit.

Conclusion This study shows clear evidence that the momentum effect persisted throughout the twentieth century and continues to hold good. As the authors point out, “Those who assert it is a chance pattern, or that it was a passing episode, are taking positions that are counter to the evidence”. Members of the STA will no doubt readily agree with their conclusion that “based on our evidence, it would seem that active managers, who ignore the momentum effect, do so at their peril”.

123


124

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

The MACD Reversal Indicator Gerald Ashley

Article originally featured in Market Technician 14 (July 1992)

The Moving Average Convergence/Divergence Indicator developed by Gerald Appel has a wide following amongst technical analysts. It has proven ability to pick changes in trend and when viewed as a histogram is also good at showing the first signs of a move fading. In his paper on MACD*, Appel included a formula that allows you to determine the future price level at which the macd line will cross the signal line.

through the spot price. It may be of some value to consider the macdrev line as a form of stop/loss tripwire. However at the start of moves the line will move a long way from the price action, and will give a very loose protection level. Once the price move slows and the macdrev squeezes in towards the spot price it may well make good sense to use it for stop/loss purposes. Chart 1

This is expressed as follows:Reversal level = Signal-0.85*0.15 exponential + 0.925*0.075 exponential (0.15- 0.075) As a Teletrac user I have entered the following instructions to reproduce this reversal formula. exp_ma

exp_ma (last, 12.33)

exp_ma2

exp_ma (last, 25.67)

macdrev (Signal-(1-0.15)*exp_ma+(1-0.075)* exp_ma2)/(0.15-0.075) The exponential moving averages of 12.33 and 25.67 days equate to the two smoothing constants of 0.15 and 0.075 respectively. This can be shown with the following formula:-

2-

b b

i.e.

2-0.15 0.15

=

12.33 days

Knowing the number of days, the formula to calculate the beta value is:-

2 1

i.e.

2 12.33+1

=

0.15

The macdrev figure will give the exchange rate at that moment in time (i.e. using the latest spot quote), at which the macd/signal lines would cross over. I find it useful to plot this as a line on the spot bar chart, together with the normal macd/signal lines and a histogram of their values. In addition I have included a histogram of the difference between the macdrev value and the latest price. This virtually replicates the other histogram. From the weekly US$/Yen chart below you will notice that at the start of the trend change the macdrev will move AWAY from the direction of the market, and as the pace of the move slows, it will whip round and start closing back towards the spot price. At the point of crossover in the macd/signal lines the macdrev cuts

In many ways the best use for this reversal value lies in knowing the actual level at which the indicator will change, it allows the analyst to concentrate on markets where a potential macd/signal cross is within “striking distance” of the current price. I have only investigated using this indicator for currencies, where weekly values seem to be particularly valid. As always with technical analysis it is this ability to plan ahead that is of the most value. * The Moving Average Convergence-Divergence Trading Method by Gerald Appel Scientific Investment Systems, Inc. 62 Wellesley Street West, Toronto.

Gerald Ashley is a sought-after speaker, advisor, broadcaster and writer on change, risk and decision making. Covering risk analysis, strateg y, behavioural economics and change management, he speaks on topics that affect any business at any level.


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

Momentum, Trending and Sentiment Richard Adcock

Article originally featured in Market Technician 58 (March 2007)

Throughout my years within the financial industry and, in particular, my involvement with technical analysis, it’s become apparent to me that there are three distinct areas that dominate my market appraisal - sentiment, momentum and trending. I have found that a consistent approach to each gives me a reliable insight into where the market stands at the present and, more importantly, where it is likely to go over the coming days, weeks and even months.

Sentiment It is always difficult to develop a true feeling for sentiment, as it often comes down to an individual’s own subjective reading of how a market has been trading (which in turn actually reflects how that person is positioned). Too bearish - and the directional risk is likely to swing to the upside; too bullish - and more often than not, you have to think very carefully about being long. True, a number of services are available (the traders’ positioning report by the CFTC and Stone McCarthy, to name but two), however, using these takes time to allow the data to be collated and reported. For example, consider the scenario where data is collected for a period ending on a Tuesday, yet won’t be published until the following Monday and an event such as the U.S. payroll number or another important release occurs during that “between” time. Again, it comes down to a subjective reading of how the market has traded between data collection and release. So is this data really all that useful in determining sentiment and, as such, directional risk?

Candlestick analysis, with pattern names such as Piercing Line, Morning Star and Doji, initially appears quite confusing to a beginner. But it really is very easy, and best of all, provides a very objective approach to all markets (not just rice!). If the criteria of a particular reversal aren’t met, then it isn’t a reversal. The candle itself simply represents the difference between the opening price on any given day and the closing price. If the day closes higher than the opening, then the body of the candle is white; if it’s lower the candle body is black. The day’s high and low are marked by vertical lines extending above or below the body (the “shadow”). Why is the open/close relationship so important in terms of sentiment? Quite simply, the opening price represents the first opportunity to trade, and gives us the first reference of what the market feels is ‘value’. The close is the last reference, and the further the closing price is from the open, the more significant that day is in terms of bullish or bearish sentiment. The only other type of candle is a Doji, which occurs when there is no difference between opening and closing prices, so no body is evident. This highlights trader indecision as to just where directional risk lies, and often occurs after a strong move either up or down, signalling that a consolidation is due. Chart 1: Types of Candlesticks

I firmly believe that these services are helpful for the longer term, since why should a few data releases change a developing/ on-going trend in positioning that can’t be picked up at the next publication date. However, my own analysis is mostly short term (3-7 days), which is why I have a problem with these particular positioning services. So does it again come down to only a subjective reading of how the market is positioned - which can work for a while but never on a consistent basis - or is there a more objective way to approach the subject of sentiment? For me, the best measure of day-to-day sentiment is to look at how the market is actually trading and set clear objective criteria to price action. Get that right and at least half the battle is won. To do this, I use Japanese candlestick charts.

Candlestick Analysis The Japanese were the first to use candlestick analysis, in trading the rice futures market. In the 1700s, a Japanese trader named Homma established that, while there was a supply:demand link for rice, prices were also influenced by the emotions of the traders. Homma realised that he could gain an edge for his trading from understanding how emotion helped predict future price movements.

Also important when gauging sentiment within a candlestick chart is how a market closes in relation to the day’s high or low. A trader always has a decision to make coming into the close, whether to hold a position overnight (or the weekend). If one is confident that a position is right and that the market will continue to move in one’s favour, there’s no position change and the trader looks for the trend

125


126

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

to remain in force. If this feeling of confidence is reflected across the market, no substantive closing of positions will develop, so the settlement will be towards the day’s high or low (depending on the general trend). But consider the situation if that same decision process resulted in traders deciding they were losing confidence in their view, and thus felt it prudent to close positions before departing for the day. In such a case, buying or selling (depending on market positioning) will materialise coming into the close, causing prices to either sell off from the day’s high or rally from the low. If the market has shown a strong directional move with large candle bodies and settlements towards the limits of the day’s range - but then closing prices suddenly start to fall back from the highs - we can assume that bullish confidence is waning, increasing the risk of a more prolonged consolidation phase. [NOTE: I review this further within stochastic momentum tools.]

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Chart 3: Ending the Consolidation

Chart 2: Sentiment Turns Bearish

posted and the close is actually the session high, so unlike the previous eight days there’s not been any selling into the close, reflecting a clear change from the neutral sentiment to a much more bullish position, from which fresh price strength developed. So we can see that, by studying the open/close relationship and how the day’s representative candlestick body compares to either the high or low, we can get a very good feel as to whether sentiment is changing. It should also suggest if we should begin looking for either a consolidation phase or a resumption of the ongoing directional trend (of course, that doesn’t even begin to look at reversal patterns that can form).

Candlestick Reversal Patterns Bearing all of this in mind, consider the examples in Charts 2 and 3 (where sentiment changes have been flagged by how the market closed in relation to the session high); these gave important signals to just where the directional risks lay for the longer term. In Chart 2, at the May 2nd low, no bullish reversal pattern was evident, so the view was that price strength was limited before the overall bearish trend was resumed and new lows scored. Price action went on to post five consecutive bullish white candles as the consolidation developed and traders’ perception of ‘value’ was higher at each close compared to the opening price. That said, it’s clear from the actual size of each candle body that sentiment was never aggressively bullish. Now look at the blue arrow-highlighted day when prices actually opened on a bullish gap higher, then hit a new recovery high. At that point there was nothing to indicate the consolidation/rally was ending, until, for whatever reason, selling pressure materialised, the high was rejected, and the closing price was below the day’s opening price - thus sentiment had turned bearish again with a black candle posted, breaking the pattern of white candles. The more bearish themes were developed further the following session, when an opening gap lower developed, confirming the negative sentiment and resumption of the ongoing bearish trend. In Chart 3, we see how this type of sentiment monitoring can signal the end of a more balanced sideways trading range, and prompt a new, higher, aggressive trade. For the initial phase of the highlighted area, it is clear that no dominant force is in place (the candle bodies’ relatively small and long shadows above and below the open/close relationship reflect rejection of the attempted higher and lower prices; i.e., sentiment is balanced and prices are moving sideways). However, on the final day of the highlighted area, we can see that a large white candle has been

There are a number of reversal patterns that work well on candlestick charts; each can be very important in highlighting a sentiment and directional change over a period of just 1-, 2-, or at the most 3-days. I won’t review every pattern, but I will discuss one pattern from each of these short-term periods, explaining why I see them as critical in flagging trend changes. For anyone wanting to look at candlestick patterns in greater detail, I recommend Steve Nison’s book, Japanese Candlestick Charting Techniques1. Chart 4 shows a bearish reversal pattern - it means the market must have been previously trading within a strong directional up

Chart 4: Shooting Star Pattern - Bearish 1-Day Reversal


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

trend, with limited corrections, new highs being consistently posted and, predominately white candles - all reflecting the ongoing bullish sentiment. On the day the reversal forms, everything appears as if the trend is being extended - prices have gapped higher at the open, and further strong support developed to post a new price high. However, at some time during the day, a rejection of this new extreme develops and prices sell off, leaving a large shadow above the open/close relationship, which in turn is within the lower third of the day’s entire range. It doesn’t matter if the market closes lower than the day’s opening trade (thus the candle body can be black or white); the activity highlights a clear rejection and shift in sentiment. If that is confirmed by a black candle the following day, this marks the end of the bullish trade and the start of at least a more extended consolidation/correction phase.

Psychology and Markets

Systematic Trading

Chart 6: Morning/Evening Star Pattern - Bullish/Bearish 3-Day Reversal

Chart 5: Engulfing Pattern - Bearish 2-Day Reversal

sellers fail to extend the trend lower, and the open/close relationship is relatively small. The third day of the reversal is undoubtedly bullish, the session opens on a gap higher (bullish sentiment; traders want to buy as soon as possible) and a strong rally ensues (thus a large white candle that closes above the midpoint of the first day’s open/close relationship, confirming the pattern is in place and that the market is setting into a new bullish trend).

On Chart 5 as is always the case with any bearish reversal pattern, the market had been trading within a bullish and consistent trend. The first day of the pattern appears like any other within the uptrend, it’s a bullish white candle and a new recovery high is scored. The second day then sees prices beginning the session in a bullish way (with an opening gap higher, and possibly but not necessarily a new high), but a rejection develops during the day and prices sell off, closing lower than the opening trade (thus a black candle body) as well as below the prior day’s opening price - so this open/close relationship has been completely ‘engulfed’. Clearly, this is a very quick shift in sentiment, going from bullish on the open to bearish at the close, and it represents the end of an up trend and a much deeper, prolonged sell-off. For the pattern to be a bullish engulfing reversal, - the market must have been trading in a downtrend. The first day of the pattern thus has a black candlestick body, and prices are making a new low. The next day’s opening price is lower on a bearish gap, followed by a strong rally that closes above the previous day’s opening price, creating a white candle and engulfing the first day’s open/close range. Again, for any candlestick reversal pattern to be valid, the market must have been trading within a clear trend, as the smaller the recovery or correction, the less significant the signal. This is just as true (if not more so) for the Morning Star pattern, which is a 3-day reversal, signalling a more important shift in long term sentiment. As with the prior examples, in the case of the Morning Star pattern, the first day’s chart appears as a normal continuation of the directional trend, with a new low scored and a black candle posted, as sentiment remains negative with the first reference of ‘value’ lower at the close versus the open. However, on the second day, the market opens on a gap lower (maintaining the bearish sentiment) - but we begin to see sentiment changing, in that

The bearish version of the Morning Star pattern is perhaps not too surprisingly called “Evening Star”, and is a mirror image of the bullish pattern. The first day is a large white candle, maintaining the drive to new highs within the bullish trend, followed by a gap higher on the second day and a small candle body, as sentiment begins to show the first sign of change. The third day of the bearish pattern leaves no doubt that the directional shift is complete, with a gap lower followed by strong selling pressure that creates a large black candle settling below the mid-point of the first days’ open/ close relationship. [NOTE: While some technical analysts don’t require it, in any reversal pattern within candlestick work, I have to see the following day to confirm the sentiment change (via a white candle following a bullish reversal and a black one after a bearish pattern). If no confirmation is seen, I declare the pattern invalid and of no significance to the daily and weekly view.]

Momentum I consider momentum-based indicators extremely valuable in determining important turning points for markets, but they can be quite infuriating, given their habit of highlighting overbought readings, yet the market continues to stay that way, with prices continuing to power ahead. In certain circumstances I would even argue that ‘overbought’ momentum readings are actually a buy signal, as they often confirm that sentiment is bullish (which is what will drive the market higher). That said, my use of momentum tools is based more from a timing perspective, with any cross up or down often an important signal, whether it be against ‘overbought’, ‘oversold’ or neutral readings.

Stochastic Momentum Within his work, George Lane2 highlighted that as prices rise (or fall) strongly within a bullish or bearish trend, closes tend to be towards the high or low of the day’s range. This reflects that sentiment remained strong (traders not looking to take profits coming into the close, and content to carry risk overnight). Stochastics extend

127


128

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

this approach to a slightly longer period by taking the latest price and comparing it to the range of the last X number of days (X = any value; for me, 10 days gives the best signals). So as prices move within a directional trend, the stochastic indicator line will rise or fall, until the current closing level begins to move away from the extreme of the 10 day range, reflecting changing trader sentiment and a recovery/correction about to develop from oversold/overbought readings.

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

can be useful, it is often late, hence I tend to use crosses in the MACD as confirmation signals, preferring to use the MACD indicator more to signal the direction of the main trend. So if the MACD is >zero the market is in a bullish trend; if it’s <zero, the market is trending bearish. Chart 8: The MACD Indicator - Signals Can be Lagging

Chart 7: The Problem with Momentum Indicators (in this case Stochastics)

Momentum/Trending Relationship However the main problem with momentum indicators is that they can get over-extended and stay that way for prolonged periods as prices maintain the ongoing directional move (Chart 7). So I am not really interested in the crosses lower from high stochastic readings, or the turn higher from low levels, as these could simply reflect a period of noise as the indicator fails to unwind overextended conditions as we know can often happen. Instead, what is even more important to me is when the indicator crosses back to the upside in a bullish trend (or to the downside in a bearish one), which signals that the correction against the main trend is ending, and the original directional move is about to be re-established, thus presenting a strong buy/sell signal.

Trending In theory, it should be easy to tell if a market is trending bullish or bearish, but that’s not always the case. To help establish where risks lie, a number of trending-based indicators are available to the technician. In my opinion, the best for ease of use and confirmation of the directional trend is “Moving Average Convergence Divergence indicator” (MACD). Used alongside the stochastic momentum, the MACD can provide some very useful signals.

MACD Trending Indicator This indicator measures the gap between a short-term and a longer term (I use 10- and 20-day) exponential moving averages, and delivers an indicator line (the blue line on Chart 8) together with a 3-day average of the indicator itself (the red line on Chart 8). As prices move aggressively in one direction or the other, the 10-day average will follow action much faster than will the 20-day, so the gap between the two averages widens. Some use the MACD in a similar manner to the stochastic; a cross lower from ‘overbought’ readings is a sell, while a cross higher against ‘oversold’ conditions is a buy. I find this difficult, as any moving average based indicator is ‘lagging’, so price movement is required to actually see a signal. Therefore, So while the MACD

Thus, we have an indicator that gives good directional signals but lags market movement, taking time to turn higher or lower (the MACD) and one that is not always reliable at highlighting overbought/oversold signals, but shows when a consolidation has ended and the directional trend set to be established again (the stochastic). What if we combine the two; maybe the problems of each will cancel out. So we use the MACD to confirm in which direction the market is trading (any cross from negative to positive will give a buy signal, and from positive to negative a sell signal) and the Stochastic to time when consolidations are over (if momentum turns higher with the MACD >zero it’s a buy signal, or lower with the MACD <zero, it’s a sell). In other words—we are buying when momentum turns up in a bullish trend and selling when it turns down in a bearish one, which I believe are the strongest and most reliable technical signals we can see. Let’s look at Chart 9 (below) in closer detail. The important thing Chart 9: Buy and Sell Signals Using the Momentum/Trending Relationship


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

to remember in using the momentum/trending relationship is that if Stochastics cross, we filter it against the MACD. If momentum crosses down and the trending tool is >zero we ignore the signal completely; if the MACD is <zero it’s a new sell signal (red arrows on Chart 9). On the flip side, if the stochastic crosses up with the MACD <zero that is ignored, but if the trending indicator is >zero, it’s a buy signal (blue arrows on Chart 9). The brown arrows highlight when the MACD crossed its own moving average to confirm a similar turn in stochastics (in other words, when we have both momentum and trending measures highlighting the risk of a more extended consolidation/recovery phase). In this situation, we do not want to hold risk, and sell all positions, waiting to see how the situation develops. Either the MACD will cross zero to give us a signal, or stochastics will be crossing up or down (confirmed by the MACD being > or < zero) to highlight that the directional trend is about to resume. December 2006 - January 2007 was a very interesting phase for markets, with prices under significant pressure for no obvious fundamental reason. Using the momentum/trending relationship, we actually saw our combined approach give some strong signals. The three long recommendations seen during the late October/ early November 2006 consolidation period (the last 3 blue arrows on chart 9 when the MACD crossed >zero on Oct 31st followed by Stochastics crossing higher on Nov 10th and Nov 21st) were closed when the stochastic cross was confirmed by a turn lower in the MACD on Dec 7th 2006, we had the indication that a more extended consolidation/correction was about to occur, so we closed the long recommendations. We didn’t recommend shorts, since we didn’t know at the time if the correction would be limited before the up trend extended again (signalled by momentum crossing up with the MACD >zero) or a new downtrend was about to develop, reflected by the MACD closing <zero, to give us a new shorting signal on December 18th at 108.12).

Psychology and Markets

Systematic Trading

Don’t forget that any indicator is a derivative of price and so I use Candlestick analysis alongside the momentum/trending relationship. As such, a candlestick reversal outweighs momentum/trending conditions. By that, I mean that if a bearish engulfing pattern is formed and confirmed, I would reverse long positions to go short, even if momentum is rising at the same time, as trending tools are bullish. The sentiment change reflected by the reversal will be mirrored, in time, by the indicators.

Summary There is a confusing array of technical tools and techniques widely used within the market these days. Most are very successful, but at the end of the day, it is the investor’s responsibility to assess what suits his trading style and what can be trusted to deliver the type of signals to follow with confidence. I have found that a consistent approach combining sentiment measures with momentum and trending tools can help establish the trend direction, when to add risk within that trend, and when to reduce exposure and actually sit on the sidelines because I simply don’t know which way the market is going. In that latter situation, I would rather wait for confirmed signals than make a coin toss (50/50) call.

Japanese Candlestick Charting Techniques: A Contemporary Guide to the Ancient Investment Techniques of the Far East, by Steve Nison (who is widely considered the expert on candlestick charting in the Western world). 1

George Lane is known as the “Father of Stochastics” for his work on the stochastic oscillator. 2

Subsequent bounces in price prompted the stochastic to cross higher, but without these confirmed by the MACD (which continued to fall <zero), each following cross lower in momentum gave us the signal to add to short recommendations - on Dec 22nd at 108.05; Jan 11th at 107.10; Jan 23rd at 106.285 - as we were confident each time the downtrend would continue having seen momentum turn bearish again within a downtrend. We established a total of 4 short recommendations within that particular sell-off, only closing all positions on February 1st 2007 at 106.215, when the MACD finally crossed above its own moving average, following the stochastic signal seen earlier on January 30th. The idea behind the momentum/trending relationship is that it keeps you in a position and adds to it even though momentum can be at overbought or oversold readings. We all tend to question how sustainable a move will be, on the ill-conceived perception that overbought prices means strength can’t continue, and oversold prices highlight that selling pressure will end and a rally is about to occur. But this just simply is not the case. By following these type of signals highlighted in the table below, we can still add to positions when we have confirmation that the trend will continue, even though prices are ‘over-extended,’ and when to stand aside and close all positions because we simply don’t know how long a consolidation is going to take, or if it is going to turn into a trend reversal. Signal MACD closes >zero Stochastic crosses down with MACD >zero Stochastic crosses up with MACD >zero MACD crosses down with stochastic falling MACD closes <zero Stochastic crosses up with MACD <zero Stochastic crosses down with MACD <zero MACD crosses up with stochastic rising

Action BUY signal Do nothing BUY signal CLOSE all positions SELL signal Do nothing SELL signal CLOSE all positions

Richard Adcock, Director Fixed Income Technical Strateg y, UBS Investment

129


130

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Variable MACD: Adapting to Financial Market Dynamics (left to right) G. Gandolfi, M. Rossolini and A. Sabatini

Article originally featured in Market Technician 57 (December 2006)

This article is a short summary of a paper presented at the IFTA Conference, Lugano, 2006

Abstract One of the major limits of technical analysis tools is its ability to supply operative signals in all types of market conditions. Lagging Indicators are efficient primarily in trending markets, whereas leading Indicators exhibit better performance in sideways market conditions. This article introduces a new technical instrument - Variable MACD. The advantage of Variable MACD is that it demonstrates high efficiency both in trending and non-trending markets. The method for constructing Variable MACD is shown below. Empirical results and considerations based on historical data of major indices and financial time series data are demonstrated and show encouraging results: Variable MACD is not only more efficient than MACD, it also performs very strongly in a range of market conditions.

Introduction In order to better appreciate the effect of Variable MACD on financial markets time series, it is important to outline the concept of phase lagging and leading. A wave, and in general, any dynamic series whose motion is oscillatory, is defined by the following three parameters: 1. AMPLITUDE: measure of the height or intensity of the wave; 2. PERIOD: measure of the time interval between two minima (or two maxima); 3. PHASE: measure of the positioning along the time axis of two minima.

Figure 1: Leading and Lagging Curves

Leading and Lagging Curves

SIN(t) LAGGING Curve

COS(t) LEADING Curve


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

Given the time series x = x1, x 2, x3,...xn, the time series moving average (MA) is defined as: N

1

MA =

Sx

N

i

i=1

MA is a lagging indicator. On the other hand, the Rate of Change (ROC) indicator, defined as ROC = [100(x2-x1)/ x1]% is a leading indicator. Mathematically, for a function ƒ(t), if every occurrence of t is replaced by t-a or t+a, “a” being a constant, ƒ (t-a) is a lagging function with delay “a”; ƒ (t+a) is a leading function with lead “a”. Operators, also, may affect the phase. For example, integration has lagging effects, whereas differentiation has leading characteristics.

Method The two indicators used in the construction of Variable MACD are the MACD (Moving Average Convergence Divergence) and VIDYA (Variable Index Dynamic Average). The aim of the VIDYA dynamic indicator is to automatically reduce the moving average time period when volatility increases and automatically increase the moving average time period when volatility decreases. In this work, volatility is defined as a measure of trending vs. trading range behaviour. High volatility indicates a strong trending market. By contrast, low volatility indicates the market is range bound. Market Volatility can be measured by the use of Chande’s proprietary indicator called Chande Momentum Indicator (CMO) as follows: CMO = [(Su - Sd)/(Su + Sd)]; where: Su = Sum of Moves up of n previous bars; Sd = Sum of Moves down of n previous bars; Moves up = difference between two price bars when current close is higher than previous close; Moves down = difference between two price bars when current close is lower than previous close; The CMO indicator ranges between +1 and -1 The absolute value of the CMO will, therefore, vary as follows: Abs(Cmo) approaching 0 when there is Low Volatility Abs(Cmo) approaching 1 when there is High Volatility Vidya is defined as: Vidyat = * k * Ct + (1- * k) * Vidyat-1

a

a

where: Ct= Current Close; K = Volatility Indicator (|CMO|); = constant determined by Analyst ( = 0.5); Vidyat-1 = Previous-bar-Vidya.

a

a

True Vidya Length, N, is: N = (2-k* ) / (k* )

a

a

Inferring that If |CMO| ➝ 1 Then N decreases; If |CMO| ➝ 0 Then N increases;

VARIABLE MACD = Vidya[12] - Vidya[26] = ( * |CMO| * Ct + ((1- ) * |CMO| * Vidya[12]t-1) - β *|CMO|* Ct + (1- β) *|CMO|* Vidya[26]t-1)

a

a

where: = 0.154; β = 0.074;

a

|CMO| = Absolute value of CMO; Ct = Current Close; Subscript t-1 indicates previous bar;

a =2/(N+1) with N=12, a=0.154 β =2/(N+1) with N=26, β =0.074

131


132

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

To calculate the True Length, N, we have to consider volatility: N12 = (2 - |CMO| * ) / (|CMO|* ) for 12-day-Vidya N26 = (2 - |CMO| * β) / (|CMO|* β) for 26-day-Vidya

a

a

If |CMO| ➝ 1 Then N12 ➝ 12 N26 ➝ 26 If |CMO| ➝ 0 (i.e. 0.005) Then the two Vidya will have a higher Length N12 ➝ 2596 N26 ➝ 5404

Interpreting the Signals Signal interpretation is the same as for MACD. The advantages of using Variable MACD are the following: • • • •

False signal reduction; Good trading range management; Less frequent trading; Better performance than MACD.

The major disadvantage is exhibited by a delay in triggering the beginning of a new trend.

Results Signal

Number of Trades

Performance (%)

Max Drawdown (%)

MACD on NIKKEI 225

12

+29.71%

-15.31%

VARIABLE MACD on NIKKEI 225

3

+49.76%

-6.00%

MACD on MIB30

18

-21.92%

-22.06%

VARIABLE MACD on MIB30

5

+10.52%

-7.39%

Figure 2: NIKKEI225 MACD trade signals (vertical grey lines)

Figure 3: NIKKEI225 Variable MACD trade signals (vertical grey lines)

Conclusion Variable MACD adapts better to market conditions. It modulates the phase lagging and leading effect efficiently.

• G. Gandolfi, Professor of Financial Markets and Institutions - University of Parma • M. Rossolini, Ph.D Candidate in “Banking and Finance” Tor Vergata University Rome; Researcher and Lecturer - University of Parma • A. Sabatini, Electrical Engineer (MIT B.S. - MS), Portfolio Manager - MIT EC, Finbest, CEO, CFO, Florence.


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

Bollinger Bands John Bollinger

Article originally featured in Market Technician 17 (July 1993)

Trading bands are simply mirror images of an average shifted up and down on the screen by some given percentage. It was immediately apparent to me these bands were extremely useful, but there were several drawbacks.

obtained by answering 2 to the prompt.

First, you had to determine empirically what the proper band width was for the price series under study. Also, you could not easily vary the widths of the bands to suit changing market conditions. Despite these problems, trading bands were and are a very useful concept. The task I set for myself was to improve generic trading bands. The result, after a long period of research, is Bollinger Bands.

There are several different approaches to using these bands. The following are several examples. (All examples assume a 20day simple average and 2 sigma Bollinger Bands).

Bollinger Bands eliminate both of the above problems. First, the band width is determined by the volatility of the underlying data series. Second, the band width varies over time as the volatility of the data series varies. Bollinger Bands refer to bands drawn above and below a simple moving average whose distance is two standard deviations (sigmas) from the average. For each point on the average, the standard deviation of the data (the number of data points used equals the length of the average) is calculated. This number determines the distance of the bands from the average. As a moving average moves forward in time, so does the standard deviation calculation for Bollinger Bands. Thus it may be thought of as a moving standard deviation of the same length as the average.

(Editor’s Note: this refers to the prompts in CompuTrac software) are most useful by far.

1. The upper bands may be thought of as resistance zones and the lower bands may be thought of as support zones. 2.

Moves beyond the bands can be thought of as “thrust” or “power” moves and therefore likely to continue. A corollary of this is that tops or bottoms are very unlikely to be made outside the bands.

3. The ability of a data series to stay at the upper (or lower) band for an extended period of time can be viewed as a demonstration of strength (or weakness). 4. When descending (ascending) from the upper (lower) band, the average often provides support (resistance). 5. During a primary move up (down), minor corrections can be expected to stop at the average and major corrections at the lower (upper) band. 6. The market can be thought of as overbought (oversold) in the region of the upper (lower) band.

In order to use Bollinger Bands, first you must graph a data series on your screen. The second step is to choose an arithmetic average and plot it. For stocks and the stock market as a whole, I suggest you start with a 20-day moving average. You will see a pair of lines drawn across the screen, one above and one below the moving average. These are Bollinger Bands. The bands can be drawn at virtually any distance from the average you choose. However, the 2 sigma (Editor’s Note: standard deviations) bands

7.

The bands encompass an area that under “normal” circumstances should contain 95 per cent of all trading activity. The confidence in this assumption can be shown to reach statistically highly significant levels for averages of 30 periods or greater.

8.

Last, and perhaps most important, the bands are extraordinarily useful when combined with several other technical indicators. For example, a second push to the the upper band accompanied by a lower reading on a stochastic or RSI than the reading obtained on the first push (a divergence) could be an important clue to a top.

John Bollinger is editor of the Capital Growth Letter. For more information on Bollinger Bands or for a complimentary copy of his newsletter write to: P.O. Box 3358, Manhatten Beach, CA 90266.

133


134

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

CHAPTER SEVEN

Elliott Wave and Fibonacci Articles in this chapter 136 The Science of Socionomics Robert R. Prechter Jr., CMT (compiled by Deborah Owen)

140 Oil - A real time application of the Elliott Wave Model Robert R. Prechter Jr

143 The British bear market of 1720-1940: A Socionomic Examination Tom Denham

150 The Golden Ratio and its presence in financial markets Tony Plummer

156 Elliott Wave, Chaos Theory and Fractals Julien Camberlin


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

CHAPTER SEVEN INTRODUCTION

Murray Gunn MSTA, CEWA

Introduction “If I have seen further it is by standing on the shoulders of Giants.” So said Isaac Newton in 1675 and the metaphor applies to all fields of research. The history of the Elliott Wave Principle and Fibonacci analysis is a case in point. Ralph Nelson (R.N.) Elliott spent most of his career as a successful accountant and having retired, made a detailed, empirical study of U.S. stock market data in the early 1930s. That study discovered that the market exhibits a series of distinct movements which he called waves. Elliott found that these waves are patterned and, crucially, repeat at every timescale from intraday movements to those that occur over many years. Not only that, with the help of Charles J. Collins, one of the leading investment newsletter publishers, he discovered that the proportionality of these waves are often related to the Golden Ratio of 1.618 (known since the time of Euclid around 300 BCE) and its related number sequence discovered by the 12th century Italian mathematician, Leonardo of Pisa, or Fibonacci as he is famously known. What Elliott had discovered was that the stock market is a dynamic, fractal system. A few decades later, this was proven by the scientific and mathematical community. It is this discovery from Elliott, that the stock market is a robust fractal, which makes the Elliott Wave Principle so important. This is because it inherently contains a forecasting element of both trends and turning points, and one that is totally different to the linear extrapolation used elsewhere. That is what so excited Elliott when, in 1934, he wrote to Collins, saying that he had discovered three novel features of market action: recognition of wave termination, classification of wave degree and time forecasting, which were “a much needed complement to the Dow Theory.” Elliott was one of the first subscribers to Robert Rhea’s “Dow Theory Comment” newsletter and Dow had also used nautical themes such as tide and waves to describe price action. But whereas Dow had theoretically discovered, Elliott had done so empirically. In 1938, The Wave Principle was published. Forty years later, in 1978, Collins wrote in the foreword of Elliott Wave Principle (Frost & Prechter): “Elliott, in developing his theory through observation, study and thought, incorporated what Dow had discovered but went well beyond Dow’s theory in comprehensiveness and exactitude. Both men had sensed the involutions of the human equation that dominated market movements but Dow painted with broad strokes of the brush and Elliott in detail, with greater breadth.” In other words, Elliott had discovered the detail behind Dow’s theory, as well as much more. After his death in 1948, Elliott’s Wave Principle was

Psychology and Markets

Systematic Trading

championed by Hamilton Bolton, founder of the Bank Credit Analyst research firm. Clearly, Hamilton saw that the Wave Principle is a way to track economic cycles, particularly those in credit that wax and wane according to society’s mood. After Bolton, Richard Russell was the leading proponent of Elliott Wave, again clearly seeing a deep link with Dow Theory. But it was the 1978 publication of Elliott Wave Principle by A.J. Frost and Robert Prechter that gave the modern rebirth to Elliott’s work. Since then, Prechter has been the undisputed dean of the Elliott Wave “school,” bringing the subject to millions of people through his company Elliott Wave International (www.elliottwave.com). Robert Prechter has continued the evolution of Elliott’s work by discovering that social mood, the driving force behind market and economic cycles, ebbs and flows according to the Wave Principle. His discovery and development of this brand new social science of Socionomics is providing fascinating links between stock market cycles and trends in politics, popular culture, health and many other fields. Socionomics, born out of Elliott’s Wave Principle, is now providing technical analysis a true gravitas in academia. It has been my experience that, no matter whether they love it or hate it, people always want to know what the Elliott Wave message is. In large part, this is because “Elliotticians” are passionate about their subject and that devotion truly comes across through the authors of the articles in this chapter. Robert Prechter is undoubtedly one of the greatest market analysts of all time, making his initial mark in the late-1970s by using Elliott Wave analysis to anticipate the roaring stock bull market of the 1980s, at a time when the consensus was overwhelmingly bearish. His article, taken from a 2010 presentation, “Oil - A real time application of the Elliott Wave Model” shows just how powerful the Wave Principle can be, especially at major turning points. Bob’s contribution to the subject of technical analysis is unmatched and his article on Socionomics gives an overview of this new and exciting field. Tom Denham’s Socionomic study of Great Britain provides further insight into how social mood drives stock market and economic cycles. I had the honour and privilege of working with Tony Plummer at the beginning of my career, and his work inspired me into technical market analysis. Tony’s unique discoveries mean that he deserves to be considered a giant of technical analysis, alongside the likes of Dow, Elliott and Prechter. His article in this chapter discusses the Golden Ratio and the reason why it governs financial market structure. Julien Camberlin’s article goes a long way to scientifically prove the existence of a link between Elliott waves and Fibonacci ratios. The articles in this chapter give a good flavour of Elliott Wave and Fibonacci analysis but, given the space constraints for the book, can only scratch the surface. I would encourage the interested reader to delve deeper. By doing so, a lifetime of fascination, incredible revelation and insight can be enjoyed.

135


136

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

The Science of Socionomics Robert R. Prechter Jr., CMT (compiled by Deborah Owen)

Article originally featured in Market Technician 50 (July 2004)

Men have tried for millennia to forecast human events. In the long history of social forecasting, the chronic propensity for immense error has resulted from linear thinking, the extrapolation of current trends into the future. This nearly ubiquitous approach is a result of the assumption that laws governing billiard ball behaviour apply to human behaviour. Most people believe that markets and societies share the property of an object in motion i.e. it will continue along a calculable path until some new outside influence - a force or obstruction - alters its trajectory. Successful anticipation of future events is possible. However, it is possible only with the knowledge that human behaviour changes as a result not of external forces but of internal ones. Understanding this dynamic is at the core of socionomics.

What is socionomics? Socionomics is the science of history and social prediction. It examines and forecasts market and social trends from the perspective that political, economic, cultural and financial trends are all the product of the collective human psychology which stems from an unconscious herding impulse in the pre-rational portion of the brain. Understanding socionomics requires comprehending the contrast between the following postulations: 1. The standard presumption: Social mood is buffeted by economic, political and cultural trends and events. News of such events affects the social mood, which in turn affects people’s penchant for investing. 2. The socionomic hypothesis: Social mood is a natural product of human interaction and is patterned according to the Wave Principle. Its trends and extent determine the character of social action, including financial, political and cultural trends and events. The contrast between these two positions comes down to this: The standard presumption is that in the social setting, events govern mood; the socionomic hypothesis recognises that mood governs events.

The Wave Principle Underpinning socionomic analysis is the Wave Principle. This is an endogenous human social dynamic that generates a specific sequence of progress and regress that regulates the complex system of social mood interaction. It is not in the social nature of mankind to accept and be content with stasis. If there is one constant regarding social mood, it is its continuous flux. However, the fact that social mood is everchanging is not, as many would assume, an impediment to forecasting; it is the key to it. Investigation by R.N. Elliott in the 1930s and 1940s yielded the crucial knowledge that social

behaviour changes not randomly but according to a hierarchal fractal pattern. Social mood does have constancy but it is a dynamic one. While the extents of social mood, experiences and conditions vary from time to time and place to place, the patterns of behaviour that lead to a reversal in trend do not. In order successfully to anticipate changes in society reliably, one must understand the consistent pattern of society’s internal dynamics.

The dynamics of the social mood Many people assume that mood forms a feedback loop with events, which in turn reinforces the mood. Some reflection reveals this idea to be erroneous. If events formed a feedback loop with mood, then social trends would never end. Each new extreme in mood in a particular direction would cause more reinforcing actions and those actions would reinforce the same mood, and so on forever. This is an untenable idea. The only feedback loop that must occur involves the synchronization of Elliott waves through human minds. In order to participate in social moods, individual minds must interact with others. All levels of communication media, from face-to-face discussion to satellite television, serve to effect this interaction. This interaction creates the trends of social mood, which stimulate social actions, which are reported as events. These actions and events are end results with no consequences of their own in terms of the waves. This must be so, because Elliott waves exist. Events affect minds to the extent that they may shape specific actions that owners of those minds take, but they do not alter or affect the trends in aggregate mood. The sociological dynamic unfolds regardless of whether humans create a gauge of it for observation and reaction, but when a gauge (such as the Dow Jones Industrial Average) is created and widely observed, people incorporate the gauge itself into the process of mental interaction. It is a conscious reference point for the individual participants and an unconscious reference point for the social dynamic. Consciously rendered changes in the components of the gauge (such as occasionally replacing one or two of the stocks in the Dow) are irrelevant to the more important fact that the unconscious dynamic uses the gauge as a reference regardless of its components. People do not notice the operation of this sociological dynamic because they are not looking for it. Indeed, given the opportunity, they typically reject the idea outright because such dynamics are contrary to the natural assumptions that people make about how societies (and more narrowly, markets) progress and regress. That these processes must be unknown to, or rejected by, virtually all is a prerequisite for their operation. Only people who are blind to the principles that govern the social dynamic can behave so as to produce it, because only then can they be passionate, active participants in it.


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

Chart 1: Elections

Because social mood change, as revealed by the stock market’s form, is patterned according to the Wave Principle, we can propose a larger socionomic hypothesis, that the Wave Principle ultimately shapes the dynamics underlying the character of all human social activity. We will now investigate some presumed “outside forces” from the socionomic perspective.

Elections The standard presumption is that election results are a key determinant of the stock market’s trends. As an election approaches, commentators debate the effect that its outcome will have on stock prices. Investors argue over which candidate would likely influence the market to go up or down. “If so-and-so gets elected, it will be good/bad for the market” we often hear. If this causal relationship were valid, then there would be evidence that a change in power from one party’s leader to another affects the stock market. On the contrary, there is no study that shows a statistically valid connection. A socionomist, on the other hand, can show the opposite causality at work. Examine Chart 1 and observe that strong and persistent trends in the stock market determine whether an incumbent president will be re-elected in a landslide or defeated in one. In all cases where an incumbent was rejected by a landslide, the stock market’s trend was down. In not one case did an incumbent win reelection despite a deeply falling stock market or lose in a landslide despite a strongly rising stock market. The social psychology that accompanies a bull or bear market is the main determinant not only of how voters select a president but also of how they perceive his performance. Correlation with the stock market, consumer confidence, economic performance

and other measures suggests that social mood is by far the main determinant of presidential popularity. What a leader does is mostly acausal with respect to the public’s opinion of him. There are two reasons for this fact. First, his actions, despite their endless analysis in the press, do little to affect his popularity. Second, his popularity is dependent upon a social mood and economy over which he can exercise no countertrend influence. If you are new to these ideas, they may be hard to swallow. Aren’t some presidents fools or rogues and others statesmen? Don’t some presidents affect the economy for good or ill? As to the first question, the answer is, certainly there are presidents of high or low character and ability. However, that does not affect their popularity. For example, President John Kennedy blew the only military conflict in which he engaged the country, attacked the steel industry out of pique to no result, and continually committed adultery. He is revered. Why? Because the country was in a state of euphoria for all but a few months of his term, euphoria that morphed three months later into Beatlemania. The same may be said of President Bill Clinton, whose escapades brought about impeachment yet whose popularity stayed high despite his troubles, peaking along with the stock market in the late 1990s. As to the second question, Republicans claim that the laissez-faire, low-tax policies of President Ronald Reagan were responsible for the economic recovery of the 1980s. The Democrats claim that the social regulatory, high-tax policies of President Franklin Roosevelt were responsible for the economic recovery that brought the US out of the Great Depression. Can they both be correct? No, they’re both wrong. The economy recovered each time because a rising social mood drove the expansion. Policies did not cause the recoveries. Policies do have effects, but they do not change the patterns of social mood. Rather, changes in social mood affect policies.

137


138

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

War

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Chart 3: Demographics

The conventional view is that war affects social mood and the stock market. But this assumption is unsupported by argument or history. As to argument, many people assume that war is a dangerous enterprise that would cause concerned investors to sell. Many historians, on the other hand, argue that war is good for the economy, which by conventional logic would make it good for the stock market. As this reasoning is contradictory, so is the historical record. Chart 2: War

By contrast, socionomics would argue that social mood governs the character of social activity, so a persistently rising stock market, reflecting feelings of increasing goodwill and social harmony, should consistently produce peace while a persistently falling stock market, reflecting feelings of increasing ill will and social conflict, should consistently produce war. Chart 2 showing the US stock market (spliced to the English one prior to 1789) over 300 years bears out this thesis. Major mood retrenchment produces war, as humans finally express their collective negative mood extreme with representative collective action. As with economic output, the size of a war is almost always related to the size of the bear market that induces it. The three biggest wars involving North Americans followed the three largest stock market declines. The Revolutionary War began near the end of the 64-year bear market in British stock prices that started in 1720. The Civil War followed the 24-year bear market that ended in 1859 after posting a 74 per cent decline. World War II began six years after the 89 per cent collapse in stock prices that bottomed in 1932. Long rises in the stock market unerringly result in climates of peace, while sharp declines result in major wars.

Demographics What is the socionomic position on demographic causality? If social mood determines the trend of the economy, politics and the conditions for peace and war, might it not also determine demographics? Chart 3 shows the stock market prices plotted against birth rates from 1909 to the present. There is a fairly noticeable correlation between the two sets of data. Why would births and the stock

market trend together, if they do at all? Sometimes answers can be found in subtleties. Notice that the deepest low in births this century came in 1933, the year after the deepest low in the stock market this century. Notice that the second most important low in births occurred again in 1975, one year after the second most important stock market low of this century. Why would there be a one-year lag? Well, can you think of any activity that always precedes a birth by about a year? If so, could this activity be correlated directly with people’s moods and therefore the trend and level of the stock market? We now have a tenuous basis for a socionomic explanation for demographic trends. As people in general feel more energetic, confident and happy, they conceive more children. Conversely, as people in general feel more sluggish, fearful and unhappy, they conceive fewer children. Thus social mood determines aggregate procreational activity. Chart 4: Films


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Films The social mood percolates through to cultural trends as well. An example of this can be seen in the production and success of certain film genres. The Walt Disney Company released its first feature-length cartoon in 1937, the year which saw the top of the roaring five-year bull market that accomplished the fastest 370% gain in US stocks ever. As can be seen from the titles listed on the top side of Chart 4, these films stayed popular for 30 years, culminating with the ultrasunny Mary Poppins in 1964, and to a lesser degree, The Jungle Book in 1967. The end of this period of success was essentially coincident with the great stock market top of 1966. For the next 16 years, as stock prices fell along with social mood, most people thought Disney’s feature cartoons were silly and sentimental. Indeed, the studio’s productivity fell by more than 50 per cent. Not one cartoon film from this period is considered a classic. When the bull market returned in the 1980s and 1990s, so did feature-length Disney cartoons that have been both acknowledged classics and box-office blockbusters. In the most recent 11 years of bull market, Disney produced 10 feature cartoon films. In brief, Disney cartoons are bull market movies, reflecting the shared mood of both their creators and their viewers.

Psychology and Markets

Systematic Trading

the 1980s through the 1990s horror films became increasingly derivative, muted or comic, just as in the years following 1942. The effect of social mood on cultural trends is not just restricted to the US. Chart 5 shows the explosion of horror films that were released in Japan during its bear market.

Music A collector of old 78-rpm records writing in The Wall Street Journal noted that music reflects “every fiber of life” in the US. Chart 6 shows that popular musical themes have been virtually in lock step with the social mood, as reflected by the major trends in the Dow Jones Industrial Average. Chart 6: Music

At the other end of the spectrum are horror movies. Some of the most notable of these productions are shown on the bottom side of Chart 4. Horror movies descended upon the American scene in 1930-1933, the very years that the Dow Jones Industrials collapsed. Five classic horror films were all produced in less than three short years. Frankenstein, Dracula and The Mummy premiered in 1931. Dr. Jekyll and Mr Hyde was released in 1932, and King Kong in 1933, on the test of the low in stock prices and right at the trough of the Great Depression. These are the classic horror films of all time. Ironically Hollywood tried to introduce a new monster in 1935 during a bull market but Werewolf of London was a flop. When film makers tried again in 1941, in the depths of a bear market, The Wolf Man was a hit. From the latter half of Chart 5: Japanese film titles

A useful perspective An individual who can rise above the social trends that sweep along his fellows so that he can observe their operation has an incalculable advantage over all those who do not. He has a basis upon which to anticipate the future, not every time and not perfectly, but well enough to have immense value. Anticipating the social future has always appeared to be a gift unavailable to humanity. That has been true until now, not because the task is impossible but because it requires detailed knowledge of the patterns of social behaviour and an ability to think and act in a fashion contrary to standard assumptions. Armed with knowledge of society’s patterned behaviour, a wise individual can train his reasoning to recognise emotionally driven social thought processes and use that knowledge to make decisions about how to harness or avoid coming trends in social mood and social action.

Robert R. Prechter is known for developing a theory of social causality called socionomics, for developing a new theory of finance and for his long career applying and enhancing R.N. Elliott’s model of financial pricing called the Wave Principle. https://robertprechter.com

139


140

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Oil - A real time application of the Elliott Wave Model Robert R. Prechter Jr., CMT

Article originally featured in Market Technician 69 (April 2011)

In light of the recent sharp rise in oil prices, we thought it would be interesting to publish the first part of Bob Prechter’s talk from February 2010, in which he analysed what Elliott wave analysts were saying, in real time, at each major turning point of this market over the past decade. He has added an update at the end of this article.

Chart 1: Crude Oil Prices (yearly/monthly 1859-1993)

I do not know very many people who, 10 years ago, correctly forecasted what the oil market was likely to do. Much less do we know many people who, when the price was up at $147 a barrel, were ready to call for the biggest drop in the history of oil - a 78% decline. So, what I am going to do is take you through a real-time struggle with the wave structure and show you what people at Elliott Wave International were saying during this period. The first chart shown above was produced by Al Graham, our oil analyst back in the 1990s. This is the picture of oil at that time, and it is interesting to see what he said: ‘Price should fall into the general area of $10 before the final leg up takes prices to new historical highs’. This was a dual forecast in that he predicted oil was first going to go down to around $10, then rally to a new all-time high. Would you be able to figure that out from this picture? Most people would not have any clue what was about to happen next. Even if you are an Elliott Wave practitioner, you are going to have to spend a little bit of time with this chart to see why you would make that forecast. Chart 2 shows what Graham was looking at. From way back in the early part of the 20th century, a five-wave structure appeared to be building on a very longterm basis, and each one of those waves on the upside subdivided into five-waves. This is a fractal, so you are going to see five-waves in each one of the impulses as you go along. But, having drawn a parallel trend channel, you can see that the corrective process begun in the 1970s was not over yet, because the price had not gone down to the lower end of the channel. There was not a good, completed pattern to count, either. That’s why he said, first, oil should go back down to the area of that low, but no lower, and then turn up to make new all-time highs. It took a few years but, eventually, in December 1998, the first of those targets was reached and oil got down to $10.35 a barrel. Undoubtedly, at that point, there was extreme bearishness about oil: ‘We have got a glut,’ ‘supply will continue to exceed demand,’ and so forth - you know how the fundamentals always fit in at the turns. Well, of course, oil then took off, tracing out a wave-one count, going all the way from $10 to the upper $30s - more than tripling in price. In December 2000, Graham wrote, ‘We are soon likely to see a wave C decline, where oil tests the 50% retracement to the $19.13 level. One relatively conservative projection would take crude up to $61 a barrel in coming years’. So, now he was saying that oil would go down into the teens from about $37 and then turn around and go, not only to a new high, but as high as $61 - and that was a conservative forecast.

Forecast December 1993: “Price should fall into the general area of $10.00 (to end a fourth wave move from 1980) before the final leg up takes prices into new historical highs.”

Global Market Perspective, December 1993 issue.

Chart 2: Long Term Waves in Crude Oil (yearly/monthly 1859-1993)


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Here is what actually happened: the price fell to $16.70, thereby completing wave two. It was not a 50% retracement, as Al expected; it was probably closer to 62%. But the point is, at $37 we knew we did not want to buy it; we were looking for a low under $20 a barrel - a pretty substantial drop. Once in this range, we would start looking for a bottom. We did get that bottom, and oil started to rally again. Chart 3 shows where oil was when I travelled over to Vienna to make a speech for the Kenos Conference in March 2006. We did not yet have five-waves within primary wave three, so oil needed to complete that part of the fractal, after which it would have a substantial correction, balancing out the wave two of primary degree, before another run up to new all-time highs. Chart 3: Price Projection (Crude Oil $ per barrel)

Chart 4: Oil: Five-waves up from 2001 (Crude Oil $ per barrel)

In the July 2006 Elliott Wave Theorist, I noted that we could now count internally to the peak of that primary degree third wave, completing five-waves up. In fact, because oil was in an overlapping situation called a diagonal, we were pretty certain that that was the end of primary wave three. Newsweek came out with oil on its cover page, which coincided with the very top of the market and, at the same time, many articles and books were being published about ‘peak oil’, so we knew it was due for a very powerful primary degree correction. As you can see toward the lower part of chart 5, that’s what happened. The market sold off substantially from $79 a barrel to under $50 over the next few months. When it got down to that level and bounced, Steve Craig, who was our analyst in the oil chair (and still is today), predicted the following: ‘I have adjusted the count to show the decline as zig-zag wave

Psychology and Markets

Systematic Trading

141


142

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

A of the primary wave four retracement. A strong B advance would be the norm. Given the depth of the decline, however, a strong alternate is that wave four is complete’. So you see some ambivalence - it could be over, but it may be only part of a correction; even that is more than most people knew at the time. It turned out to be the entirety of the correction, and from there oil started its primary fifth wave up. This is when it started going vertical. I was getting really excited at this point, because we knew that this was primary wave five. If you identify wave five, you certainly do not want to be chasing that market, because when it reverses it is likely to be a really powerful move. Commodities in fifth waves tend to be longer and stronger at primary degree than the other impulse waves in the structure. Commodities tend to ‘blow off’, and the reason they do is that the emotion driving the market is usually one of fear rather than confidence, which dominates when the stock market is making a high, so stock and commodity peaks are completely different in that respect. I was trying to catch the peak in order to short it. Many of the commodities had big runs in the 1970s, and the fifth wave of these moves was frequently related by the Fibonacci ratio to the run-up from the bottom to the peak of wave three. So, I projected that relationship and forecast that, if oil gets up to the $160-189 area, that should be the extent of the upward phase. But, whatever happens, Chart 5: Wave 5 in Crude Oil

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

it is a fifth wave, and when it is over we are going to have terrific downward move. It is going to be a blow-off situation, followed by a reversal. Chart 5 is a weekly chart of crude oil showing the fifth wave all the way up from $49 to the peak of $147.50. In the final impulsive stage, this move held remarkably well within two lines containing all the major touch points - showing how well even a blow-off can adhere to Elliott’s channels. If you are using the Elliott model, and you are channelling the waves, you will find a lot of clues along the way as to where prices are going to reverse, even on a short-term basis. The aftermath of the top in July 2008 was a 78% collapse in five months, down to $33.50. Why were we pretty sure that the decline would be so vicious after the top? If we were counting five-waves up only from the 1990s, we would not have expected it to drop that far; it would have pulled back only to the approximate area of the fourth wave within that rise. But I was convinced that oil was topping out on a very long-term basis, going back 90 years, as shown in Chart 6. And at the top it did what Elliott called a ‘throw over’ - it rallied up to the upper end of the channel, burst through it briefly and then fell back below. This was how the top was made in 1929 in the stock market.

Chart 6: Long term waves in crude oil (yearly/monthly 1859-2009)

Update as of March 23, 2011: We labelled the early 2009 low near $33 as wave A of a large A-B-C corrective pattern. Oil has since more than tripled to $107 and should be near the peak of wave B. Though the Federal Reserve has all but guaranteed that it will create inflation, this wave labelling implies that the dollar price of oil will fall below the 2009 low sometime this decade.

Robert R. Prechter is known for developing a theory of social causality called socionomics, for developing a new theory of finance and for his long career applying and enhancing R.N. Elliott’s model of financial pricing called the Wave Principle. https://robertprechter.com


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

The British bear market of 17201940: A Socionomic Examination Tom Denham

Article originally featured in Market Technician 52 (April 2005)

It’s wintertime in 1940. A family in London cowers under their kitchen table behind drawn black-out curtains, quaking as they listen to the air raid sirens and fearing the sound of German bombers overhead. But, 220 years earlier, all that their ancestors had to worry about was how to spend the newfound wealth created when the price of their South Sea Company stock rose nearly 700% in eight months. What happened? One reason to look back at historical stock charts is to see the patterns in the past so as to project patterns in the future. Another reason is to learn more about the social mood of the time, because social mood is reflected in stock prices. More important, while fundamental events do not cause changes in wave patterns, changes in wave patterns can forecast changes in underlying fundamental conditions. On 2 July 2004 we observed a contracting triangle spanning 220 years from 1720 to 1940, followed by a thrust from 1940 to 2000. This view confirmed that the period in British history from 1720 to 1940 formed a Grand Supercycle wave IV. (See Figure 1.) In other words, it was a long bear market. This article observes how social mood was reflected in stock prices during this period, by comparing social events in the United Kingdom with the highs and lows of the British equity markets.

Figure 1: 300 years of British Equity Data

143


144

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

South Sea Mania Kicks Off British Wave IV

Socionomic Highlights, 1720-1940

The South Sea Company, which was formed in England in 1711, was the primary vehicle of financial speculation by 1720. The company used its government-granted monopoly on trade with the New World to attract investors. As prices surged, aristocrats and common folk alike borrowed on margin to secure as much stock as possible. The price of South Sea Company stock rose eight-fold from £128 in January 1720 to £1000 in early August, then collapsed to £150 by the end of September. Thousands of people were ruined. An investigation by Parliament discovered widespread fraud among company directors.

As the 220 years from 1720 to 1940 wore on, anger manifested itself in various ways as people in the United Kingdom lived through wars, violent rioting, political repression, and famine. However, they also experienced the economic progress of the Industrial Revolution, expanded voting rights, broader educational opportunities, and greater protection for workers.

As Bob Prechter observed in “Bulls, Bears and Manias,” an Elliott Wave Theorist Special Report (May 1997): “Historians characterize manias as a kind of madness that takes hold of a population. The widely shared illusion of endless huge profits that propels a mania also produces another kind of madness: anger... Regardless of extent, every mania is followed by a decline that ends below the starting point of the advance. Considering the heights that manias reach, this is an amazing fact.” The mania of 1720 was not limited to South Sea Company stock. The entire London stock market rose and fell with the mania. The Financial Times All Share Index started from a low of 24.66 in 1719 and shot up 323% to the bubble top at 104.34 in 1720. Two years later, in classic form, the index was down 83% to a lower low at 18.13. This was the kickoff to a series of broad swings in stock prices and historical events over the next two centuries.

In other words, even though it was a Grand Supercycle wave IV bear market, prices (and social mood) didn’t simply move in a straight line. They moved in a common Elliott formation called a contracting triangle. In particular, a contracting triangle pattern usually reflects a balance of negative and positive social forces. Consistent with the five-waves of a contracting triangle, British stocks went through three Supercycle bear markets [(a), (c) and (e)] and two Supercycle bull markets [(b) and (d)] between 1720 and 1940. Moreover, within these Supercycle swings, there were a number of Cycle degree bull and bear markets. Let’s review in brief some of the events that shaped this period of time. We will group the events according to whether the stock market (and, thus, social mood) was on an upswing or a downswing within wave IV. The broad strokes of the period can be seen in Figure 2. For a more detailed description of these events, please see the appendix to this report.

Wave (b) up: 1816-1825 What is Socionomics? R.N. Elliott discovered and popularized the Wave Principle in the 1930s to forecast financial markets. Although Elliott recognized that the Wave Principle could also be used to assess social trends, it was Bob Prechter who offered an articulate demonstration of how human behavior is moodbased and patterned, according to the Wave Principle, in his 1999 book, The Wave Principle of Human Social Behavior and the New Science of Socionomics. Bob observed that the stock market records the vicissitudes of far more than stock prices. It records the ups and downs of society’s mood state. More than is generally recognized, speculation in financial markets results from emotional, rather than rational, decisions. Stock market investors can respond almost immediately to social mood changes; investors in the aggregate buy when their emotional state is positive and sell when it is negative. Their transactions are meticulously tabulated and, therefore, provide voluminous insight into the net emotional state of society. Because social mood determines the character of social events, a rising stock market serves as a leading indicator of positive social events and a falling stock market forewarns of negative social events.

The British economy began a period of rapid expansion in 1815 after the Napoleonic wars. Some of the positives during that time: • Exports to the newly independent countries of Latin America boomed. • Large-scale infrastructure projects (e.g., gas lighting, canals, and railroads) were funded. • Factory legislation in 1819 limited those aged nine and above to a 12-hour day. • Robert Peel became Home Secretary in 1822 and led reforms of the legal system that removed the death penalty from more than 100 crimes. • On the cultural side, the Royal Academy of Music opened in 1823 in London. The Bank of England exercised easy monetary policy during this upswing, and the stock market boom became a bubble as investors bid up the prices of real and imaginary stocks, such as bonds from the imaginary South American Republic of Poyais. Another century, another South Sea Company-style mania. This bubble burst in 1825.

Wave (a) down: 1720-1816 After the 1720 collapse, a long period of conflict ensued that had both local and global features. The global conflict included war with Spain, the Seven Years War, the American Revolution, and the Napoleonic wars. From a socionomic perspective: “Major mood retrenchment produces war, as humans finally express their collective negative mood extreme with representative collective action... The size of the war is almost always related to the size of the bear market that induces it.” (The Wave Principle of Human Social Behavior, page 266)


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

Figure 2: Grand Supercycle Wave IV in Britain

Wave (c) down: 1825-1849 Famine and illness swept through the United Kingdom during the period from 1825 to 1849. •

Two great Cholera Pandemics ravaged the country, the first reaching its height in 1831 and the second in 1848. Cholera killed as many as 1,000 people per day when the disease’s devastation in England was at its worst.

Life in Ireland reached a low point as the Potato Famine tightened its grip. Between 1845 and 1851, one-and-a-half million people died during the Irish Potato Famine, and another million emigrated to escape starvation. The government was slow to respond when the potato blight first emerged and, in the end, the population of Ireland decreased by 20%.

The French sociologist, Gustave de Beaumont, visited Ireland in 1835 and wrote: “I have seen the Indian in his forests, and the Negro in his chains, and thought, as I contemplated their pitiable condition, that I saw the very extreme of human wretchedness; but I did not then know the condition of unfortunate Ireland... In all countries, more or less, paupers may be discovered; but an entire nation of paupers is what was never seen until it was shown in Ireland.” (from www.historyplace.com)

Wave (d) up: 1849-1929 The 80-year long Supercycle wave (d) was a series of Cycle degree bull and bear markets, some lasting even longer than their Supercycle cousins. All the same, strongly positive events clustered

145


146

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

near Cycle degree highs and strongly negative events tended to occur near the Cycle degree lows. Figure 3 takes a closer look. Figure 3: Focus on Supercycle Wave (d) in Britain

From the time of the stock market bottom in 1849, the social tide began to turn positive. • The Factory Act of 1850 restricted all women and young people to no more than 10½ hours work a day. • The Education Act of 1870 provided for genuine mass education on a scale not seen before. By 1874, more than 5,000 new schools were founded.

By 1906, Parliament began a series of ambitious social reforms: medical examinations for schoolchildren, free meals for the poorest students, a program for slum clearance, and the introduction of a basic old-age pension scheme.

• In 1918, the government granted women over the age of 30 the right to vote, and, in 1928, the right to vote was extended to women over 21.


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

And, negative events showed up strongly near the lows. • The last public hangings were conducted in 1868. • The takeover of the Transvaal Republic and the first Boer War occurred between 1877 and 1880. • The worst rioting over Irish Home Rule took place in Belfast in 1886. • “Jack the Ripper” terrorized London as a serial killer in 1888 and became a cultural icon. • The First World War began on the Continent, and Britain entered World War I in 1914. After the war, demobilized soldiers came home to a poor economy. More than two million people were unemployed by 1921, and strikes increased. •

Frustration over unmet demands for Home Rule led to an armed uprising in Dublin in 1916. The British subsequently executed 15 nationalist leaders and interned 3,000. Civil war continued until 1923.

Wave (e) down: 1929-1940 Britons experienced some of the worst hardships in modern history towards the end of the Grand Supercycle wave IV. • Unemployment peaked just below three million in 1932, up from the two million unemployed in 1921 in the aftermath of World War I. • The Irish Republican Army engineered terrorist bombings in London, Liverpool, Birmingham, Manchester, and Belfast in 1939. • The Second World War officially began in Europe in 1939 when Britain and France declared war on Germany after deciding that Hitler could not be allowed to seize Poland. • Children were sent to the countryside for safety, while adults huddled in basements and feared their country would be invaded.

Psychology and Markets

Systematic Trading

Appendix How well did the stock market serve as a leading indicator of events? We did a small and admittedly unscientific study by sorting through the timelines of Britain and England published at www.bbc.co.uk/history/timelines/ to identify positive and negative events. The Victorian Web project of the National University of Singapore at www.victorianweb.org and a few other sources were consulted as well. The final list included 68 events, 37 of which we labeled as “negative” and 31 “positive” (included in the Appendix). The included events are not of uniform importance. For example, the Napoleonic Wars surely had a greater impact on society than Elizabeth Garrett Anderson becoming the first licensed female doctor in 1865. A historian who specializes in this period could undoubtedly improve upon our list, but we think it presents an adequately accurate picture of the times. The important thing to know is that we did not exclude any event in order to improve consistency with socionomic expectations. Some important developments extended over a long period and could not easily be associated with a specific date. One example is the textile industry, which began to flourish in Britain in 1733 as a series of inventions made it possible to industrialize spinning and weaving. The development of the textile industry was a social positive in the sense that people worked together to create more wealth. On the other hand, owners violently repressed early attempts to unionize. Underlying fundamental conditions improved, because England went from an agrarian to an industrial economy. But new technology did not rehabilitate the social mood, and the bear market continued until it had run its negative course. Here’s what we learned from plotting the positives and negatives on the All Share Index chart to see how consistently historical events lined up with socionomic expectations: Approximately three-quarters of the time, negative historical events occurred near market lows, and positives occurred near market tops. See Figure 4. While not a perfect fit, it’s far better than a 50-50 chance. Figure 4: Plotting Social Events Across 220 Years of British History

Conclusion British social mood reached its low point in 1940. Rationing began in January. The news filtering in was all bad: Germany bombed the Scapa Flow naval base near Scotland in March. Germany invaded Denmark and Norway in April. Germany invaded France, Belgium, Luxembourg, and the Netherlands in May. In July, German U-boats attacked merchant ships in the Atlantic and the Battle of Britain began. But it appears that a critical point in social mood shifted in July 1940. The Financial Times All Share Index bottomed that month at 18.58 while the war escalated. Ironically, in a historic big-picture version of “sell the rumor, buy the news,” no bomb touched London until after the low was in place. This low held even though the Battle of Britain wore on through May 1941 and despite Germany’s attempt to destroy British morale with nightly bombing raids on London and other major cities. The conclusion of the triangle wave pattern confirmed the wisdom of buying stocks while air raid sirens still blared. Stock prices gained 117% from the 1940 low to May 1945 when the war ended.

147


148

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

British Social Events, 1720 - 1940 N E G AT I V E S 1

1735

People objected to paying toll road fees, and there were serious outbreaks of rioting in 1735 and 1750.

2

1739

Competition in trade between Britain and Spain grew increasingly angry until war broke out.

3 1745

Catholic factions tried to reclaim the monarchy. Bonnie Prince Charlie led an unsuccessful uprising to make his Catholic father, James Stuart, king.

4

1750

Turnpike riots - see 1735 entry above.

5

1756

Seven Years War begins.

6

1775

7 1780

Britain’s North American colonies revolted, and eventually France, Spain, and the Netherlands entered the war against Britain. Violent anti-Catholic riots broke out in major cities in response to the repeal of harsh anti-Catholic legislation from the seventeenth century.

8 1800 Many landowners in Scotland ‘cleared’ local people who were no longer economically useful. On occasion, this involved burning their houses above their heads. 9

1800 Legislation restricting the freedoms to speak, publish, meet, and organize suppressed hundreds of local reform societies.

10 1803 Britain resumed war against France to stop the European conquests of Napoleon. Napoleon was not defeated until 1815. 11 1811

New machines and industrial practices threatened traditional home workers. Led by Ned Ludd, people destroyed machines and factories. After 17 Luddites were executed in 1813, the movement diminished.

12 1815

The Corn Laws were passed to enforce high grain prices and kept cereal and bread expensive to support farmers who had become accustomed to high prices during the Napoleonic Wars. Consumers complained.

13 1819

11 people died when cavalry charged a large meeting of reformers gathered to hear a speech at St Peter’s Fields in Manchester. The government taxed the radical press, sent spies and provocateurs into the reform movement and empowered magistrates to suppress public meetings.

14 1825 Bust of the export and infrastructure investment boom. 15 1831

Insurrectionary rioting erupted across the country when the House of Lords rejected the Reform Bill.

16 1831

Cholera Pandemic of 1826-37. More than 21,500 persons died in England and Wales. There were 9,500 deaths in Scotland. Ireland had 25,000 deaths in 1832.

17 1837

Oliver Twist by Charles Dickens began to appear in serialized form. The popular novel was a thinly veiled protest against the Poor Law of 1834, which dictated that all public charity be channeled through workhouses. The old, the sick, and the very young suffered more than the able-bodied benefited, according to Dickens.

18 1839 Parliament rejected a People’s Charter, which demanded democratic rights. 19 1845 More than 1 million citizens died during the Irish Potato Famine of 1845-1851. Another 1-2 million emigrated. Slow government response exacerbated the problem. 20 1847

Bust of the railway mania. Railway companies lost 85% of their value and several hundred folded.

21 1848 Marx and Engels publish The Communist Manifesto in London. 22 1848 Cholera Pandemic of 1846-63. The epidemic claimed 1,000 lives a day at its height in England. 23 1870

Parliament passes the Married Women’s Property Act, which allowed women to keep their earnings, inherited personal property, and small amounts of money in a divorce.

24 1854 Britain entered the Crimean War to prevent Russian expansion into the Ottoman Empire. 25 1857

Hindu and Muslim soldiers mutinied against a series of military demands by their British commanders. It escalated into widespread rebellion in India.

26 1868 The last public hangings took place. 27 1877

Britain annexed the Transvaal Republic in southern Africa. In 1880, the Boers of the Transvaal revolted against British rule and achieved independence.

28 1886 The worst rioting over Irish Home Rule took place in Belfast. 29 1888 “Jack the Ripper” terrorizes London as a serial killer and becomes a cultural icon. 30 1899 During the second Boer War of 1899-1902, to annex the Transvaal, many Boers died under oppressive conditions in British concentration camps. 31 1914

Britain entered World War I.

32 1916

Frustration over unmet demands for Home Rule led to an armed uprising in Dublin. The British subsequently executed fifteen nationalist leaders and interned 3000. Civil war continued until 1923.

33 1918

Industrial profits and wages fell, and demobilized soldiers found it difficult to find jobs. More than 2 million people were unemployed by 1921, and strikes increased.

34 1919

British soldiers kill nearly 400 and wound more than 1,000 Indian nationalist protestors in the Punjab.

35 1932 Unemployment peaked just below three million. 36 1939 Britain entered World War II. Britain endured heavy and frequent bombing raids and feared invasion through 1940. 37 1939 The Irish Republican Army engineered terrorist bombings in London, Liverpool, Birmingham, Manchester and Belfast.


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

POSITIVES 1

1753

The British Museum founded.

2 1768

James Cook undertook the first of three exploratory voyages to the Pacific, as British interests in the wider world expanded.

3 1769

James Watt invented an improved steam engine that soon became the dominant design and helped bring about the Industrial Revolution.

4

1819

Factory legislation limited those aged nine and above to a 12 hour day.

5 1822 Robert Peel becomes Home Secretary and leads reforms of the legal system that removes the death penalty from more than 100 crimes. 6

1823 The Royal Academy of Music opened in London.

7

1824

8

1825 George Stephenson built the first public steam railway from Stockton to Darlington.

9

1825 Export and infrastructure investment boom, which started in 1815.

The National Gallery of Art established in London.

10 1826 University College, London, founded. 11 1832 The first Reform Bill, which extended voting rights and redistributed Parliamentary seats, is passed. 12 1833 Factory legislation prohibited the employment of children under nine in mills and further restricted the time children over nine could work. 13 1846 Parliament repealed the Corn Laws. 14 1847

Top of the railway mania that began to gather steam in 1835. Railways became fashionable, new railway lines were built, people poured their savings into railway stocks.

15 1848 The growing Sanitary Reform Movement led to a Central Board of Health with powers to supervise street cleaning, refuse collection, water supply and sewage disposal. 16 1850 The Factory Act restricted all women and young people to no more than 10½ hours work a day. 17 1851

The Great Exhibition celebrated British imperial and industrial might. More than 6 million visitors to the exhibition viewed over 13,000 exhibits. The profits from the event allowed for the foundation of public works such as the Albert Hall, the Science Museum, the National History Museum and the Victoria and Albert Museum.

18 1855 The Limited Liabilities Act allowed companies to limit the liability of their individual investors to the value of their shares. Prior to this, investors in a company stood to lose all their wealth if economic circumstances forced the company out of business. 19 1858 The transportation of convicts to remote penal colonies such as Australia was abolished. 20 1865 Elizabeth Garrett Anderson became the first licensed female doctor. 21 1867

The Reform Act attempted to redistribute parliamentary seats in a more equitable manner.

22 1870

The Education Act provided for genuine mass education on a scale not seen before. By 1874, more than 5,000 new schools were founded.

23 1848 Five million citizens signed a petition asking for a People’s Charter of rights, and Parliament rejected it anyway. 24 1882

Parliament passes a broader Married Women’s Property Act, which allows married women to keep all personal and real property acquired before and during marriage in a divorce.

25 1884 The Third Reform Act increased the right to vote and gave more representation to urban areas. 26 1887

Massive outpourings of public affection were displayed upon Victoria’s Golden Jubilee.

27 1897

More affection was displayed upon Victoria’s Diamond Jubilee.

28 1901

Victoria’s death was an occasion of national mourning.

29 1906

Parliament began a series of ambitious social reforms and reversed the 1901 Taff Vale judgment, which had made trade unions liable for employer’s losses during strikes.

30 1918

The government granted women over the age of 30 the right to vote.

31 1928 The right to vote was extended to women over 21.

Tom Denham is the Editor of European Financial Forecast Service, a publication produced by Elliott Wave International.

The Great Exhibition, 1851.

149


150

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

The Golden Ratio and its presence in financial markets Tony Plummer FSTA

Article originally featured in Market Technician 62 (October 2008)

Preface This is an amended and updated version of an article that was originally published in the July/August 2001 issue of the Australian Technical Analysts Association Journal. It was an invited response to the valid question: What is the justification for using Fibonacci ratios to formulate entry and exit trading rules? My own experience with the Golden Ratio - which is the proper name for the phenomenon in question - is that it has a profound role to play in tracking the evolution of a market’s primary trend. In my opinion, it is ultimately superior to signals derived from (for example) Dow Theory and trend line breaks, although it is best used in conjunction with such techniques. The purpose of the article, therefore, was - and is - to survey the philosophical background to the Golden Ratio, and to assess some of the practical implications of the Ratio for financial markets.

The divine measure Visitors to King’s College Chapel in Cambridge, England, will soon find themselves standing in front of a massive rectangular stained glass window above the altar. The window dominates the whole area known as the ‘choir’ and helps to fill it with light. Some will see the window as a relic of a bygone age. Others will see it as a beautiful piece of artwork. Yet others will see it as something mysterious, yet somehow relevant. In fact, the window is all of these things, but it is the last quality - its mysteriousness - that seems to capture the imagination. It points, somehow, to something beyond the mundane. And this is not just because it resides in a church (although that is relevant); it is because it contains within it dimensions that have long been associated with the sacred. Reflecting the nature of revelation itself, these dimensions are not obvious from the outside; they can only clearly be seen from the inside, as the light shines through. Specifically, the central panel has a width that is 61.8% of its height. This is a physical representation of a mathematical relationship that is known as the ‘Golden’ or ‘Divine’ Measure. In the diagram below, a line is divided in such a way that the ratio of the smaller part to the larger part is exactly equal to the ratio of the larger part to the whole. That is, in other words, the relationship BC/AB = AB/AC = 0.618

1

Figure 1:

A

B

C

between successive lengths - moving upwards from the smallest, through the largest, to the whole - is constant. This means that, if the length AC is given as 1 unit, then since AC = AB + BC, AC = 0.382 + 0.618

2

And it also means that 0.382/0.618 = 0.618

3

The number 0.618 is the Golden Number, the associated ratio, 0.618:1, is the Golden Ratio, and the constant proportion between the lengths, 0.382:0.618:: 0.618:1, is the Golden Proportion.1 The generic term for the phenomenon is the Golden Measure.

The importance of ratios Before enquiring more deeply into the essence of the Golden Measure, it is important to understand the role of ratios in human thought. For human beings, ratios are directly involved with the process of conscious reasoning. Indeed, the words rational (‘of the reason’) and ratiocination (‘reasoning’) are derived from the same Latin root as ratio. Older civilisations than ours had therefore obviously spotted the importance of ratios to humankind’s ability to think clearly. To them the ability to reason meant the ability to use ratios. Suppose, therefore, that we wanted to present a ratio, or set of ratios, that simultaneously represented the abilities of the human mind to perceive, reason and draw conclusions. Put another way: What would be the characteristics of a ratio (or ratios) that best represented the various types of information that a human mind can receive and generate? First, the human mind needs to be able to ‘recognise’ information. This means that it needs to be able to recognise a change, or difference, between one state and another. This is the basis of analytical reasoning. In its simplest form, therefore, if we denote one state as being X and the other as being Y, and the human mind recognises and responds to the difference between X and Y, then the situation could be represented by the ratio X:Y. But note what happens if X = Y. There is no ‘difference’ as such, and the ratio X:Y becomes equal to unity. Without difference, there is no information and no response. Hence, a ratio that represents analytical reasoning needs to consist of unequal numbers so that it is not equal to unity. Second, however, the ratio X:Y does not in itself need human intelligence in order to create a reaction. The ratio could, for example, represent an amoeba’s response to a change in its environment. So a simple number is a necessary, but not sufficient, part of the representation of human intelligence. In fact, human intelligence has a specific additional dimension. This is the ability to understand by analogy - i.e., by comparison. The ability to make


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

comparisons facilitates attraction and repulsion. Consequently, a ratio that represents the human mind and its abilities would need to include three terms. That is, the appropriate ratio would have to be able to compare not just X with Y, but also Y with (say) Z. Hence, the appropriate ratio would be something like X:Y::Y:Z. Providing that X and Y are not equal, this allows for perception via both analysis and analogy. The ability to reason analytically and analogically will ensure survival.

Psychology and Markets

151

Figure 2:

Even so, there is a third element. The proportion X:Y::Y:Z does not allow for the absolutely critical part of human intelligence that goes beyond ensuring survival. This is the faculty of creative insight, or intuition, which encourages evolution. The very real problem, however, is that we do not really know where such insight (or intuition) comes from. It is one thing to collect pertinent facts, but it is yet another thing to claim that the critical insight based on those facts was purely the result of analytical and analogical reasoning. Or, to put the same thing another way, we can create the right quantitative conditions for insights, but the essential ingredient of putting it all together and ‘seeing’ a different qualitative dimension within a data set, lies somehow beyond our ordinary everyday abilities. Some of the greatest philosophers - including more contemporary ones such as Alfred North Whitehead and William James - have suggested that intuition arises from a level of the mind’s organising power that is not easily accessible. Specifically, they have suggested that accessing intuition is very much a case of becoming passive and allowing an insight to ‘drop through’ in one whole piece, as it were. The ‘effort’ therefore needs to go towards relaxing the mind, rather than exciting it. If this is correct, then it suggests an intriguing possibility. This is that the mind is in a feedback loop with its own higher levels of organisation. The idea of feedback between hierarchical levels of organisation is a vitally important aspect of living systems. More immediately, here, the higher levels of mind can only contain lower levels of itself. This implies that, in the proportion X:Y::Y:Z that represents the human mind, the third term must consist only of the first two. That is, Z has to consist of X and Y.

F

E

D

A

B

C

The Spiral of Squares Next, we can draw a square within the smaller rectangle BCDE, with one side on, and equal to, the line BC. See Figure 3. Two things have now happened: First, since by construction BC/AB = 0.618, then the square BCGH is 61.8% of the square ABEF. Second, we have simultaneously created a rectangle GDEH. Since from equations 1 and 2, BC/AC = 0.382, then rectangle HGDE is 38.2% of rectangle ACDF. These comparisons apply to area sizes and to perimeter lengths.

Figure 3:

F

E

D

H

So, the pair of ratios that best represents the three aspects of the human mind - analytical reasoning, analogical reasoning and intuition - is the proportional relationship X:Y::Y:(X+Y). This is the Golden Proportion.

Three terms from two

A

The Golden Proportion is, in fact, the only pair of ratios that allows the proportional relationships between three terms to be expressed by the relationship between two of those terms. Referring to Figure 1, AC is equal to the sum of AB and BC. Hence, equation 1 expresses the relationship between three terms by the relationship between two of those terms. In other words, if BC is to AB as AB is to AC, and if AC = AB + BC, then BC/AB = AB/(AB+BC) = 0.618

Systematic Trading

G

B

C

The process can, again, be repeated. See Figure 4. This time the new square is drawn with one of its sides on, and equal to, GD. The rectangle GDEH then consists of a square GDIJ and a rectangle HJIE. Figure 4:

4

F

E

I

D

The Golden Square and the Golden Rectangle Moreover - and this is important - the relationship stays constant, no matter what absolute quantity is involved. That is, the Golden Proportion operates in (and despite) the processes of growth and decay. So, in Figure 1, AC can get bigger or smaller, but the internal integrity of the relationships remains unchanged. A very basic diagram can demonstrate this idea. In Figure 2, a rectangle ACDF is constructed on the line AC, derived from Figure 1. However, the dimensions of the rectangle are such that the width of the rectangle (i.e., CD or BE) is equal to the length of the line AB. In other words, the rectangle is drawn so that ABEF is a square.

H

A

G

J

B

C


152

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

The Golden Spiral

The Golden Measure and religious insight

It can immediately be seen that we are constructing a series of contracting squares, whose boundaries are touching in such a way that they appear to be spiralling towards some infinitesimally small amount. Indeed, the points B, G, I, etc. can be linked together by an appropriate spiralling arc. See Figure 5.

However, this was not always the case. It is not now known when the presence of the Golden Measure was first recognised by humankind,4 but what is clear is that when one of the older civilisations came to know about it, it was treated with a particular degree of reverence. We can only speculate about the sense of awe that contact with the concept of the Golden Measure might have aroused. But there seems to be no doubt that it was considered to be a direct reflection of the essential nature of God:

Figure 5:

First, the logarithmic spiral constructed using the ratio 0.618:1 (see Figure 5) starts and ends in infinity. The spiral thereby identifies the unattainable point where the infinitely large is harmonised with the infinitesimally small. Second, since the larger of the elements in equation 1 is equal to the sum of the other two elements, the Golden Number 0.618 is an expression of the way that The Source of All Things divides Itself into Two. Specifically, the larger of the two portions is the receptive power in creation; while the smaller portion is the generative power. Horizontally, this means female and male respectively; vertically, it means spiritual and mental respectively.

Logically, of course, we could have constructed the diagram differently, drawing ever-increasing squares on the sides of rectangles. The process would represent a spiral towards some infinitely large figure.

The Golden Measure in nature Now, this is not just a rarefied exercise in mathematics or logic. The importance of the Golden Measure lies in the fact that it appears everywhere in Nature. For example, the length from the base of a person’s foot to the navel is, on average, 61.8% of that person’s total height. The distance from the navel to the middle of the neck is then 61.8% of the distance from the navel to the top of the head; and the distance from the navel to the knee is 61.8% of the distance from the navel to the base of the foot. Meanwhile, computer analysis of a so-called ‘ideal’ face, made from an amalgam of a large number of faces, confirms that the width of the head is 61.8% of the height of the head; that the distance from the centre of the eyes to the tip of the nose is 61.8% of the distance from the eyes to the mouth; and that the distance from the tip of the nose to the chin is 61.8% of the distance from the eyes to the chin. Furthermore - and by no means finally - each bone of a person’s finger is 61.8% of the length of the adjoining metacarpal.2 Of course, there will be significant individual divergences from all of these calculations; but the larger the number of individuals whose structural proportions are measured, the more likely it will be that the average of the ratios will converge on the Golden Measure. This is not the place to go into all the evidence for the presence of the Golden Measure,3 but the point to be clear about is that it is ubiquitous in living systems. This raises very important questions, such as: Why is this so? What does it imply about the laws of Life? And should we actually be paying more attention to it? Unfortunately, although our present culture recognises the presence of the Golden Measure, it nevertheless tends to treat it as something of an accident, or evolutionary genetic outcome, and to leave it at that. Very few take the time to ponder some of the deeper implications.

Third, the Golden Proportion, 0.382:0.618::0.618:1, was seen as an expression of the Trinity - the idea of three sacred powers within one. From this perspective, the generative and receptive energies (0.382 and 0.618 respectively) are balanced and harmonised by higher-order purpose, meaning and organisation (i.e., unity). One of the original by-products of a religious interpretation of the Golden Measure was that it was treated, not so much an outcome as an influence - not so much an effect as a cause. It was considered to be, in some way, inherent in the blueprint of creation and therefore ‘prior to’ the structures of this world. However, a more ‘modern’ interpretation allows that causality can run in both directions. That is, change can be initiated from above (i.e., from the whole), through either the active pole or the passive pole (i.e., the parts); or it may be initiated from below by either the active part or the passive part, and the whole adjusts. Since the mathematical relationships require only that proportions remain constant, change anywhere in the system is automatically supplemented by change in other parts. Among many other things, therefore, a religious interpretation of the Golden Proportion would be that Man’s decisions are validated by God and that God’s Will is implemented by Man. The Golden Measure was thus both the mathematical reconciliation between the Whole and Its Parts and the archetypal mathematical formulation of the ancient saying: “As above, so below”.

The ‘Laws’ of Life On the face of it, these interpretations might be seen as extraordinarily arcane and, accordingly, as having little or no relevance to the modern world. But if we look a little deeper, there is a logic contained within them that not only bears consideration but also has a direct correspondence to modern scientific thought. Contemporary analysis has had to come to terms with three aspects of life on this planet. First, Life is a self-organising, constantly changing, process that is stimulated by (and therefore learns from) information ‘shocks’.5 Second, Life interacts with itself via circulation between opposite polarities. And third, Life is attracted either to expansion or contraction. These are fundamental ‘laws’ for the existence of Life; and there are no exceptions. However, by definition, these laws have certain mathematical requirements, not the least of which is the need to


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

reconcile random fluctuations with stable processes. And the point is that the Golden Measure may actually be part of the context within which this reconciliation can take place.

Self-organising hierarchies In many ways, the clue to this reconciliation comes from the comments already made about the Golden Proportion in the context of the human mind and religious insight: not only are lower-order parts of a living process organised by higher-order levels, but the higher-order levels also necessarily consist of, and are affected by, the lower-order parts. The difference between the two levels in this ‘hierarchy’ is that the whole is somehow greater than the sum of its parts. This simple observation goes some way towards resolving the paradox between random fluctuations and stable processes: the lower levels of the hierarchy can produce apparent randomness, but these fluctuations are contained by the higher ordered process of which they are a part. Importantly, modern scientific thought absolutely accepts the presence of such hierarchical organisation, and the resulting coexistence both of fluctuation and of order in the phenomenon of Life.6 The Golden Proportion is a perfect representation of these characteristics of Life: it demonstrates hierarchy, allows fluctuation, and asserts harmonic order. So, despite the apparently arcane origins of the Golden Measure, it precisely contains within itself the important features of a selforganising system. The Golden Measure recognises information in the form of differences; it accommodates responsiveness to that information by allowing comparison; and it allows for feedback within a system by recognising the presence of hierarchical levels of self-organisation. We are back to the three conditions for the existence of Life on this planet: self-organisation stimulated by information, circulation between opposite polarities, and attraction to growth or decay. These are profound conclusions that raise equally profound questions about the insights of those who originally worked with the Golden Measure. After all, only the terminology has changed. But, as we have already seen in relation to the human body, the Golden Measure does not appear to be just a metaphor. There seems to be something quite practical about it.

Metaphor and reality The first point to make is that metaphors and symbols are too often taken to be an irrelevance to modern life. However, psychologists are now very aware of the profound impact that metaphors and symbols can have on the individual psyche - not only as representations of inner realities, but also as triggers for inner change. Metaphors and symbols activate the right-hand side of the brain, which is non-linear and ‘holistic’. In fact, the right side of the brain has to be harmonised with the left side for the faculty of creative insight to be enabled. It is therefore hubristic to consider that the left-hand, linear side, of the brain is more important, more creative and more useful. And, as a corollary, it is mistaken to consider that historically early insights into the nature of being were misguided, or irrelevant. They were actually very likely to have been quite appropriate; it is just that they may not necessarily have been packed out with a modern degree of scientific detail.7 The second point, however, is possibly more important. This is that the Golden Measure is not just a right brain symbol. It is also clearly defined by left-brain mathematical equations. The science of mathematics rightfully belongs to the realm of ideas, but its ultimate value lies either in the accuracy of its description of known reality or in its implications for unknown reality. So, mathematics can, like metaphors, point towards a truth that lies

Psychology and Markets

Systematic Trading

beyond the boundary of known experimental techniques. Hence, one inference that could be drawn from the mathematics of the Golden Measure is that, if it is relevant to Life on earth, we really should be able to identify its presence in indicators that, in some way, reflect the behaviour of living systems. And it is arguable that one such system is a financial market.

Financial market groups This idea holds water because the phenomenon of selforganisation is universal. If every part of Life has its own individuality but belongs to a larger group, then it follows that there are forces at work in human behaviour that originate from beyond each individual person. There is nothing esoteric about this (although that, too, is an option). All it means is that human beings come together in various ways and create groupings whose influence is greater than the sum of their parts. At a simple level, there is ‘family’ or ‘tribe’; at a larger level there is ‘society’ and at a meta-level, there is ‘civilisation’. And these groupings incorporate the influences of history, so self-organisation extends over time. The point is that these groupings are psychological and not necessarily physical. They are based on beliefs, understandings, and codes of behaviour that are transmitted from one person to another, from one sub-group to another, and from one generation to another. And they are discriminatory: an individual is either in a particular overall grouping or out of it. This is the basis of all forms of group competition and group violence. Individuals behave differently in groups than they do in isolation.8 Within financial markets, the outcome of group behaviour is measurable by the fluctuation in prices. Indeed, financial markets form a wonderful laboratory for testing out ideas about group behaviour. Moreover, a simple observation of market prices confirms the obvious - i.e., that they oscillate - so it is probable that some form of swing between opposite polarities is at work. Consequently, a practical working hypothesis includes three features: (a) financial markets are a part of a wider socio-economic environment and are therefore ‘organised’ by the latter; (b) a market ‘organises’ its parts to conform with developments in the higher socio-economic environment; and (c) the various hierarchical levels are linked by the transfer of information that creates oscillations. Under this hypothesis, a group is a part of the natural order of things and, as such, its behaviour should reveal the influence of the Golden Measure.

The Golden Measure in financial markets In fact, the presence of the Golden Measure in financial markets has been known about since the early part of the last century. R. N. Elliott9 was the most celebrated exponent but others such as H.M. Gartley10 also utilised it very effectively. Elliott saw that corrections during a mature trend would tend to allow 61.8% of the whole trend to be retained; both he and Gartley saw that a retest of a high or low would likely be limited to 61.8% of the initial break away from that high or low. To put the same thing another way, both saw that corrections were somehow constrained by the Golden Proportion 0.382:0.618::0.618:1 And their observations are not without foundation. Indeed, the Golden Measure is so synchronised with financial market vibrations that it can be safely used to confirm all strategic entry and exit points. Just a few examples will suffice to indicate the principles involved. Figure 6 shows the US dollar against the Swiss franc from its 1995 secular low to its 2000 high, and the subsequent drop. Note that the sharp fall in the autumn of 1998 retraced exactly 61.8% of the rally from the 1995 low. This formed the base for the subsequent impulse wave. That wave stopped in 2000 at almost exactly 38.2% of the 1985 to 1995 dollar bear (not shown in the chart). The dollar then spent 18 months forming a

153


154

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

top pattern, which was persistently bounded on the downside by 38.2% of the 1998-2000 impulse wave. This boundary was finally broken on the monthly close in May 2002, thereby signalling that a correction had transmuted into a major downtrend. This signal and its precursors could, of course, have been identified earlier by using weekly charts. Figure 6:

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

These are: (a) a market correction will tend to be contained to 61.8% of the previous wave if it is part of a base or top pattern, and (b) a market correction will tend to be limited to 38.2% of the previous wave if the main trend remains intact. The beauty of this hypothesis is that clear information is produced if the 61.8% and 38.2% boundaries are penetrated: it says something about fundamentals.11 For example, in the autumn of 2002, having experienced a 38.2% fall, the Dow was faced with the possibility of confirming a change to the foundations of the modern economy. It declined to do so. In June 2003, the Dow retraced more than 38.2% of the 2000-02 bear and, in so doing, signalled both a strategic bull and a cyclical shift in fundamentals. What seems to happen is that the Golden Measure provides very specific constraints on a market’s ability to regress.12 A market can tolerate some slippage in relation to fundamentals, but it cannot turn into a full-blooded reversal unless the fundamentals so warrant. Indeed, it seems that the critical dividing line between a technical correction and a fundamental reversal is literally defined by the Golden Measure.

Why not?

Figure 7 shows some of the extraordinary power of the Golden Measure in relation to the equity market. The chart tracks the Dow Jones Industrial Average from January 1999 to mid-June 2008. The first thing to note is the fact that the 2000-02 bear consisted of an absolute drop of almost precisely 38.2%. This created the level from which the subsequent bull started. The second point is that the base pattern that developed prior to the renewed bull was bounded on the upside at 38.2% of the 2000-02 bear. When this level was broken in June 2003, it was the signal that a strategic bull was underway. Finally, the January to March 2008 sell-off lows coincided almost precisely with a 38.2% retracement of the 200207 bull. In my view, it is much more than just an accident that this was the level at which the index supported. At the time of writing (June 2008), this boundary remains intact; but the implications of a break would be profound. Figure 7:

The Golden Measure as a limit to decay The influence of the Golden Measure, both in reversal patterns and in subsequent corrections, is broad rather than absolutely precise. In fast moving markets, the calculated levels may act as a centre of gravity rather than as an absolute boundary. Nevertheless, the influence is so persistent and pervasive that it cannot be an accident. Indeed it suggests two strong working hypotheses.

The Golden Measure therefore combines linear logic with nonlinear symbolism to create a truly universal metaphor for change and transformation within the context of Life. Moreover, it reveals itself regularly and persistently in the non-linear dynamics of the particular self-organising system that we know as a financial market.13 The information that it can provide both in setting targets and in assessing whether a market is correcting or reversing is extraordinary. In my opinion, it plays the most crucial of roles in defining the direction of a market trend. But why is it treated with such scepticism? The first reason is that the philosophical context in which the Golden Measure originally found expression has resulted in the Measure being either sanctified or trivialised. It has been sanctified insofar as its religious connotations have resulted in it being regarded as an unknowable ‘law’. However, the Golden Measure is no more unknowable than any other mathematical concept. Its value lies in its representation of reality and the use to which it can be put. Moreover, it can be the subject of statistical research.14 Meanwhile, the Golden Measure has also been trivialised by its association with the so-called Fibonacci series. In this series (ie, 1, 1, 2, 3, 5, 8, 13, 21, 34, ....etc.), where each number is the sum of the two preceding numbers, the ratio between successive numbers tends towards 0.618. The series, which is named after Leonardo of Pisa,15 is certainly an important example of the Golden Ratio, and it can itself be found in Nature;16 but it is definitely not the ultimate statement about the Golden Ratio. Indeed, many complainants justifiably ask why they should accept that the ‘Fibonacci ratio’, which helps to demonstrate the ability of rabbits to procreate, is applicable to financial markets.17 Second, huge numbers of traders and investors are basically unwilling to accept either that financial markets are group phenomena, or that they themselves will be in any way influenced by the behaviour of others. The issue centres on cultural beliefs about rational behaviour by individuals - beliefs, moreover, that are embedded as assumptions in economic theory. Accepting the operation of the Golden Measure directly confronts these beliefs. But third, and most importantly, the existence of the Golden Measure is usually missed simply because of its own characteristics. The Measure allows a constant relationship between three variables to be expressed in terms of two of those variables - that is, the Golden Proportion can be expressed as X:Y::X:(X+Y). In a sense, therefore, the third variable (which is, somehow, more than the sum of its parts) is ‘hidden’. So, while rational observers easily recognise the existence of two of the variables, they can just as easily miss the third.18 And they do.


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

1

To put it another way, the phenomenon simultaneously expresses the idea of unity (Golden Number), duality (Golden Ratio) and trinity (Golden Proportion).

2

These proportions are taken from Trudi Hammel Garland, Fascinating Fibonaccis: Mystery and Magic in Numbers, Dale Seymour Publications, 1987.

3

See, for example, Matila Ghyka, The Geometry of Art and Life, Dover Publications, 1977.

4

Certainly there is evidence that it was known about in the third millennium B.C.E., because it was used in the construction of the great pyramid at Gizeh.

5

Such systems are called, among other things ‘autocatalytic’.

6

See Ilya Prigogine and Isabelle Stengers, Order out of Chaos, William Heinemann, 1984.

7

Metaphors and symbols - particularly those that ‘work’ - point to great truths, even though they are not themselves the truth. Since the ultimate truth (whatever that may be) is - by definition - unchanged and unchanging, old metaphors and symbols may still be applicable, provided that we can understand them.

8

For a simple analysis of this phenomenon, see Arthur Koestler, Janus: A Looking Back, Hutchinson, 1978. A more recent exploration can be found in Howard Bloom, The Lucifer Principle, Atlantic Monthly Press, 1995.

9

Ralph N. Elliott, Nature’s Law: The Secret of the Universe, Elliott, 1946. Reprinted in Robert Prechter (Ed.), The Major Works of R.N. Elliott, New Classics Library, 1980.

10

William D. Gann, Profits in the Stock Market, Lambert-Gann, 1981.

11

In its simplest form, the calculations are constructed in relation to price movements that are measured from turning point to turning point. More sophisticated analysis, however, requires that the calculations are made from the point at which important information shocks actually impact a market. But that is another subject. See Tony Plummer, ‘Some thoughts on the mathematics of turning points’. Market Technician, March 1999.

12

The Golden Measure also acts as a brake on the market’s ability to evolve. This brake reveals itself in the fact that the Golden Measure can be used to calculate expansion objectives. See Tony Plummer, op. cit.

13

Indeed, the presence of the Golden Measure in financial markets is evidence that Nature effectively treats psychological groupings of human beings as a living organism.

14

See, for example, Roy Batchelor and Richard Ramyar, ‘Fibonacci: fact or fiction?’ Professional Investor, December 2006/January 2007. Unfortunately, the authors seem not to understand the intricacies of the phenomenon that they purport to be testing. Not surprisingly, therefore, they conclude that they have “strong reservations about the value of price targets based on magic numbers.”

15

Leonardo of Pisa, writing under the name of Fibonacci, published a book called Liber Abaci (or Book of Calculations) in 1202. The book introduced the Hindu-Arabic decimal system, which includes zero as the first digit in the sequence, to Europe. The nineteenth century mathematician Edouard Lucas attached the name Fibonacci to the Golden Ratio based sequence contained in the book, and the name stuck.

16

Two outstanding examples are spiral phyllotaxis (the number of branches per circuit around the stem of a plant) and the distribution of seeds over a sunflower’s disk.

17

The Fibonacci Sequence was used to demonstrate a very contrived example of the breeding capabilities of rabbits.

18

And it is, indeed, true that people clearly recognise the existence of opposite polarities (such as night and day, male and female, etc.) but somehow manage to miss (or not understand, or somehow dismiss) the third force that links the polarities together. However, it is the third, higher-level, force that powers the circulation of a system from one polarity to the other, giving it meaning, purpose and functionality.

Copyright, Tony Plummer, 2001-2008

155


156

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Elliott Wave, Chaos Theory and Fractals Julien Camberlin, CFTe, MFTA, CEWA Lev I

Article originally featured in Market Technician 79 (September 2015)

Introduction People have always been fascinated by financial markets and the possibility of predicting their future movements. First scientific studies date back to the beginning of the 20th century. Understanding socionomics requires comprehending the contrast between the following postulations: Bachelier (1900), defended his thesis «Theory of speculation», stipulating that markets are random while Dow (1900) developed his so-called «Dow Theory». One of the most important points of his theory is that financial markets display tendencies and evolve in successive waves.

Figure 1: Elliott wave: self-similarity of a fractal

If we follow the ideas of Bachelier, markets are random and cannot be forecast. If we follow the ideas of Charles Dow, markets follow trends (meaning they are not random) and it can be possible to forecast future movements by looking at past movements. In 1934, Ralph Nelson Elliott discovered that financial markets organise in waves, respecting certain rules and guidelines. These waves are self-similar patterns (Figure 1) and it was the first mention of the idea of fractals in the financial markets which he called Elliott waves.1 In his second book,2 Elliott discussed for the first time the concept of Fibonacci relationships in financial markets. Elliott noted that Fibonacci ratios appear quite often in the relationship between waves of the same degree, although, he never established clear rules for using the Fibonacci ratios.


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

In the 1960s, two new concepts were developed that help in our understanding of financial markets: •

The first is Chaos Theory. Edward Lorenz was working on climate forecasts. He discovered that certain complex systems with two states differing by imperceptible amounts may eventually evolve into two considerably different states that, in the long term seem unpredictable.3

The second is fractals invented by Mandelbrot. Mandelbrot gave the name of fractal to the mathematical sets that have self-similar patterns. Mandelbrot explains how financial markets organise in fractals and developed a fractal model that approaches the behaviour of financial markets.4

Frost & Prechter published in 1978 the first version of Elliott Wave Principle,5 which is still today the reference for Elliott wave study. They presented a very smart system of multiple wave confirmation called multiple wave relationships.6 This system proposes that, as all levels of waves are developing at the same time, the end of waves of different levels could then be forecast with concentrations of Fibonacci ratios. Multiple wave relationships are a very interesting tool, but they recognise that “If a complete method of ratio analysis could be successfully resolved into basic tenets, forecasting with the Elliott wave principle would become more scientific”. 7 In the 1990s, Edgard E. Peters published different books on Chaos Theory in financial markets, confirming that markets are chaotic systems.8

Random system vs deterministic system A random system is a mathematical model composed of a succession of random walks. This means that future events only depend on the present event by its position. There is no memory effect and no forecast can be made - just like the toss of Figure 2: Random System

Psychology and Markets

Systematic Trading

Figure 3: Example of deterministic system: sinusoidal

Chaotic system There is no unanimous definition for chaos systems, but among the different points defining them, there are: 1

Sensibility to initial conditions. This is the butterfly effect, when a small change in initial conditions can ultimately lead to completely changed outcomes. The model will progressively deviate.

2

Memory effect: chaotic systems have a memory effect which is between 0% and 100%. If we tend to 0%, the system becomes more random. If we tend to 100%, it becomes more deterministic.

3

Strange attractors: To give a concrete example of an attractor, let us imagine a pendulum circling around its point of stability (the centre of the ‘circle’). This point is the attractor. If there is an external element, like a cat kicking the pendulum, it disturbs the model. However, once the kick stops, the system starts again to be attracted by the attractor. In most chaotic systems, more complex attractors called strange attractors are found. Strange attractors are attractors with more complicated structures. They are not single points, but organise most of the time in fractal sets. Hence they are sometimes called fractal attractors. “The solution set of a dynamic system is generally called an attractor, and the attractor of a chaotic system is specifically called a strange attractor. Strange attractors have a fractal structure.”10 Strange attractors are considered as the limit point of the system.

4

Fractals: As explained earlier, Benoit Mandelbrot gave the name of fractal to the mathematical sets that have selfsimilar patterns. They look the same from close up as they do from far away (Figure 1).

Construction of Mandelbrot financial market fractal:

a coin where no forecast for the next toss can be made from the past toss. A deterministic system is a system that always gives the same result from an input. The memory effect is 100%. An example is a sinusoidal pattern (Figure 3).

Mandelbrot11 describes how to draw a financial market fractal. The interesting point here is that he needs a start point and an end point to draw his fractal (Figure 4). In fact, if end point (A’, t’) is not known, it is not possible to draw the fractal. This means that in financial markets, the end point must also exist. This point is the strange attractor. This is true for the general vector in green on the figure below, but it is also true for the smaller ones in other colours.

157


158

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Every degree of Elliott wave developing at same time (Minuette, Minute, Minor…), moves to its own strange attractor. When the final attractor is hit, the wave is completely developed, the equilibrium between buyers and sellers forms and the price can change direction. This means that the price does not stop at Fibonacci ratios, but rather vibrates in line with their direction.

Figure 4: Construction of a fractal

Fibonacci strange attractor model As seen before, fractals are self-similar patterns. These patterns can have different forms. Elliott waves are the patterns in force in financial markets. Impulse waves (waves developing in the direction of the main trend) are always subdivided into 5 waves: 3 impulse waves developing in the direction of the main trend and 2 corrective waves developing against the main trend (Figure 1). In the Elliott Wave Principle, Frost and Prechter give different guidelines to validate Elliott wave counts. A part of these guidelines is dedicated to ratio analysis between waves. To clearly understand the organisation of the ratios, a brand new idea will be proposed here: the construction of the Fibonacci Strange Attractor Model. This model will be compared to the statistics of ratios between Elliott waves following Elliott wave guidelines proposed by Frost and Prechter.

Where is the strange attractor? In a chaotic system, a strange attractor is the set where the system is in equilibrium. A financial market is in equilibrium when buyers and sellers are in equilibrium. This is at the end of a wave (top or bottom), when price changes direction, this is where we can find the strange attractor.

Six markets have been chosen: two currencies, two indices, and two commodities. For S&P500 and Cocoa, the futures markets have been chosen. They are more liquid and more complete Waves developing after cash market close are important for the counts. For Gold, EURUSD and the USDJPY the Forex market has been used for the same reason. For the DJIA cash markets were used (futures markets did not exist between 1957 and 1966). Table 1 shows the different information about the sets of data of historical ratios:

Each wave at every degree has its own strange attractor. The strange attractor is active even if the price is slightly disturbed. If it is highly disturbed, the price will change its strange attractor and will be attracted by another one. Price formation is a compromise between determinism and randomness (the memory effect is between 0% and 100%). Table 1: Historical ratios Market

Start Date

End Date

Pivot Points

Corrective 3 waves

Triangle

Impulse X3

Impulse X1

Impulse X5

Leading/ Ending Diagonal

S&P500

24-Mar-00

18-Feb-11

121

26

1

12

2

0

1

DJIA

22-Oct-57

2-Feb-66

76

55

4

17

1

1

1

USDJPY

19-Apr-95

1-Feb-12

211

11

0

10

0

5

1

EURUSD

26-Oct-00

15-Jul-08

70

10

0

6

0

2

1

Gold

21-Sep-99

22-Aug-11

90

18

3

7

1

1

2

Cocoa

12-Dec-00

4-Mar-11

100

15

0

8

3

0

0

668

135

80

60

7

9

6

Total


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

Figure 5: Cumulative curve of ratios between waves following Elliott wave guidelines

Elliott wave counts have been done on daily and weekly charts, following Elliott wave rules and guidelines described in Elliott Wave Principle (channelling, alternation etc.). From these counts, all the ratios that appear between the waves described in the ratio analysis section are taken and added on a cumulative chart (Figure 5).

Fibonacci We can see in Figure 5 that we have spikes at certain levels. Most of these spikes are situated at Fibonacci ratios: 38.2%, 50%, 61.8% etc. Before we examine the construction of the model, it is important to review what exactly a Fibonacci ratio is. The Fibonacci ratio is also called the golden section or divine proportion. The golden section is when C is to A what A is to B (Figure 6) and the ratio is 61.8% or its inverse 161.8%. Figure 6: Fibonacci ratio - Golden section

between addition and multiplication because 1 + PHI =1/ PHI. This is the growth ratio. The model will be constructed from the Fibonacci ratios suite that is … .382, .618, 1.00, 1.618, 2.618 …

Construction of the model The following arithmetic operations from the Fibonacci ratios suite are calculated: addition, subtraction, division, multiplication, exponent, square root, inverse and opposite. In order to represent this on a chart, the calculation has not been done to infinity but a maximum of three combinations has been set. It is worth noting on Figure 7 that, instead of a random repetition, there appears a fractal repetition of the ratios. One could say here that the process of construction of the model is too vague, but this repetition is due to the fact that the golden section is the ratio which is the link between addition and multiplication, and on the long run, the results of the calculations, which we can call echoes, Figure 7: Fibonacci Strange Attractor Model

A lot of information about Fibonacci and the golden section can be found, but what is most important for the construction of the model is that the golden section is the ratio that is at the cross-roads

159


160

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

will always end up with the same repetition, with spikes at the same places, whatever Fibonacci ratios are taken or the type of calculation that will be done. This is the magic of Fibonacci! Figure 8 is the combination of the Fibonacci Strange Attractor Model and the cumulative curve of ratios between waves following Elliott wave guidelines. We clearly see the similarity of the statistics and the model. The model has two kinds of ratios: ratios that are the Fibonacci suite in blue (…, .382, 0.618, 1, 1.618, 2.618,…), that we call Fibonacci ratios, and clusters of ratios in red that are between the Fibonacci suite, that are formed by echoes of the calculations and that will be called derived Fibonacci ratios (.447, .472, .486, .5, .514, .528, .553, .724, .764, .786 …). Figure 8: Combination of the Fibonacci Strange Attractor Model and the cumulative curve of ratios between waves following Elliott wave guidelines

The Fibonacci ratios have much more importance than the derived Fibonacci ratios. Ratio .50 is not a precise Fibonacci ratio, but a cluster from .447 to .553. The inverse of 2.00 being .50, it is also a cluster around 2.00 and not a precise Fibonacci ratio. Concerning the ratio of .764, it is also a cluster of different ratios between .724 and .809 and the inverse is between 1.236 and 1.382. This means for practical use that we will allow more uncertainty for 50% or 76.4% retracement than for 61.8% or 38.2% retracement. This model is the representation of the strange attractor that is in force in financial markets. It can be used to improve guidelines for the validation of Elliott waves.

1

Elliott, R.N., The Wave Principle, 1938

2

Elliott, R.N., Nature’s Law - The Secret of the Universe, 1946

3

https://en.wikipedia.org/wiki/ Edward_Norton_Lorenz

4

Mandelbrot B., Hudson R.L., The (Mis)behaviour of Markets: A Fractal View of Risk, Ruin and Reward, Profile Books, 2008, p 221

5

Frost, A.J., Prechter, R., Elliott Wave Principle, Tenth Edition, New Classics Library, 2005

6

Frost, A.J., Prechter, R., Elliott Wave Principle, Tenth Edition, New Classics Library, 2005, p. 145

7

Frost, A.J., Prechter, R., Elliott Wave Principle, Tenth Edition, New Classics Library, 2005, p. 146

8

Peters, E.E., Chaos and order in the capital markets, Wiley Finance Edition, 1991; Peters, E.E., Fractal Market Analysis: Applying Chaos theory to Investment and Economics, Wiley Finance Edition, 1994

9

https://en.wikipedia.org/wiki/Chaos_theory; http://fractalfoundation.org/resources/what-is-chaos-theory; https://en.wikipedia.org/wiki/Chaos_theory; https://en.wikipedia.org/wiki/Attractor

10

Brown, C.T., LieboVitch, L.S., Fractal Analysis (Quantitative Applications in the Social Sciences), SAGE Publications, First edition (April 14, 2010), p. 63

11

Mandelbrot B., Hudson R.L., The (Mis)behaviour of Markets: A Fractal View of Risk, Ruin and Reward, Profile Books, 2008, p119, p 210-211, p 220-221

Conclusion In the last 120 years, science and mathematics have helped to further our understanding of financial markets. With the new concept of Fibonacci Strange Attractor Model, it has been possible to show how Fibonacci ratios organise in a fractal way, that are the spectrum of the strange attractors in force in financial markets. This confirms the existence of two kinds of ratios: precise ratios (Fibonacci suite) and clusters of derived Fibonacci ratios (echoes). This model gave an explanation for .500, 2.000, .724, .764 and .786 ratios. Further study of Elliott waves is also of interest with the same technique applied to time ratios, which could improve market timing.


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

161


162

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

CHAPTER EIGHT

Gann Analysis, Cycles and Forecasting Articles in this chapter 164 Autumn Panics: A Calendar Phenomenon Christoper Carolan

168 Channels and Cycles Brian Millard

171 William Gann’s Law of Vibration Tony Plummer

180 Fibonacci - Lucas Numbers, Moon Sun Cycles and Financial Timing David McMinn


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

CHAPTER EIGHT INTRODUCTION

Patricia Elbaz MSTA

Introduction William Delbert Gann (1878-1955) was said to be an outstanding stock and commodities trader. He was highly praised by Richard Wyckoff, The Ticker Magazine, 1908, a respected Wall Street trader in the 1900s who recorded gains made by W.D. Gann trading commodities (See Chapter 2 for Dow and Wyckoff). Gann was also a teacher on how to make speculation a profitable profession using his theory. He wrote seven books including ‘45 Years in Wall Street’, 1949 and produced two courses on trading the stock and commodity markets. While many books support William Gann’s success there are also controversies regarding his work and some critics argue that there is no apparent evidence of Gann being a successful trader. Alexander Elder in his book ‘Trading for a Living’ 1993, was one of the critics, quite sceptical about Gann’s claimed success. There are many aspects of Gann’s work: Gann geometric angles and percentages, Square of 9, swing charts, astrology, the squaring of Price and Time. The Gann fan lines have proved to be highly significant in all markets, FX, equity, commodity & bond. The Gann lines, computerized today rather than drawn on arithmetic paper, still give a powerful indication of how impulsive a trend is and can be used as support and resistance levels. A recent analysis on GBP/USD weekly charts, using Gann lines from the major peak of Nov 2007 above $2.10 show that the 1x1 line acted as major resistance in 2014 through to 2018. It has been challenged in Sep 2019. The level at $1.31 is seen as pivotal for the long term trend.

Selected Articles Our first article ‘Autumn Panics: A Calendar phenomenon’ by Christopher Carolan in 1998 focuses on short term equity panics. The question asked is: Is there a correlation between the lunar calendar and the stock market panics of 1929 and 1987? Carolan concludes that calendars are complex mechanisms and that calendar research ‘must recognise the importance of both lunar and solar calendars’. From lunar calendars we move on to ‘Channels and Cycles’ by Brian Millard in 1999, which looks at the relationship between the average length of a trend and the profit which can be expected from this trend. Millard expands on his model of price movements, using moving averages with channel analysis. The model is then applied to the FTSE 100 in detail. Millard’s scientific background explains his strong interest in cycles in market data and in developing a software to allow traders to apply his method. Our next author has over 40 years’ experience in financial markets and his book ‘Forecasting Financial Markets’ in its

Psychology and Markets

Systematic Trading

6th edition continues to be extremely popular for students as well as professionals in finance. Tony Plummer’s article ‘William Gann’s Law of Vibration’ is based on his book ‘The Law of Vibration’ published in 2013. The article looks at the Law of Vibration in US output and touches on the 54-year Kondratyev price cycle and the Kitchin cycle. One of the conclusions was that ‘every 35 years a shock emerges to end a whole socio-economic era’. Our final article ‘Fibonacci - Lucas Numbers, Moon Sun Cycles and Financial Timing’ by David McMinn looks closely at the relationship of the above. Can Phi and Fibonacci numbers provide a scientific basis in market forecasting? This question is explored in depth. When reading articles on W.D. Gann we often come across this quote from Ecclesiastes: ‘That which has been is what will be, that which is done is what will be done, and there is nothing new, under the sun.’ It is highly relevant to Gann’s beliefs that the law of action and reaction means that history must repeat itself. Markets are driven by human beings and they repeat their behaviour cycle-aftercycle. Applying Gann’s theory has proved to be a strong signal and used with other technical analysis indicators continues to be highly relevant today. In this case, there is certainly cause for saying that there IS something new under the sun.

163


164

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Autumn panics: A Calendar Phenomenon Christopher CaroIan

Article originally featured in Market Technician 32 (July 1998)

The crash of the Hong Kong stock market in October 1997, with its obvious parallels to similar events in the U.S. in 1987 and 1929, once again raises the spectre of October as a dark and ominous month for stocks. Is it merely a coincidence that these three crashes all occurred in October? Is there a timing pattern among autumn panics that would be useful to market participants? This article expands upon the observation, originally contained in Chapter 1 of the author’s book, ‘The Spiral Calendar’1, outlining the correlation between the lunar calendar and the stock market panics of 1929 and 1987. This paper examines how the 1997 Hong Kong panic conforms to that earlier model, as well as examining the great autumn panics of the 19th century. Finally, a look at the peculiar international character of panics, and its implications for the possible causes of these panics.

Figure 1: Autumn Panics, Lunar Aligned

Definition of Terms Panic: The focus of this article is on short-term equity market panics. The crashes of 1929 and 1987 are obvious examples. I define these panics as one-to-three day, freefall drops of approximately 20% in the major averages. The term “panic” is preferred over “crash” as the definition of panic stresses the suddenness and irrationality of the event. Panics were originally ascribed to the god Pan simply because there were no obvious fundamental causes for their occurrence. Collapse: Collapse is used to signify the larger macro market decline lasting weeks or months within which the panic occurs. An example would be the Hong Kong panic of October 1997, occurring within the larger Asian equity and currency collapse that ran from July 1997 to January 1998. Annual Lunar Calendar: The annual lunar calendar used here is based on the Babylonian calendar, which was the model for the later Jewish calendar. This annual lunar calendar labels the date of the first new moon following the spring equinox as month one, day one; or 1-1. The following date is 1-2. The date of the second new moon after the spring equinox is 2-1, etc... The difficulty with annual lunar calendars, and one of the reasons for their abandonment, is that the solar year does not have an even number of months. Thus, some years in an annual lunar calendar have 12 months, others 13. For our purposes, which focus on the Autumn months, this issue is inconsequential. All calculations use Eastern Standard Time to determine the dates of the lunar phases. In 1992, this author demonstrated how the panic dates of “Black Tuesday,” October 29, 1929, and “Black Monday,” October 19, 1987 occurred on the same annual lunar calendar date, 7-28. Additionally, the other similar points in the comparisons of those

two years, the spring lows, summer highs and autumn failure highs all occurred within one day on the lunar calendar. Figure 1 shows those years in a chart aligned with the lunar calendar, where similar lunar dates are juxtaposed above each other. The panics are marked with arrows. The other similar features are denoted with dashed lines. The chart also includes Hong Kong’s Hang Seng index for the panic year 1997. Table 1: Largest 1 day % decline 1

26 October 1987

-33.33%

Hang

P

2

18 October 1987

-22.61%

DJIA

P

3

5 June 1989

-21.75%

Hang

4

20 October 1987

-14.90%

Nikkei

P

5

28 October 1997

-13.70%

Hang

P

6

28 October 1929

-12.82%

DJIA

P

7

16 October 1989

-12.81%

DAX

8

29 October 1929

-11.73%

DJIA

P

9

19 October 1987

-11.12%

Hang

P

10

22 May 1989

-10.78%

Hang


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

These price moves are extraordinarily large over a very short period of time. Are these panics the largest such declines, or do we selectively remember the October panics and forget those of other months? A scan of daily data of the Dow Jones Industrial Average from 1915, the Hang Seng index from 1980, The Japanese Nikkei index from 1950, and the German DAX index from 1960 for the 10 largest, single-day percentage drops is shown in Table 1. Seven of those ten declines were days associated with one of the three panics. Two of the others, the spring 1989 declines in the Hong Kong market, were tied to a fundamental news event, the Tiananmen crisis in China. The final entry is from the German market during the “mini-crash” of October 1989, an October event similar to the others, but smaller in magnitude. The point to stress here is that in their breadth and ferocity, these panics lie outside the boundaries of normal price action. There are no other comparable one-to-three day declines of this magnitude in the data. They represent the very largest percentage drops in the database. This is not normal market behaviour. What else ties these events together? The panics occupy virtually identical positions on the annual lunar calendar. Table 2: Autumn Panic 1 day % change Lunar Month

7

7

7

7(8)

Lunar Day

27

28

29

30(1)

DJIA

1929

-12.8%

-11.7%

12.3%

5.8%

DJIA

1987

Closed

-22.6%

5.8%

10.1%

Hang Seng

1997

-5.8%

-13.7%

18.8%

-3.7%

Table 2 shows the percentage declines for each panic in the key four-day time span around the lows. The lunar dates 7-27 and 7-28 are the “dark days,” encompassing the various Black Tuesdays of N.Y. in 1929 and Hong Kong in 1997, and the Black and Blue Mondays in N.Y. in 1987 and 1997 respectively. In each case, lunar date 7-28 marked the end of the panic and the next two days, 7-29 and 7-30 (or 8-1, some lunar months have 29 days, others 30) saw significant retracement rallies. Table 3 groups the data into two-day segments and includes the percentage of these retracement rallies. This table shows the striking similarity of these panics and how that similarity conforms to the annual lunar calendar. Table 3: Autumn Panic 2 day % change Lunar Month

7

7

Lunar Day

27-28

29-30

% retrace

DJIA

1929

-23.6%

18.9%

62%

DJIA

1987

-22.6%

16.6%

57%

Hang Seng

1997

-18.7%

14.4%

61%

Table 4: Spike Low-New Moon Differential Panic Low EST

8-1 New Moon

Diff in hours

1929

29-Oct

14:45

1-Nov

8:00

-65

1987

20-Oct

11:30

22-Oct

13:26

-50

1997

29-Oct

9:15

31-Oct

6:01

-45

Table 4 pinpoints the precise timing of the panic lows on the lunar calendar. The timing from 1929 is gathered from the news

Psychology and Markets

Systematic Trading

accounts that described stock prices as rallying sharply off their lows in the last fifteen minutes of trading on Black Tuesday, October 29. The 1987 and 1997 times are from available databases for the Dow Industrials and are Corrected to Eastern Standard Time. The table also shows the date and time of the nearest lunar phase, the eighth new moon on the annual lunar calendar, as well as the difference in hours between the stock market’s low and the moon’s phase. The timing of these three great panic lows is within twenty-four hours of each other. In other words, all three lows fall within the same one-half of one percent of the calendar year.

A review of the Pre-1915 Autumn Panics: The Panic of 1907 The so-called panic of 1907 does not fit our short-term panic criteria. There was no market decline of approximately 20% in the span of one-to-three days. The largest, single-day declines were 3% in the Dow Jones Industrial Average during the collapse. There was a collapse and coincident banking panics, most of which occurred in October of that year. Sobel, in Panic on Wall Street2, describes the ending of the collapse. J.P. Morgan put together his plan to save the banking system on November 3-5, 1907, 7-28 through 7-30 on the annual lunar calendar. After being closed for Election Day on November 5 (7-30), stocks rallied strongly on lunar 8-1. The crisis was over. The timing of the end of the crisis is consistent with the lunar panic model. The day Morgan realised the banking system was not going to fail, he put into motion a plan to save the banks, which ultimately arrested the decline. That day was lunar 7-28, the same date as the lows for the later 20th century panics. The Crash of 1873 September 18 and 19, 1873 were labelled “Black Thursday” and “Black Friday” in the collapse of 1873. The Friday selling took prices of major stocks from 5 to 25% percent below Thursday’s already collapsed levels. This panic was considered the greatest on Wall Street until 1929. The news accounts describe the same type of free fall and despair as the 20th century counterparts. The annual lunar calendar dates of “Black Thursday” and “Black Friday” were 6-27 and 6-28, one month earlier, but exactly the same lunar days as the 20th century examples. News accounts describe a temporary bottom late on Friday. Saturday, September 20 brought renewed selling and the closure of the exchange after a shortened two-hour trading day. The stock exchange remained closed for a week thereafter. Though on Monday 22nd September prices rose sharply in trading in the streets. The timing of the 1873 Autumn panic is consistent with the 20th century results, though exactly one month earlier. The Crash of 1857 The collapse of 1857 was not a stock market free-fall in the sense of the 20th century panics outlined above. It was a very sharp drop in stocks over a period of nine weeks, accompanied by a number of runs on banks, persistent pressure on the banking system, and sharply rising interest rates. Also, it was international in scope, a fact we’ll address later. Though the selling in the equity markets did not climax in a free-fall panic, the pressure on the banking system did, as the N.Y. banking panic broke out on October 13 and mayhem continued for two days thereafter. Sobel, in Panic on Wall Street3, quotes George Strong writing on October 15, “Wall Street blue with collapse. Everything flaccid like a defunct Actina.” On the annual lunar calendar, October 13 and 14, 1857 are 7-27 and 7-28, the same “dark-days” as the 20th century examples. Causation: The correlation between the annual lunar calendar and the timing of the three 20th century panics as well as the supportive data from the 19th century does not prove that an annual lunar calendar position is the cause of those panics. A few examples of anything cannot statistically prove a hypothesis. However, it should be

165


166

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

realized that each occurrence is not a 50-50, or true/false proposition. If the Hong Kong panic had occurred on any of the 360 days of 1997 other than lunar 7-27, 7-28, 6-26, or 6-28; then this model would be effectively discredited. Yet the 1997 Hong Kong panic climaxed 5 hours after the timing of the 1978 panic and 20 hours after the 1929 panic on the lunar calendar. Previous theories explaining panics have not fared well when the next panic came along. In the 19th century, it was widely believed that panics occurred in October specifically because banks’ cash positions were weakened as farmers were paid for the new crop. Yet today, agriculture makes up a much smaller fraction of the world economy than before, yet October panics are still with us. The Federal Reserve System was set up in the belief that if banking panics were prevented, stock market panics would cease to exist as well. That casual theory was disproved by the 1929 crash. The 1929 panic was blamed on low margin levels, yet 1987 happened anyway. In 1987, the finger was pointed at program trading. However, the 1997 panic occurred without any appreciable role by program traders. The lunar calendar model of panics, alone among theories, not only survived the next panic intact, but its basic tenet was remarkably affirmed by the precise timing of the 1997 low. The timings of financial collapses do not show a pattern. The 1997 Asian collapse began in July, while the crisis of 1987 and the collapse of 1857 began in August. The 1929 and 1873 examples began in September. Yet in each case, the start of the collapse did not result in immediate widespread panic. Those panics seem to wait for a particular time period on the calendar, the 27th and 28th days of the autumn lunar months, usually October, but in one instance September.

The International Question: The international character of financial crises has been a difficult problem for those who have sought to ascribe causes to collapses and panics. Kindleberger, in Manias, Panics and Crashes writes, “Time and again, observers like Juggler, Mitchell and Morgenstern have observed that financial crises tend to be international, either running parallel from country to country or spreading by one means or another from the country where they originate to other countries4.” And “What is remarkable is that securities prices do the same even when only a few securities can be said to be truly international, that is, are traded on several markets, their prices joined by arbitrage. In 1929 all stock markets crashed simultaneously; the same was largely true in October 1987... It is striking that share prices behaved in parallel almost sixty years apart, even though share prices were thought not to have been integrated in the 1920s as they were in the 1980s5.” The panics of 1987 and 1997 highlighted the international quality of panics. Traders the world over saw these markets dive and then rally in unison. In this wired world, that interconnection is not so extraordinary, though Kindleberger is surprised by the international nature of the 1929 collapse. An examination of the 1857 collapse is more revealing. Kindleberger notes, “What is striking is the concentrated nature of the crises... Clapham observes that it broke out almost at the same moment in the United States, England, and Central Europe, and was felt in South America, South Africa and the Far East.”6 Aside from the international nature of the macro collapse, the 1857 collapse affords a unique, controlled database of market behaviour in the “dark days” of lunar 7-27 and 7-28 on two continents. In 1857, the Atlantic cable linked America with England by telegraph. In the early days of the collapse, the telegraph cable failed and all communication was done by ship for the remainder of the crisis. The London Times and The New York Times from the period leading up to and through the N.Y. banking panic provide striking

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

evidence of two markets in distress. Wall Street began its rally from the depths of the collapse on October 13, 1857 (lunar 7-27) at the same time the banking panic broke out in N.Y. Table 5 is reprinted from The N .Y . Times of October 14, illustrating the sharp rise in prices underway as contrasted with the lows of October 13. I’ve added the column on the right showing the month’s-end prices. Some issues had made their lows earlier in September, but others were at or near their lows on October 13. What’s clear is that prices began to rally from their depressed levels on lunar 7-27, coincident with the outbreak of the banking panic. This sequence parallels the 1987 Table 5: N.Y. Stocks, October 1857 Aug 22

Low Since

Oct 13

Oct 14

Oct 31

N.Y. Central

77

50

58

61

65

Galena

86

52

55

62

64

Rock Island

90

53

58

64

66

Delaware

115

77

78

81

94

Panama

90

60

65

67

72

Reading

67

24

33

36

30

N.Y. States Sixes

112

90

90

90

97

Missouris

78

59

60

67

68

Virginias

90

66

67

81

78

III. Cen. Bonds

98

50

51

59

68

Erie’s 1883

80

50

50

50

Erie Share

28

7

8

11

13

Source N.Y. Times, Oct 14. fractions omitted

experience, when U.S. bond prices began a sharp rally from their lows on lunar 7-27, coincident with the outbreak of the stock panic. At this same time, Europe was aware of, and sharing in, the collapse in America. In the week leading up to October 13, the Bank of England raised their discount rate twice, while Paris, Hamburg and Amsterdam each raised their rates once. Though debt and equity prices traded down sharply, there was no free-fall panic. London stocks and debt bottomed decisively on October 13 at the beginning of the trading day. The London Times of November 2, 1857 summarized the events of October and printed the table of prices labeled here as Table 6. To that table is added the date of the month’s low for each security. Here is the commentary accompanying the table. “The range of Consols (government debt) has been unusually extensive, showing a difference of 4 percent between the highest and lowest prices, although at the conclusion (of the month) the market has returned to the precise position in which it stood at the commencement... In railway shares the fluctuations have also been violent, and the rebound, except in a few cases has not been equal to that of the funds.”


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Table 6: London Stocks, October 1857 Sep 30

Oct High Oct Low

Oct 31

Low Date

90.25

90.5

86.5

90.5

13-Oct

Brighton

102.5

103.5

100

103

13-Oct

Caledonian

85

86

76.5

83.5

29-Oct

Eastern Counties

57.5

58.5

51.5

54.5

13-Oct

Great Northern

98

98.5

92

93.5

13-Oct

Great Western

54.5

55.25

50.5

51

29-Oct

London & NW

97.5

98.5

93.5

96.75

13-Oct

Midland

83

83.25

79.5

83.25

14-Oct

Lancs and Yorks.

95.5

97.25

91.5

93.25

13-Oct

North Staff.

13.25

13.75

12.75

13.75

15-Oct

South Eastern

65.5

66.25

61.5

64

14-Oct

South Western

90.5

91

87.5

89.5

13-Oct

North E. Berwick

92.5

93.5

88.5

91.5

13-Oct

Ditto York

73.5

80

75.5

78.5

13-Oct

British Debt Consols Railway Shares

Source: London Times

The London Times offered this account of the trading in debt on October 13 in its October 14 edition. “The fluctuations in the funds today (Oct. 13) have again been most rapid and extensive. The market opened with a great weakness at a fall of nearly one and a quarter percent from the heavy prices of last evening. But there was subsequently a considerable reaction and a more healthful tone became apparent in all departments of business.” Now, here’s the account of stock trading on October 13 from The New York Times of October 14. “The stock market this afternoon advanced from I to 3 percent, the conviction being general that the basis of business would be changed tomorrow and that a large amount of money held in abeyance since the panic first paralysed confidence will be set free now that the worst is known...” The cause of the market low in New York on October 13 is ascribed to the banking panic, yet London bottomed on the same day. The selling, motivated by fear, was pervasive on both sides of the Atlantic leading up to October 13. That selling ceased and a vigorous rally commenced on the same day, continents apart, with neither market having access to any timely information from the other. Word of the N.Y. banking panic did not reach London until October 26, and was then reported in The Times the following day. The sudden, international cessation of distress selling that is a hallmark of 20th century panics also occurred in the crises of 1857, at a time when no timely communication existed. The international character of panics has been a stumbling block to those who subscribe to local, “fundamental” causes for these panics. Contrarily, a lunar-based model for panics would seem to require an international manifestation of the phenomenon. If the moon is affecting market participants, it should affect them the world over. All the panic examples cited here, from 1857 through 1997, have been international, yet the dearth of communication technology in 1857 provides a datum that cannot be explained as a serial reaction. The international character of panics is distinctively supportive of the lunar model.

Uses: Put simply, every market participant should have his calendar marked with the “dark days” of lunar 7-27 and 7-28. Even better, everyone should calculate the time of the eighth new moon and subtract 55 hours from that point. A time window of plus or minus

Psychology and Markets

Systematic Trading

twelve hours from that point is the lunar calendar model for an Autumn panic’s low point. There may not be another October stock panic for sixty years or longer. And the lunar model offers no clues as to which years will see a panic. Yet there can be no doubt, as the trillions of dollars lost during these panics make plain, market intelligence that can pinpoint when an unfolding panic will climax is invaluable. In 1997, as worldwide markets become unglued in October, the lunar calendar model provided by 1987 and 1929 pointed to late Monday October 27 as the ideal low point. The dramatic early Tuesday morning low of October 28 demonstrated the model’s effectiveness in real time. Calendars are complex mechanisms. Calendar research must recognise the importance of both lunar and solar calendars. The annual lunar model for panics points to the 27th and 28th days of the lunar month as the dark days, yet that is only true in the autumn season, the 6th or 7th lunar month. Past studies that purport to find no lunar relationship in markets have treated all lunar phases alike, lumping spring and fall together as well as summer and winter. Likewise specific seasonal analysis tends to ignore the concurrent lunar calendar. Those who dismiss that October may be a rough month for stocks cite that overall, it is not the worst month for stocks statistically, falling on average .5% since 1915. A proper approach to calendar research suggests that distinctions should be made among Octobers based on the lunar calendar. Here’s the lunar distinction. When there is no full moon between October 3 and 19 inclusive, the Dow has been up 1.5% in October since 1915. In those years with a full moon between those days, the Dow’s average change is a loss of 1.9%. Seasonal analysis should recognise the lunar distinctions and vice versa. The annual lunar calendar makes those distinctions. When autumn panics are viewed through its prism, the results are remarkable.

‘CaroIan. The Spiral Calendar. New Classics Library. 1992 ² Sobel. Panic.On Wall Street. Dutton. 1988. pp 318-320 3 ibid. p 106 4 Kindleberger. Manias Panics and Crashes. Basic Books. 1989. p 131 5 ibid. p131 6 ibid. p 143 1

Autumn Panics: A Calendar Phenomenon ©1998 Christopher Carolan Editor’s Note: This article was awarded the 1998 Charles H. Dow Prize by Dow Jones, Barron’s and the Market Technicians Association.

167


168

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Channels and Cycles Brian Millard

Article originally featured in Market Technician 34 (March 1999)

Introduction This talk will start with the relationship between the average length of a trend and the profit which can be expected from this trend, showing how closely the UK and US markets behave in this respect. A model of price movement will then be presented in which the major components are random point-to-point movement and partially random cyclic movement. The discussion will show how the random movement can be eliminated and the cyclic movement isolated either as a narrow band or as a broad band which will include all cycles greater than a specified wavelength. The latter leads quite naturally to the technique of channel analysis. Finally some idea will be given of the way in which the Sigma-p probability program carries out its calculations.

Profit Potential The starting point for this exercise is to recognise that trends must have a time-scale attached to them. Trends can be expressed with an exact time, such as a 5-day trend, or inexact, such as ‘short term’, ‘medium term’, etc. An up trend must start at a trough in the security price and end at a peak in the price, while the opposite is true for a down trend. Trends with a time scale within a narrow window around a specified value can be isolated by means of a centred moving average, i.e. an average which is plotted so as to lag by half of its span behind the latest data point. The span of the average is given a value which is the same as the nominal time scale of the trends which it is desired to isolate. The turning points in the centred average give the approximate locations of the starting and ending points of the trend. The exact starting and ending points are then found by scanning for a few points either side of the turn in the average for the trough or peak in the data. The actual trend lengths will be found to be close to the nominal value used for the centred average. By this method the average gains and losses for a nominal trend length can be determined. It is interesting that similar results are obtained for the constituent companies of the Dow Jones Index and for those of the FTSE100 Index, as shown in Table 1 for nominal 5, 25 and 51-day trends. Table 1: Average length of rising and falling trends in Dow and FTSE constituents Days falling

Average % fall

Days rising

Average % rise

Dow 5-day

5.2

-4.6

5.5

5.5

FTSE 5-day

5.4

-5.2

6.0

6.5

Dow 25-day

20.8

-10.3

26.7

14.6

FTSE 25-day

19.9

-10.3

27.1

15.8

Dow 51-day

35.7

-12.2

58.8

21.1

FTSE 51-day

36.9

-13.3

59.0

24.2

Once the average gain from a specified nominal trend-length has been determined, it is possible to determine the effect of re-investing the gains from one trade into the next. Obviously we can carry out more short term trades than long term trades over a period of say one year, so that the compounding effect will be larger, but on the other hand the gain per trend will be less. If we subtract say 5% from each trade to cover spreads, dealing costs and the fact that we will never enter and leave the trade at the exact starting and ending points, then we will arrive at the important conclusion that a plot of trend length versus this adjusted annualised gain will pass through a maximum, as shown in Figure 1. The optimum trend length taken as an average for the FTSE100 shares lies between 50 and 100 days. Thus remaining invested in one of these shares for longer than this before moving to the next trade is not recommended.

Figure 1:

A model of price movement This model assumes that price movement is made up of two elements: 1. point-to-point movement which is almost totally random and hence unpredictable 2. cyclical movement which is composed of a number of cycles of different wavelengths, amplitude and phase. Point-to-point movement ‘Point-to point’ movement refers to the change from one sampling point to the next, so that it can be hourly, daily, weekly, etc. The proportion of this point-to-point movement in the overall price movement will vary from one share to another, but between 5 and 10% seems to be the norm. At the 5% level, this random


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

movement can be almost totally eliminated by the application of a moving average with a span of at least 9 points, but a larger span is required at the 10% level. Cyclical movement It is possible to isolate a narrow band of cycles by using the difference between two centred moving averages. The span of one average should be about half that of the other, but both should have odd spans so that they can be individually centred. It is an advantage if the second average is weighted. By using this technique the nominal 201-day cycle in British Steel can be plotted as shown in Figure 2. The time axis is labelled in days rather than with specific dates for ease of interpretation of wavelengths. This Figure illustrates two important points about individual cycles. First the wavelength varies within fairly narrow limits as the time axis unfolds, and second, the amplitude varies within very wide limits over the same time period. It can be seen that the amplitude of British Steel is greatly reduced in the middle section of the plot. The correlation between the cycle peaks and troughs and the movement of the share price is obvious. The vertical scale on the cycle plot is in pence, representing the contribution made by that cycle to the overall price movement.

Psychology and Markets

Systematic Trading

have the same wavelength as its span. It will also remove most of the random point-to-point movement and will greatly reduce those cycles with wavelength less than the span. But it will have only a marginal effect on cycles with a wavelength greater than about one and a half times the span used. The plotting of a moving average with no lag, while useful for those indicators which rely on the crossing of one average by another or by the data itself, ignores a much more important aspect. Since the average represents the net result of all those cycles which have a wavelength of more than about 1.5 times the span being used and a trend can be considered to be one half of a cycle (either the rising or falling half), then the average is a good representation of the sum of all those trends which have a time scale of about 0.75 times the span of the average. However, this is true only if it is plotted as a centred average, as shown in Figure 3 where a 53-week average of the weekly closing prices of PowerGen is shown as the solid line through the data. Figure 3:

Figure 2:

Note that because centred averages have been used, a section equivalent to half of the span of the longer span average is missing. Thus the path taken over this last half span has to be estimated using the principle of symmetry, as has been done for the section from day 900 to day 1000. While the turning point in the cycle can be estimated quite closely because of the limited variation in wavelength, the amplitude, i.e. the importance of the cycle to the overall price movement, is subject to a very large error. The lowest error occurs when a cycle has been prominent for a period of only about one or two wavelengths; after this point the probability increases that the cycle will suddenly become greatly diminished. It can be seen from the failure of the price in the lower panel to make a substantial gain from around day 920 that the magnitude of the cycle over this period as shown in the upper panel is probably greatly overestimated, thus emphasising the unreliability of narrow band cycles in price prediction.

Moving averages as templates for channel analysis Of much more use to the investor is the isolation of the combined effect of all cycles greater than a specified wavelength. The random variation in many cycles will often cancel out, leaving a much more predictable result. The isolation of these cycles is achieved simply by the application of a moving average. While a moving average is an imperfect tool for this purpose, it is still extremely useful. A moving average will remove those cycles which

Once plotted in this way, another property becomes obvious - the excursions of the data on either side of the centred average are limited by an unseen boundary. These boundaries form a channel, and they can be drawn either by eye or by using the centred average as a template, moving copies of it vertically above and below the centre. Since channel analysis requires that the vertical depth remains constant, it is essential that the copies are at exactly the same distance above and below the average, i.e. the centred average remains at the exact centre of the channel. The boundaries are moved in or out until a small number of peaks and troughs in the data just touch or only marginally penetrate these boundaries which are shown by the dashed lines in Figure 3. The boundaries therefore represent areas of high probability that the price movement will reverse direction. This drawing of boundaries by using the centred average as a template is the easy part of the process, and is readily done by a computer program that allows a predetermined number of points to lie outside the boundaries. The difficult aspect is that, since a centred average has been used to draw the channel, it will terminate half a span back in time. The behaviour of the channel from that point in the past up to the present and on into the near future can only be estimated. The estimation starts by extrapolating the channel boundaries onwards at the same rate of curvature that they had just prior to the last calculated point in the average. This is done either by eye or by the computer using a curve fitting routine. The position of peaks and troughs in the data becomes critical in deciding whether the channel should be bent to prevent any of these penetrating the extrapolated boundaries to any great extent or to allow a significant peak or trough to approach a boundary. This latter

169


170

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

process can be seen in Figure 4 where the boundaries have been made to curve downwards while an extrapolation of their previous curvature would have them running more or less horizontally. Once this channel has been drawn and the investor is reasonably confident of its position, then an approach by the price to either boundary at the present time represents a high probability of a change in direction. Under such a circumstance a trading decision can be taken. As new data comes in, the analyst will constantly re-assess the position of the boundary in order to be ready when a decision point is reached.

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Figure 5:

Figure 4:

While space does not permit an extended discussion of the nesting of channels, it should be noted that channels can be drawn outside existing channels by drawing new boundaries that allow only a small number of penetrations by the inner channel. An outer channel will act as a control on the inner channel, forcing the inner to change direction in order to prevent violation of the boundary. This can help enormously when it comes to deciding if an inner channel might have changed direction.

The Sigma-p probability program The problem with the above method of extrapolating channels is that the predictability depends primarily upon the mathematics of curve fitting; there may be an absence of information from peaks and troughs and outer channels to provide evidence of any additional bending. Even when these features are present the evidence may be equivocal. The Sigma-p program overcomes this problem by taking the difference between each average point and one some distance in the past to obtain a ‘trend pressure’. The cyclical nature of this enables it to be predicted up to the present and into the future. The reverse mathematics then enables a predicted path to be obtained for the centred average, from which the channel can be obtained. In those cases where the average is dominated by only one or two cycles, not only the nearest turning point in the average can be predicted, but a turn well into the future. This is shown in Figure 5 for the FTSE100 Index. The program is predicting a fall in the FTSE100 53-week channel through 1999 until October, followed by a rise until April 2001. The chevrons indicate the width of the probability window which has to be applied to these predictions. It is also shown that the channel is probably falling as at 29th January 1999.

Brian Millard is the author of five books on technical analysis published in the UK with a sixth, Channels and Cycles - a tribute to J.M.Hurst, about to be published in the US. He owns Qudos Publications Ltd which carries out independent research and publishes investment software.


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

William Gann’s Law of Vibration Tony Plummer FSTA

Article originally featured in Market Technician 75 (November 2013)

William Gann William Delbert Gann (1878-1955) has a reputation both for being an extraordinary trader and for using uncommon knowledge about markets. There has always been clear evidence that Gann was ahead of most of his financial market contemporaries in terms of his use of specific entry and exit criteria. But it has also been very difficult to obtain details of his underlying knowledge. This is now changing. Accessible evidence is emerging that Gann built his trading systems on the threefold foundation of: (a) astrology; (b) sacred geometry; and (c) a ‘law’ of vibration. It seems that he was unable to be open about his use of such tools in the religiously hostile environment of the first half of the 20th century. Indeed, it is believed that he was at odds with his own family over the matter. Fortunately, the environment is now more favourable. Mr. Gann’s use of astrology, for example, has become more widely acknowledged in recent years, and many traders and analysts such as Bill Meridian in Austria and Olga Morales in Australia - are pointing to incontrovertible relationships between the planets and financial market activity. Mr. Gann’s use of sacred geometry embedded in his ideas of angular trend lines and ‘squaring price with time’ - have also achieved some degree of market penetration, although their origins, and therefore their full meaning, remain elusive. It is, however, Mr. Gann’s Law of Vibration that has been the most incomprehensible element in his triad of understandings. He mentioned the existence of such a law on a number of occasions in his various writings, but made his literary alter-ego, Robert Gordon, say in 1927: “The general public is not yet ready for it and probably would not understand or believe it if I explained it.”1 It was eventually assumed by many either that it was an exaggeration or that it was a reference to the structure of vibrations that could be found in sound waves and electricity oscillations. It can now be shown that Mr. Gann’s Law of Vibration was not just a propaganda tool or a figment of his imagination. It does exist. The problem, however, is that it is not a law that has been deduced from conventional scientific research; it is, instead, something that can more readily be found in esoteric literature.2,3 This leaves researchers facing a mountain of prejudice. But what if there is something in it? What if it actually works?

St. Matthew’s Gospel In Tunnel Thru The Air Mr. Gann pointed to Chapter 12, verses 38 to 40, of St. Matthew’s Gospel as being one of the sources of his information. He states quite explicitly that he understands what was meant by the phrase, “No sign shall be given, but the sign of the prophet Jonas”. He also maintains that there is “a secret meaning” in the phrase “The Son of man be three days and three nights in the heart of the earth”. And he argues that it “is the key to

the interpretation of the future”.4 We’ll probably never know how Mr. Gann came to recognise the importance of three short verses in St. Matthew’s Gospel, but the fact is that it really does contain something quite extraordinary. Specifically, it holds a number of geometric diagrams, each with great depth of meaning. This geometry is hidden from general view because it can only be deduced by the application to the original Greek text of the ancient science of numbers, known as ‘gematria’. This involves assigning a number to each letter of the Greek alphabet,5 and then calculating the numerical values of words and phrases in a particular piece of text. In this way, certain words can be identified with one another; words and phrases can be mathematically related to one another; and important ideas can be given a geometric structure.6 Once the number code is applied to St. Matthew’s text, it is possible to deduce at least three diagrams. The first is based on circles and squares, and is the configuration of the sacred linking of heaven and earth, known as the ‘vesica piscis’. The second is a potential revelation about the very nature of Time itself, which is the keystone of Mr. Gann’s sacred geometry. And the third is the pattern that is central to Mr. Gann’s Law of Vibration. I have shown how to unravel these aspects of St. Matthew’s text in my book, The Law of Vibration,7 and so will not go into any detail here. However, the important fact is that the text contains a pattern consisting of a three-wave advance into an early peak, followed by an elongated three-wave fall. This is, literally, the “sign of the prophet Jonas”, and it turns out to be very similar to the one that is central to Mr. Gann’s Law of Vibration. On this basis, we can Figure 1: William Gann’s pattern of vibration

171


172

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

conclude that knowledge of the Law of Vibration goes back (at least) 1,900 years.

Mr. Gann’s Law of Vibration Mr. Gann’s book, The Tunnel Thru The Air, is not particularly convincing as a novel. The basic storyline is one of love in a time of war, but the two lovers are rarely together. The romance is, instead, largely pursued by the artifice of an exchange of letters. In addition, the story relies heavily on the sentiments and advice contained in poems, excerpts from books, and passages from the Christian Bible. What Mr. Gann has done is use the letters, poems and quotations to control the number of pages that are assigned to each chapter. This aspect of Tunnel is, of course, eliminated - and the crucial information therefore completely obliterated - in more recent, and heavily revised, editions of the book.

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

‘somatic’ (or structural) changes designed to restore the system’s flexibility. One of the important features of the GGM pattern is that it shows this second stage as a contraction in the energy dynamics of the system. What happens is that, at some level, system energy has to be diverted towards making the necessary somatic adjustments. In the human brain, for example, the process involves the transfer of information from short-term memory into long-term memory. This takes energy and, while the transfer is occurring, the ability to do other tasks is reduced. Figure 2: The GGM pattern and its role in the processes of adjustment

There are 36 chapters in Tunnel,8 with the number of pages in each varying from a low of four to a high of 30. The relationship between the pages in a chapter and the chapter number itself can be plotted graphically. The result is as shown in Figure 1. The pattern is stunning. It is a six-wave cycle (three waves up and three waves down, which I have marked 1-2-3/A-BC), each of the larger waves consists of three phases; and the resulting configuration has a left-biased peak and a major mid-cycle low. Moreover, a closer inspection reveals that the whole diagram resonates with the harmonies of the Golden Ratio, 0.382:0.618.9 There is very little chance that the precision of the relationships is an accident and it stands in stark contrast to Mr. Gann’s unwillingness to reference the Golden Ratio in any of his writings. On its own, however, the diagram is almost meaningless. The same can also be said of St. Matthew’s construction. Fortunately, the pattern can be found in a third source, where the purpose and mechanics of the pattern are made explicit. The source is a book written by the self-development teacher George Gurdjieff (18731949), entitled All and Everything.10 The book itself is very difficult to read because the story is presented in the form of an allegory, and sometimes in a seemingly impenetrable language.11 But within the story, Mr. Gurdjieff describes a cyclical pattern by which energy travels through a living system. And then he hides the exact form of the pattern to which he was referring within the physical structure of the book (or at least within the structure of the first edition of the book):12 it is a six-wave cycle with a left-biased peak and a major mid-cycle low.13 Moreover - and most revealingly - the method that Mr. Gurdjieff used to conceal this pattern is exactly the same as the method used by Mr. Gann. It seems highly likely that the two men were in some way linked. The important practical point is that Mr. Gurdjieff’s description can be applied to Mr. Gann’s pattern. The conclusion can then only be that Mr. Gann’s pattern represents the vibrations - up to the level of an overarching life cycle - that are experienced by a living system. And the critical additional insight here is that Mr. Gann must therefore have thought that economic activity and financial market speculation somehow represented living systems. In other words, collective human behaviour is - quite literally - a natural phenomenon.

An archetypal pattern of shock and response The specific pattern offered by Mr. Gann, Mr. Gurdjieff and St. Matthew (hereafter, GGM) shows the initial impact of an information shock on a system (including the birth of a system itself), followed by the inner response of that system. The first stage of the response will involve the internalisation of stress and, therefore, the introduction of some degree of rigidity into the system. The second stage of the response will, however, involve

These two adaptive stages are shown as waves 1 and 2 in Figure 2, which is a simplified version of Figure 1. Once the adaptation has occurred, the system can advance with maximum energy. This is wave 3 in the diagram. There are, however, limitations on the system’s ability to make somatic/structural changes. Once this limit is reached, the system has to receive a new input of energy, and jump to a new higher structure, or it will contract. This is the point at which, with an appropriate stimulus, genuine evolution in the form of a system mutation can occur. Without the stimulus, however, the system reverses its polarity by dropping into what I have called an “energy gap”14, here represented by wave A. It is very important to recognise that it is the absence of extra energy in the approach to a potential energy gap that keeps the system locked into a recurring cycle. The energy gap reverses the system’s polarity from expansion to contraction or - more prosaically - from activity to rest; but the rest phase will nevertheless include a positive period (i.e., wave B). During the polarity reversal, the system undergoes changes that disrupt its inner integrity, releasing resources for other purposes, and/or actually destroying resources. This means that the ‘positive’ stage of the contraction period will not necessarily be dynamic because energy has to be diverted to incorporate some degree of inner reorganisation. These ideas can quite clearly be applied to the behaviour of financial markets. The initial recovery from an important low (wave 1) is usually a surprise. The bears will be squeezed, and the rally can be swift and aggressive, but will not generate self-fulfilling optimism. The implications of the bear squeeze will, however, subsequently be analysed and absorbed. The market will drift back (wave 2), and can even make new lows, but there will be an increase in the number of investors who buy into weakness and a decrease in the number of investors trying to sell. This reflects the introduction of more psychological flexibility into collective perceptions. Once this has happened, three things happen: ‘fundamentals’ start to catch up, buying becomes more


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

aggressive, and the market moves ahead (wave 3). This is the bull trend, where traders and investors have ‘learnt’ that the environment will support an advance. Consequently, any ‘bad’ news is regarded as being temporary and any weakness is treated as a buying opportunity. Eventually, of course, the market will become fully bought. It can only advance further if more investors can be induced by new information to part with their capital. If this does not happen, then Figure 3: The double GGM pattern and Elliott’s 5-3 configuration.

Psychology and Markets

Systematic Trading

the market will drop into an energy gap (wave A), and the polarity of the market will reverse. The market will not be properly based until it has completed another two waves after the energy gap (i.e., B and C). Then another cycle will start. If, on the other hand, new information arises just at the right moment, then the market has the potential to evolve/mutate to a new higher level structure. This necessarily means that the potential energy gap in the lower level structure is circumvented. See Figure 3. The gap is ‘jumped’. The movement after the new information is, in effect, wave 1 of the new structure (which can be denoted wave 1N). Then, after the necessary period of absorption (wave 2N), the market can advance again (wave 3N). Finally, the market will again become fully bought on the basis of the new information, and an energy gap (wave AN) will materialise. This will be the first wave of a conventional three-wave correction (AN-BN-CN). Importantly, the correction will therefore be related to the strength of the information shock at the top of the original wave 3 and, therefore, the dynamism of wave 1N within the new GGM pattern. The evolution of each stage of a market movement will depend very much on its position within the price-time hierarchy. That is, the actual shape of an advance and the depth and structure of the subsequent correction will depend on the presence or otherwise of higher level trends. I have covered this phenomenon in detail elsewhere,15 but the important conclusion is that the overlay of two levels of the GGM pattern produces a configuration that is similar to that observed by Ralph Nelson Elliott (1871-1948).16 What emerges is a 5-3 pattern, with the rising impulse wave consisting of waves 1, 2, and 3, which extends into 1N, and is followed by 2N and 3N. This rising wave, by definition, is going to be the most dynamic within the whole structure, and it will be followed by a three-wave correction consisting of waves AN, BN and CN. Because the correction is related to a ‘higher’ version of the underlying GGM

Figure 4: The 54-year Kondratyev cycle and Mr. Gann’s pattern

173


174

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

pattern, it is likely to be significant in the context of the whole pattern. On this analysis, a five-wave movement definitively indicates that some degree of structural evolution has occurred. And it also anticipates the emergence of some form of threewave correction. Even so, the whole pattern has to be seen as the outcome of an underlying three-wave structure.

The Law of Vibration in US wholesale prices: historical evidence The central thesis behind Mr. Gann’s pattern is that a shock of some sort will result in a predictable outcome. So the real question is: can Mr. Gann’s pattern be found in practice? The answer is, most definitely, yes. Once an analyst knows what he or she is looking for, it is possible to isolate Mr. Gann’s pattern in some - but by no means all - economic and financial market data. Figure 4 shows the average pattern of price behaviour over the four 54-year beats of the great Kondratyev cycle in US wholesale prices that occurred between 1785 and 2002.17 Price momentum is measured in terms of two-year rates of change, which helps to smooth away short-term fluctuations; each cycle beat is measured from trough to trough; and the (marginally) shorter cycle beats are mathematically stretched to match the longer ones. Each particular beat of the Kondratyev cycle will have its own characteristics, depending on contemporary influences. Nevertheless, the match between the average of four beats of the cycle and the expected pattern based on Mr. Gann’s prototype is truly remarkable. It accurately picks out the left-biased cycle peaks in 1813, 1864, 1917 and 1974 (all of which were related to major military conflicts); it identifies the disinflation that seems to precede such price explosions; and it isolates events such as the initial (1930-32) slump phase within the Great Depression (1930-39). Figure 5: The current Kondratyev price cycle

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

The surprise, however, is that this compelling correspondence does not occur over the whole 54 years of a Kondratyev cycle; it happens over just 48 years. Hence, for example, the first Kondratyev cycle in the diagram evolved between 1785 and 1840, but Mr. Gann’s pattern matches it only between 1792 and 1840. The important point, however, is that a similar divergence then occurs in the subsequent cycles. Moreover, the duration of this particular period is very consistent, at six or seven years. So we can hypothesise the existence of a recurring period, which has a constant duration and its own dynamic, that occurs between each manifestation of Mr. Gann’s pattern. One explanation for this phenomenon is that it actually takes time for collective psychology to recognise that an old cycle has finished. After all, the length of the Kondratyev cycle makes it unlikely that many individuals will have experienced the cycle as a complete whole. Indeed, it is far more likely that collective memory will be focused on either the most recent or the most traumatic part of the cycle that has involved a disruption. In some sense, therefore, the six to seven-year period that precedes the start of Mr. Gann’s pattern represents a time for cleansing and reassessment; it is the period where market operators get used to the idea that disinflation has ended.

The Law of Vibration in US wholesale prices: the current cycle It is therefore possible to amend Figure 4 to include the idea of a period, prior to the firing up of Mr. Gann’s pattern, that is in some way a preparation for change. In addition, we can superimpose the performance profile of the wholesale price index in the current Kondratyev cycle, which started in 2002. These amendments are shown in Figure 5.


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

In the early stages of the contemporary 54-year cycle (i.e., starting in 2002), the US authorities provided an aggressive stimulus to the domestic economy. At the time, the Fed misread the ending of a benign price disinflation as being the start of a malign deflation. This error - which seems to have been based on ignorance about the influence of the digital revolution and of globalisation - was compounded by official spending programmes related to the US’s military endeavours in Iraq and Afghanistan. The resulting stimulus created a bubble in wholesale prices and in financial markets - that is, the new Kondratyev cycle started with an unexpected burst of inflation - and it was the inevitable bursting of this bubble that was the essence of the 2008-09 financial panic. This seven-year period of government stimulus and financial collapse undoubtedly formed the preparation - or gestation period from which Mr. Gann’s pattern of vibration can be expected to emerge. What is important for the future evolution of the new cycle is the nature of the psychological changes that took place during this period. The simple fact is that it became almost universally accepted that low interest rates and plentiful liquidity were essential prerequisites for economic health. Ironically, the financial collapse of 2008-09 merely intensified this belief. This is why the ‘shock’ phase of Mr. Gann’s pattern (i.e., wave 1 of Figure 2), which had its main impact between 2009 and 2011, incorporated a collapse in interest rates and a resort to a form of money printing that is euphemistically known as ‘quantitative easing’. This means that (a) a monetary shock arrived on time, that (b) a 48-year pattern of vibration, which will be similar to the profile shown in Figures 4 and 5, has now started, and that (c) a major inflation episode is scheduled for 2020 to 2027. It is relevant that the ‘ideal’ pattern of vibration indicates that the precursor to the inflationary episode could be an outright fall in prices (see Figure 5). If this weakness was exacerbated by (for example) another banking crisis then the predictable outcome would be another bout of monetary easing.18 History suggests, however, that one of the main drivers of a major Kondratyev-type price inflation is military activity, and it hardly needs to be said that the global political environment is becoming increasingly unstable.

Figure 6: The transformed version of Mr. Gann’s pattern

Psychology and Markets

Systematic Trading

The Law of Vibration in terms of shocks and sub-cycles The beauty of Mr. Gann’s pattern of vibration in the context of wholesale prices is somewhat overshadowed by the fact that it is not always possible to find it in other data series. On the face of it, this raises questions about the pattern’s ubiquity. Deeper digging, however, provides the answer: the original pattern of vibration can be transformed into another, momentum-orientated, series that has a wider application. Figure 6 shows the relationship between Mr. Gann’s original pattern expressed in absolute terms (the upper part of the diagram) and that pattern translated into a one-period momentum series (the lower part of the diagram). Ignoring, for the moment, the ‘shock’ element of the upper diagram, the original pattern is then seen to contain three beats of a sub-cycle, where each beat: (a) has a duration equal to that of each of the other two; (b) has its own specific pattern; and (c) includes a significant intra-cycle contraction. There are two further points about the sub-cycle patterns that are worth registering: the first sub-cycle has a shape that is very similar to the shape of Mr. Gann’s original pattern; and the average of all three sub-cycles (not shown) is dominated by - and therefore very similar to - the shape of the first subcycle. The former aspect emphasises the non-random nature of the phenomenon that we are dealing with, and the latter aspect is of great use when trying to validate the presence of Mr. Gann’s subcycles within a data series.19 It follows from the two patterns, however, that the three subcycle beats of the lower pattern are triggered by whatever it is that constitutes the ‘shock’ in the upper pattern. This can help an analyst to determine the nature and location of the actual shock in a data series. Often the actual event (or series of events) is obvious; but it is very helpful to have it confirmed.

The Law of Vibration in US output: historical evidence Once the lower level momentum pattern is applied to data series such as the rate of change in US output, or the rate of change in the Dow, an extraordinary world of organised behaviour emerges. Figure 7 shows the average profile of US industrial production over four beats of the US’s dominant 35-year secular cycle (or ‘era’). The average relates to the 144 years between 1798 and 1942, and the cycles are estimated using one-year percentage rates of change. In order for the cycles to be directly comparable with each other, the shorter cycles are mathematically stretched to match the length of the longest cycles. As with the 54-year Kondratyev price cycle, each new pattern of oscillation does not start immediately after an old pattern has finished. There is a ‘lead in’, or a preparation phase, which seems to establish a psychological environment that is conducive to change. There is a variation of two to four years in the duration of this preparatory period but, on average, it lasts about one beat of the 40-month inventory cycle proposed by Joseph Kitchin.20 When it has finished, Mr. Gann’s pattern fires up. This latter consists of three phases (or sub-cycles), each with its own signature, and each containing a significant intra-cycle contraction. Each subcycle lasts for 10- to 11-years, which is similar to the periodicity of Clement Juglar’s capital expenditure cycle.21 Again, the correlation between the average pattern and Mr. Gann’s ‘ideal’ pattern is quite remarkable. Each of the three 10- to 11-years sub-cycles will inevitably have its own idiosyncrasies because of contemporary influences.22 Consequently, significant strength or weakness in any one cycle can distort the average. This is

175


176

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Figure 7: The 35-year output cycle and Mr. Gann’s transformed pattern

often particularly noticeable in the second (middle) cycle where major structural changes are evolving. In Figure 7, for example, the downward distortion was caused by the financial panic of 1893 - which occurred at the end of a Kondratyev price cycle (see Figure 4). Nevertheless, the average profile accurately identifies the economic slumps at the end of each sub-cycle and - very significantly - also picks out the major intra-cycle contractions. It is apparent, both from their shape and from their correlations with historical periods, that the three sub-cycles incorporate the three stages of evolution. These stages involve: differentiation from an old paradigm (or set of understandings); identification with a new paradigm; and then an integration of the best of the old paradigm with the innovations associated with the new. In Figure 7, therefore, the three cycles are therefore labelled separation (or transition), change (or innovation), and incorporation (or absorption), respectively.23 The separation cycle is volatile and forces a break with old beliefs and methodologies; the change cycle involves the creation of new ideas, new technologies and new attitudes; and the incorporation cycle normally starts well but ends in the throes of a great deal of fear. Figure 8 shows the average pattern of oscillations in US industrial production over the two 35-year cycles that evolved between 1946 and 2012. Mr Gann’s pattern isolates the major turning points and the intra-cycle contractions over the period - even though, as might be expected, some of the shorter-term fluctuations diverge from the ideal. The intra-cycle collapse in each of the incorporation cycles occur almost precisely where they should, in 1974 and 2009, and the subsequent recovery in momentum is followed by an end-cycle slowdown. This latter is the usual location of the phenomenon of a ‘second dip’ recession. In fact, the intra-cycle contraction during the incorporation cycle

registers as a ‘shock’ to the system. The unpalatable truth is that it exposes the excesses in the system and marks the start of the ending processes of the whole 35-year era. If the data from Figures 7 and 8 are taken together, something important is revealed. This is that end-of-era shocks to the US economic and financial system do not occur randomly in terms of time. The shocks of 1865 (American Civil War), 1904 (Rich Man’s Panic), 1937 (Depression within the Great Depression), 1974 (the Great Inflation), and 2009 (Global Financial Panic) happened at virtually the same stage of each 35-year socio-economic era. This conclusion then raises profound questions about the role of government. The end-of-era shock is initially viewed as being only temporary because very few people believe that the system has reached its limits. However, as the subsequent ‘preparation’ stage becomes more embedded, reality is unlikely to fulfil optimistic expectations. Instead, disappointments increasingly force collective beliefs to absorb the fact that changes are inevitable. Figure 8 includes the profile of output during this phase, as experienced in 1942-46 and 1978-80. The shorter period is mathematically stretched to match the length of the longer phase, and then the two series are averaged. The period starts well, but then deteriorates. The end of the preparation phase is accordingly reflected in a recession. There is some evidence that, in late 2012, US output reached the end of the socio-economic era that began in 1978. This still has to be confirmed, particularly since slow growth in early 2013 suggests that an early 2013 end date might be more appropriate. In either case, however, the evolving economic recovery needs to be viewed as the prelude to a very much more difficult period.

The Law of Vibration in financial markets Figure 9 shows the behavioural pattern of the Dow Jones


Indicators and Momentum

Elliott Wave and Fibonacci

Figure 8: The post-war experience

Figure 9: The 35-year cycle in the DJIA

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

177


178

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Industrial Average, measured in terms of six-month rates of change, over three secular cycles, starting in 1907. Each cycle consists of an initial ‘preparation’ phase of between two and six years, followed by a set of three 10- to 11-year cycles. On average, the total length of a secular cycle is 35 years. One of them, however, is longer than the others (i.e., 1941-1978), so in order to calculate the average for the chart, the shorter ones have been mathematically stretched to match it. The average is calculated after this has been done. The result is as clear as in the case of US output: there is a very close match between Mr. Gann’s pattern and the average DJIA pattern. In particular, the ‘ideal’ pattern picks up the bear phases at the end of each 10- to 11-year cycle. It therefore picks up major bears, such as those of 1920-21 and 1930-32. Encouragingly, it also picks up the intracycle contractions. So it identifies the 1987 equity crash (in the first subcycle), and it identifies the equity traumas of 1937-38, 1973-74, and 2008-09 (in the third sub-cycle). The latter are particularly important. They confirm that (a) there are end-of-era equity collapses, that (b) these collapses occur with a periodicity of 35 years, and that (c) they are not therefore unforeseeable and random shocks. A closer analysis will also reveal the malign influence of government policy in each of these events. The chart also includes the average historical behavioural pattern for equities during the ‘preparation’ phase. Obviously, the exact pattern that starts each era will depend on contemporary circumstances, but the average pattern indicates a positive start that is then followed by a deep retrenchment. In the current era, US equities are only a few months into this preparation phase, but it seems wise to anticipate that the period will end with a cyclical bear. At the moment, such a bear looks likely to conclude in 2015/16. This would coincide with the bottom of the final economic recession before the next triad of Mr. Gann’s pattern of vibration starts.

Helmsman Economics’ models and Mr. Gann’s pattern of vibration Throughout the decade of the Noughties, I used model patterns for prices and for output that were almost identical to those that were hidden by Mr. Gann. These models accurately identified all the major historical turning points in prices and output. They also anticipated the dual recession and Kondratyev price low of 2000/02, and the economic disaster of 2008/09. The point is that it was quite possible to deduce behavioural patterns that were similar to Mr. Gann’s from the available data. All that was required was a shift in assumptions away from rational behaviour by independent individuals towards the concept of collective learning within the context of evolutionary change. As Henry Mills showed in 1967, the process of learning expresses itself in a very specific three-phase energy pattern.24 So, treating markets and economies as dynamic learning systems encouraged a search for three-wave patterns. The result was a basic model that assumed the presence of six-phase cycles (three waves up and three waves down), leftbiased learning patterns (with a peak about one-third of the way through the process), and the clustering of cycles in groups of three (reflecting the stages of evolution). 25 In constructing this model, however, I did not recognise the specific relationship between a six-phase, left-biased, learning cycle and a set of three evolution-orientated oscillations. The ‘secret’ was that the two constructs belonged to different hierarchical levels: an overarching learning cycle produced a lowerlevel triad of cycles. Moreover, the former made the initiating shock explicit, while the latter left that shock implicit. Despite the limits of my own models, the associated research and analysis prepared the ground for new information. This was why it was so easy for me to ‘recognise’ Mr. Gann’s vibrational

William Delbert Gann pattern when it was unearthed. In addition, my research had already generated a set of ideas an concepts that could be used to interpret Mr. Gann’s pattern. Some of these understandings have been applied in this article.

Conclusion A great deal of research still needs to be done on the Law of Vibration and on its potential applications to financial markets and to economic forecasting. Nevertheless, the early results are both stunning and compelling. The first insight is that collective financial and economic activities have a powerful cyclical component. The second insight is that so-called ‘shocks’ that emerge suddenly and unexpectedly in financial markets and/or the economy actually have a non-random periodicity. And the third insight is that, every 35 years or so, one of these shocks emerges to end a whole socioeconomic era. The global financial crisis of 2008-09 accordingly marked the beginning of the end phase for the era that began in 1978. The evidence suggests that (give or take a few months) the era actually ended in late 2012. We can look back on the whole era and see that it encompassed an economic revolution, based on digital technology. Although not widely acknowledged, however, the digital revolution is now reaching a degree of maturity and the primary product markets are becoming satiated. Innovation has transmuted from genuinely new products (personal computers, and mobile phones) to improvements and artificial enhancements (e.g., social networking, new look hardware, and enhanced software). Crudely put, an energy gap is appearing in product innovation. Using Mr. Gann’s Law of Vibration as a guide, the unavoidable inference is that western economies are unlikely in the next 10 years to experience either the same degree of economic dynamism, or the same flow of tax revenues into government coffers, that were experienced in the 10 years prior to the 2008-09 crisis. Moreover, the solution to a supply-side slowdown is not a demand-side stimulus. Indeed, the inflation-orientated obsession with monetary and fiscal stimuli currently being pursued by politicians is likely only to make the inevitable adjustment significantly worse.


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

Notes 1

William D. Gann, The Tunnel Thru The Air (Financial Guardian Publishing, New York, 1927).

2

Nevertheless, my own research on economic and financial cycles, much of it published in the Market Technician, clearly identified the specific patterns and relationships that can now be seen as being intrinsic to Mr. Gann’s Law of Vibration. I had already proposed that cyclical oscillations arise in groups of three; that each cycle beat has its own natural pattern; and that each such beat has a serious intra-cycle contraction. See: The phenomenon of cycles, Market Technician, No. 37, February 2000; Pattern and periodicity in financial cycles, Market Technician, No. 45, October 2002; and The Kondratyev Wave: Heretical thoughts and practical understandings, Market Technician, No. 59, July 2007. The main difference between my findings and Mr. Gann’s ‘law’ relates to the existence of an intra-cycle period that prepares collective psychology for the evolutionary changes that are due to come.

3

For some, this is at least consistent with Mr. Gann’s status as a 32nd Degree Freemason of the Scottish Rite Order. See Constance Brown, in David Keller (Ed.), Breakthroughs in Technical Analysis (Bloomberg Publishing, New York, 2007). If Freemasonry was, indeed, the source of Mr. Gann’s knowledge, then significant questions are raised about why such information is being kept hidden.

4

William Gann, op cit.

5

There are some accompanying rules that introduce an important degree of flexibility into the relationships between words and numbers.

6

The word ‘gematria’ is derived from the Greek word for ‘geometry’.

7

Tony Plummer, The Law of Vibration: The Revelation of William D.Gann (Harriman House, Petersfield, Hants, UK, 2013)

8

This is almost certainly an allusion to the 360 degrees of a circle.

9

Further details can be found in Tony Plummer, op cit.

10

George I. Gurdjieff, All And Everything: An Objectively Impartial Criticism of the Life of Man (Routledge & Kegan Paul, London, 1950).

11

Mr. Gurdjieff’s ideas/teachings are presented in a more accessible form in a parallel book. Pyotr D. Ouspensky, In Search of the Miraculous: of an Unknown Teaching (Harcourt, Brace & World, New York, 1949)

12

As with The Tunnel Thru The Air, more recent revisions to All and Everything have distorted the hidden pattern.

13

Tony Plummer, op. cit.

14

Tony Plummer, Forecasting Financial Markets (Kogan Page, London, 1989-2010).

15

Ibid.

16

The lives of William Gann, George Gurdjieff and Ralph Elliott obviously ran in parallel.

17

The Kondratyev cycle is here defined only in terms of prices. Most analysts conflate prices and output in trying to track the Kondratyev cycle. This presumes a functional relationship between the two variables that runs from output to prices. In practice, however, there is a relationship that runs from prices to output. This was particularly true during periods of innovationdriven disinflation. The price-output conflation also results in a very long and very variable ‘wave’ that is of very little use in terms of practical prediction. Concentrating just on prices yields a very accurate cycle of 54 years, plus or minus 2 years.

18

It is impossible now to be specific about the likely cause of an outright deflation prior to the inflation event. Another debt crisis within the Eurozone is a candidate, as is a military event. But there can be little doubt about the response from the authorities: monetary expansion will be significant.

19

The idea is that three cycles in a data series should, when averaged together, look like the average of the theoretical construct.

20

Joseph Kitchin, Cycles and trends in economic factors, Review of Economic Statistics, 1923.

21

Clement Juglar, Des crises commercials en leur retour périodique en France, en Angleterre, et aux États-Unis (Paris, Libraire Gillaumin et Cie, 1862).

22

It also needs to be pointed out that the accuracy of some of the earlier data may be suspect.

23

These labels accord with some of the concepts that were introduced into evolutionary psychology in the 1980s and 1990s - particularly by Ken Wilber. See, for example, Ken Wilber, The Essential Ken Wilber (Shambhala Productions, Boston (Ms.), USA, 1997).

24

Henry Mills, Teaching and Training (Macmillan, London, 1967).

25

See note 2. See also Tony Plummer, op. cit.

Tony Plummer was formerly a Director of Hambros Bank Ltd. and responsible for their trading positions in the London Gilt-Edged market.

179


180

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Fibonacci - Lucas Numbers, Moon Sun Cycles and Financial Timing David McMinn

Article originally featured in Market Technician 75 (October 2013)

Abstract Fibonacci numbers and Phi ratios are used in various techniques to predict market outcomes and have long been accepted in technical analysis. The reason why these factors arise in financial patterns has never been clarified, thereby creating a credibility problem. Fibonacci - Lucas numbers and Phi ratios also appeared in the timing of historic May and autumn panics, as well as Moon Sun cycles. Thus, it was hypothesized that there were interrelationships between these phenomena - trading activity, lunisolar tidal harmonics and additive numbers. If proven, it would provide a sound scientific basis for using Phi and Fibonacci numbers in market forecasting, thereby supporting a fundamental tenet of technical analysis.

Introduction Fibonacci numbers and Phi (1.618) have long been used in technical analysis to help predict financial outcomes. Various market forecasting techniques are based on such factors, for example the Elliott Wave (Pretcher, 1980) and the Spiral Calendar (Carolan, 1992). Why additive numbers and Phi ratios arise in trading activity has never been explained, despite their common usage. This paper examines Fibonacci - Lucas numbers and inverse Phi ratios in relation to Moon Sun cycles, as well as the timing of historic financial panics. Numerous correlates can be achieved to support a strong Moon Sun effect in market activity (Dichev & Troy, 2001; Yuan et al, 2006, McMinn, 2006, 2010), while Fibonacci - Lucas numbers and Phi ratios are evident in financial patterns. Based on these two strands of thinking, additive numbers and Phi were hypothesized also to show up in Moon Sun cycles. This was a reasonable speculation, assuming the various factors were valid and interrelated. On assessment, Fibonacci - Lucas numbers and Phi could be firmly established in patterns of lunisolar cycles. Lunisolar tidal effects are believed to influence human physiological cycles, which in turn determined the prevailing mass mood and thus market outcomes. Periods of optimism lead to rising markets, while periods of pessimism result in declining indices and depressed markets. The crisis occurs when there is a sudden shift in sentiment from greed to fear. The collective mood is viewed as fluctuating through cycles of optimism - crisis - fear, in harmony with lunisolar cycles. A connection between Moon Sun effects, physiological cycles and market outcomes can be supported by various studies. Hormone levels of animals and humans have been shown to fluctuate over the lunar month (Endres & Schaad, 2002; Zimecki, 2006), while studies have linked hormone levels to market trading success (Chen & Ozdenoren, 2005; Coates & Hebert, 2008; Coates et al, 2009). Moon Sun data was timed at 12 noon in the financial centre where the crisis or panic occurred (daylight saving ignored). Data on the Dow Jones Industrial Average (DJIA) index was based on the daily closing values throughout the text. The abbreviation A° was used to represent the angular degree between the Moon and Sun (lunar phase), while the ecliptical degree is denoted by E°. This was to

prevent confusion between two different concepts. Robert van Gent provides an excellent coverage of the various eclipse cycles discussed in this paper and is recommended as background reading.

Fibonacci - Lucas numbers Readers with a background in technical analysis will already be familiar with Fibonacci numbers. These are an additive series in which each number is the sum of the previous two, beginning 1, 1, 2, 3, 5, 8, 13, 21, 34, 55, 89... Lucas numbers are another additive series beginning 2, 1, 3, 4, 7, 11, 18, 29, 47, 76, 123... Both series are interrelated. Lucas numbers may be derived by adding or subtracting two Fibonacci numbers as shown in Table 1. The Fibonacci series commencing 1, 1... and the Lucas series commencing 2, 1... are the simplest additive number series, both of which show up in Moon Sun cycles and financial patterns. Table 1: Relationships between Fibonacci and Lucas Numbers Adding Fib Numbers

Subtracting Numbers

Lucas Number

1+1

1 - (-1)

2

0+1

2-1

1

1+2

3-0

3

1+3

5-1

4

2+5

8-1

7

3+8

13 - 2

11

5 + 13

21 - 3

18

8 + 21

34 - 5

29

13 - 34

55 - 8

47

The golden ratio Phi (1.618) is denoted by the symbol Φ and is produced between two successive numbers in any additive series. For example, the ratio of any two successive Fibonacci numbers


Indicators and Momentum

Elliott Wave and Fibonacci

is alternately greater or less than 1.618 as follows - 1/1 gives: 1/1 - 1.000; 2/1 2.000; 3/2 - 1.500; 5/3 - 1.667; 8/5 - 1.600; 13/8 - 1.625; 21/13 - 1.615; 34/21 - 1.619; 55/34 - 1.618 and so forth. For the larger Fibonacci numbers, the ratio increasingly approaches 1.618. The important Fibonacci ratios are 0.382, 0.618, 1.382, 1.618, 2.382, 2.618 and so forth. In technical analysis, these ratios are used to help forecast future turning points in market patterns. Inverse Phi ratios are given in Table 2 below. Table 2: Inverse Phi Ratios Inverse Sqrt Phi Ratios*

Inverse Phi Ratios

0.618 1/Phi

0.486 1/sqrt Phi3 0.382 1/sqrt Phi4

0.382 1/Phi2

0.300 1/sqrt Phi5 0.236 1/sqrt Phi6

0.236 1/Phi3

0.146 1/sqrt Phi 0.115 1/sqrt Phi

0.146 1/Phi

4

9

0.090 1/sqrt Phi10

1847 Oct 23

+10

174 A°

1857 Oct 14

+50

324 A°

1907 Oct 22

1937 Oct 18

+30

196 A°

+50

164 A°

1987 Oct 19 324 A°

+10

1997 Oct 27 320 A°

Table 4: October panics and inverse Phi ratios 1857 Oct 14

1907 Oct 22

+50

324 A°

+30

196 A°

1937 Oct 18

+50

164 A°

1987 Oct 19 324 A°

1857 + 50 1907 + 30 1937 1907 + 30 1937 + 50 1987 50 divided by 80 = 0.625 (1/Phi). 30 divided by 80 = 0.375 (1/Phi2). 3, 5 & 8 are all Fibonacci numbers. 1857 + 50 1907 + 80 1987 1857 + 80 1937 + 50 1987 80 divided by 130 = 0.615 (1/Phi). 50 divided by 130 = 0.385 (1/Phi2). 5, 8 & 13 are all Fibonacci numbers. 1907 Oct 22

+60

174 A°

+30

196 A°

1937 Oct 18

+60

164 A°

1997 Oct 27 320 A°

1847 + 60 1907 + 30 1937 1907 + 30 1937 + 60 1987 These are in the ratio of 2:1:2 1, 2, 3 & 5 are all Fibonacci numbers. 1929 Oct 28

+58

1987 Oct 19

+10

324 A°

1997 Oct 27 320 A°

10 divided by 68 = 0.147 (1/Phi4). 58 divided by 68 = 0.853 (1 - 1/Phi4). 10, 58 & 68 divided by 2 equals 5 & 34 (Fibonacci numbers) and 29 (Lucas number). The intervals are based on the equation (29 + 5 = 34) x 2.

0.090 1/Phi5 1907 Oct 22

0.071 1/sqrt Phi11 0.056 1/sqrt Phi12

Systematic Trading

British and US October Panics

313 A°

0.186 1/sqrt Phi7 8

Psychology and Markets

Table 3: October panics and Fibonacci numbers

1847 Oct 23

0.786 1/sqrt Phi 0.618 1/sqrt Phi2

Gann Analysis, Cycles and Forecasting

0.056 1/Phi6

* Square root Phi equals 1.272

October panics Major historic October panics occurred in a notable 10 - 50 - 30 - 50 - 10 year series (see Table 3). The six events happened in the two weeks to October 27, with lunar phase around the full Moon (between 160 and 200 A°) or before the new Moon (between 320 and 325 A°). The intervals between the 1857, 1907, 1937 and 1987 panics occurred in Fibonacci numbers (3, 5, 8 and 13 multiplied by 10) and thus yielded inverse Phi ratios (see Table 4). NB: Years in bold contained major financial crises as listed by Kindelberger (Appendix B, 1996). How the 1847 and 1997 panics integrate into the overall pattern is puzzling. These panics give intervals in a 5:1 ratio (both Fibonacci numbers) that adds up to 6 (neither a Fibonacci or Lucas number). Even so, the series 10 - 50 - 30 - 50 - 10 years appeared too neat to be coincidental. October panics in 1929,

+22

196 A°

1929 Oct 28

+8

313 A°

1937 Oct 18 164 A°

8 & 22 divided by 2 gave 4 & 11 (both Lucas numbers) respectively.

1987 and 1997 can also be adjusted to give inverse Phi ratios and double Fibonacci - Lucas numbers (see Table 4). Double Lucas numbers (11 x 2 & 4 x 2) also appeared between the 1907, 1929 and 1937 October panics.

September panics Kindleberger (Appendix B, 1996) only listed three September crises/panics in recent centuries. 1763, September. Amsterdam panic. 1873, September 19. US Black Friday. Jay Cook & Co failed. 1931, September 20. Britain announced that it would go off the gold standard. The intervals between these three events produced the ratio 55:29 - a Fibonacci and a Lucas number respectively (see Appendix 1). Seven major annual one day (AOD) falls (=> -4.50%) for the DJIA have taken place

in September since 1896. These events yield Fibonacci - Lucas numbers and inverse Phi ratios, as shown in Appendix 1. The notable anomaly was the 31 year interval between the AOD falls in 1955 and 1986. NB: The AOD fall is taken as the biggest % DJIA one day fall during the year commencing March 1. It represents the biggest one day shift in negative trader sentiment during a given solar year.

May panics A listing of US and Western European May panics were sourced from Kitchin (1933) and Kindleberger (Appendix B, 1996). Historically, May panics have clustered between the 9th and the 21st of the month. If placed in chronological order, the intervals between May panics were nearly always in Lucas numbers (see Table 5). The exception was the 1884-1920 interval of 36 years, which was a double Lucas number (18 x 2). Unlike

181


182

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

October panics, May panics showed no lunar phase emphasis. Lucas intervals between key historic May panics yielded many inverse Phi ratios (see Table 6).

Table 5: May panics and Lucas numbers Historical US and European May Panics 1819 May

+18

1837 +29 May 10

1866 May 11

1873 May 9

+7

+11

1884 1920 +36 +11 Oct 19 May 19(a)

1931 May 11

(a) The biggest % one day DJIA fall in 1920 was used as the maximum panic intensity, as no panic date was given by Kindleberger (1996). NB: The US 1893 panic was not regarded as a May event, because Black Wednesday occurred on July 26.

Table 6: Inverse Phi ratios derived from May panics 1819 May US panic

1837 May 10 US panic

+ 18

1866 May 11 British panic

+ 29

+ 29

1866 May 11 British panic

1884 May 13 US panic

+ 18

18 divided by 47 = 0.383 (1/Phi2). 29 divided by 47 = 0.617 (1/Phi). 18, 29 & 47 (all Lucas numbers). 1837 May 10 US panic

+ 36

1873 May 9 Austrian panic

+ 58

1931 May 11 Austrian panic

36 divided by 94 = 0.383 (1/Phi2). 58 divided by 94 = 0.617 (1/Phi). 36, 58 & 94 divided by 2 = 18, 29 & 47 respectively (all Lucas numbers). 1873 May 9 Austrian panic

1884 May 13 US panic

+ 11

+ 47

1931 May 11 Austrian panic

11 divided by 58 = 0.190 (1/sqrt Phi7). 47 divided by 94 = 0.810 (1 - 1/sqrt Phi7). 11 & 47 (Lucas numbers), 58 (double Lucas number)

Table 7: Lucas numbers and eclipse cycles n

Phin

The German Black Friday (May 13, 1927) was not listed by Kindleberger (Appendix B, 1996), but this event gave Lucas numbers 4 and 7 when inserted between the 1920 and 1931 May crises. The anomaly was the US Black Thursday (May 9, 1901), which failed to produce Lucas numbers when inserted between the 1884 and 1920 panics. Additionally, the interval between US stock market panics of May 14/21, 1940 and May 28, 1962 was 22 years - a double Lucas number.

Lunisolar eclipse cycles

18 divided by 47 = 0.383 (1/Phi2). 29 divided by 47 = 0.617 (1/Phi). 18, 29 & 47 (all Lucas numbers). 1837 May 10 US panic

Moving Averages and Trends

Lucas No

Eclipse Cycle

Lunar Months

Solar Years

0

1.000

02

Hexon

035

2.830

1

1.618

01

Half Lunar Yr

006

0.485

2

2.618

03

Hepton

041

3.315

3

4.236

04

Octon

047

3.800

4

6.854

07

Tzolkinex

088

7.115

5

11.089

11

Tritos

135

10.915

6

17.942

18

Saros

223

18.030

7

29.030

29

Inex

358

28.945

8

46.971

47

47 YC

581

46.975

9

75.999

76

Short Calippic (a)

939

75.920

10

122.966

123

Half 246 YC (b)

1520

122.895

(a) One Calippic equals 76.0 solar years (940 lunar months) or four Metonic cycles of 19.0 solar years each. The Short Calippic is equal to the Calippic minus one lunar month (939 lunar months). (b) Robert van Gent listed a 246 year eclipse cycle (unnamed) of 3040 lunar months, which divided by two gave the 123 year cycle of 1520 lunar months. Abbreviation: YC - Year cycle. Source of Eclipse Cycle Data: Robert van Gent Source: McMinn, 2006.

Fibonacci - Lucas numbers can be directly linked to Moon Sun cycles. The additive series commencing 35 and 6 lunar months gave Lucas numbers (in terms of solar years) for the following eclipse cycles - Tzolkinex (7 years), Tritos (11 years), Saros (18 years), Inex (29 years), 47 year cycle, Short Calippic (76 years) and the 123 year cycle (see Table 7). For cycles less than 7 years and over 123 years, the link with Lucas numbers peters out, as solar years align less precisely at integral numbers. According to van den Bergh (1955), the interval (in terms of lunar months) between two solar or lunar eclipses can be derived from the formula: T = a.Inex + b.Saros where T is the interval between successive eclipses in numbers of lunar months. a and b are integral numbers (zero, negative or positive). The Saros equals 18 solar years (223 lunar months), while the Index equals 29 solar years (358 lunar months) with 18 and 29 both being Lucas numbers. Based on van den Bergh’s formula, the series commencing 35 and 6 lunar months is composed in multiples of the Inex and Saros in patterns of Fibonacci numbers (see Table 8). The Saros or Inex number may be positive or negative for eclipse cycles below 135 lunar months. For eclipse cycles of 223 lunar months or more, the Saros and Inex numbers are always positive. Another additive series may be produced commencing 62, 37 lunar months, which was equivalent to a series beginning 5, 3 solar years (two Fibonacci numbers) (see Table 9). In this series, the same lunar phase repeats at the same time of year over hundreds of years and included the important 8 year Octaeteris and 19 year Metonic cycles. Applying van den


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Table 8: Fibonacci numbers and eclipse cycles Lunar Months

Eclipse Cycle

Inex

Saros

a.Inex + b.Saros

35

Hexon

-8

13

-8 I + 13 S

6

Half Lunar Year

5

-8

5I-8S

41

Hepton

-3

5

-3 I + 5 S

47

Octon

2

-3

2I-3S

88

Tzolkinex

-1

2

-I + 2 S

135

Tritos

1

-1

I-S

223

Saros

0

1

S

358

Inex

1

0

I

581

47 YC Unnamed

1

1

I+S

939

Short Calippic

2

1

2I+S

1520

Half 246 YC Unnamed

3

2

3I+2S

2459

199 YC Unknown (a)

5

3

5I+3S

3979

322 YC Unknown (a)

8

5

8I+5S

6438

521 YC Unnamed

13

8

13 I + 8 S

(a) Eclipse cycle not listed by Robert van Gent.. Abbreviations: S - One Saros cycle of 223 lunar months. I - One Inex cycle of 358 lunar months. YC - Year Cycle. Source of Eclipse Cycle Data: Robert van Gent Source: McMinn, 2006.

Table 9: The additive series commencing 5, 3 solar years Named Cycle

Octaeteris

Metonic

Lunar Months

Solar Years

5,3 Year Series

van den Bergh’s Formula a.Inex + b.Saros

62

5.012

5

-97 I + 156 S

37

2.992

3

68 I - 109 S

99

8.004

8

-29 I + 47 S

136

10.996

11

39 I - 62 S

235

19.000

19

10 I - 15 S

371

29.996

30

49 I - 77 S

606

48.996

49

59 I - 92 S

977

78.993

79

108 I - 169 S

1583

127.989

128

167 I - 261 S

2560

206.981

207

275 I - 430 S

4143

334.970

335

442 I - 691 S

Bergh’s formula to this series did not yield Fibonacci - Lucas numbers. How the series beginning 5, 3, 8... years integrates with the additive series in Table 7 remains unknown.

Discussion and conclusions Eclipse cycles are important because they give repeating angles between the Moon, the Sun and other lunisolar factors (McMinn, 2006). The changing angles between the Moon and the Sun in the heavens play a key role in terrestrial tidal harmonics, which are believed to influence the mass mood. Eclipse cycles, as discussed in this analysis, are separate from eclipse phenomena. Eclipse events

may appear spectacular to humans on Earth, but the author could not establish any direct links between market trends and the timing of solar/lunar eclipses. Moon Sun astronomical planes seem highly relevant - the plane of the Earth’s orbit around the Sun (the ecliptic), the plane of the Moon’s orbit around the Earth, the plane of the Earth’s equator extended out onto the heavens (celestial equator) and so forth. The points where these planes intersect are called nodes, imaginary points that seem to yield maximum significance. The random walk - efficient market hypothesis was the prevailing paradigm in academic finance during the latter

Systematic Trading

decades of the 20th century. According to this tenet, financial markets were believed to function both efficiently and randomly. This completely contradicted technical analysis, which viewed market activity as being repetitive and mathematically structured. Moon Sun correlates, additive numbers and Phi cannot arise in a random market and thus the findings of this paper completely contradict the concept of market randomness. A more realistic view is to consider markets to be mathematically structured, moving in tune with Moon Sun cycles. • Lucas numbers 7, 11, 18, 29, 47, 76, 123... in terms of solar years (see Table 7). • Fibonacci numbers using van den Berg’s formula (see Table 8). Another additive series beginning with 5 and 3 solar years (both Fibonacci numbers) produced the same lunar phase at the same time of year over hundreds of years (see Table 9). Fibonacci - Lucas numbers also show up in trends of historic May and autumn panics. Thus, close interrelationships are speculated to arise between financial trends, lunisolar cycles, additive numbers and Phi ratios. This would give strong support for the use of Phi and Fibonacci - Lucas numbers in market forecasting. There may be emerging a simple theory based on Moon Sun tidal harmonics, which would reduce the complexity of market cycles to a few basic principles. This would greatly boost our understanding of financial timing and offer the potential to make accurate market forecasts years in advance. Only time will tell. The connection between free markets, lunisolar cycles and Fibonacci - Lucas numbers can only be proposed as a hypothesis. How Moon Sun tidal harmonics actually functioned in relation to financial trends remained enigmatic. Much more research is necessary before definitive conclusions can be drawn. Lunisolar cycles potentially offer a causal explanation why financial patterns exhibit additive numbers and Phi ratios, thereby supporting a fundamental principle in technical analysis. Hopefully, this paper offers an impetus for other researchers to explore this topic more fully.

References Carolan, Christopher. 1992. The Spiral Calendar. New Classics Library. 159p. Dichev, Ilia & James, Troy. 2001. Lunar Cycle Effects In Stock Returns. Working paper. August. http://papers.ssrn.com/ sol3/papers.cfm?abstract_id=281665

183


184

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Chen, Y, Katusccak, P & Ozdenoren, E. 2005. Why Can’t A Woman Bid More Like A Man? Working paper. September 3. Coates, J M & Herbert, J. 2008. Endogenous Steroids and Financial Risk On A London Trading Floor. Proceedings of the National Academy of Sciences. Apr 22; 105 (16): 6167-72. Coates, J M., Gurnell, M & Rustichini, A. 2009. Second-to-Fourth Digit Ratio Predicts Success Among High frequency Financial Traders. Proceedings of the National Academy of Sciences. Jan 13; 106: 347-348. Endres, Klaus-Peter & Schaad, Wolfgang. 2002. Moon Rhythms in Nature. How Lunar Cycles Affect Living Organisms. Floris Books. Kindleberger, Charles P. 1996. Manias, Panics and Crashes. John Wiley & Sons. 263p. Kitchin, J M. 1933. Trade Cycles Chart. Published by The Times Annual Financial & Commercial Review. 1920, 1924, 1930. Revised chart to 1933 presented in Gold. A reprint of The Special Number of The Times. June 20. Times Publishing Co Ltd. McMinn, David, 2006, Market Timing by The Moon & The Sun. Twin Palms Publishing. 163p. McMinn, David. 2010. Market Timing Moon Sun Research 2004 to 2009. Privately Published. 185p. Prechter, Robert R. 1980. The Major Works of R N Elliott. New Classics Library. Prechter, Robert R & Frost, Alfred J. 1978. Elliott Wave Principle: Key to Stock Market Profits. New Classics Library. van den Bergh, George. 1955. Periodicity and Variation of Solar (and Lunar) Eclipses, 2 vols. H D Tjeenk Willink & Zoon NV, Haarlem. van Gent, Robert. A Catalogue of Eclipse Cycles. http://www.phys.uu.nl~vgent/ calendar/eclipsecycles.htm Yuan, Kathy, Zheng, Lu & Zhu, Qiaoqiao. 2006. Are Investors Moonstruck? Lunar Phase & Stock Returns. Journal of Empirical Finance, Vol 13, Issue 1, p 123. January. Zimecki, M. 2006. The Lunar Cycle: Effects On Human and Animal Behviour and Physiology. Postpy higieny i medycyny doswiadczalnej (online). 01/02/2006; 60:1-7. http://www.ncbi.nim.nih.gov/pubmed/1 64007788

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Appendix 1: Inverse Phi ratios derived from September AOD falls DJIA AOD September falls => -4.50% since 1896 AOD Fall

% Fall

Sun E°

Moon E°

Phase A°

Sep 24, 1931

-7.07

181

338

157

Sep 03, 1946

-5.56

161

252

091

Sep 26, 1955

-6.54

183

301

118

Sep 11, 1986

-4.61

169

264

095

Aug 31, 1998

-6.63

158

265

107

Sep 11, 2001

na (a)

169

090

281

Kindleberger’s September Panics 1763 Sep

+ 110

1873 Sep 19

+ 58

1931 Sep 20

These intervals gave a 55:29 ratio, which comprised a Fibonacci and a Lucas number respectively. Adding these numbers gave 84, which cannot be linked to Fibonacci - Lucas numbers.

Inverse Phi Ratios Derived From September DJIA AOD Falls => -4.50% 1931 Sep 24 157 A°

+ 15

1946 Sep 3 091 A°

+ 40

1986 Sep 11 095 A°

Intervals gave an 3:8 ratio comprising two Fibonacci numbers. Adding 3 and 8 gave 11, a Lucas number, while 55 was a Fibonacci number. 1946 Sep 3 091 A°

+ 40

1986 Sep 11 095 A°

+ 15

2001 Sep 11 281 A°

Intervals gave an 8:3 ratio comprising two Fibonacci numbers. Adding 3 and 8 gave 11, a Lucas number, while 55 was a Fibonacci number. 1931 Sep 24 157 A°

+ 15

1946 Sep 3 091 A°

+ 55

2001 Sep 11 281 A°

Intervals gave an 3:11 ratio comprising two Lucas numbers. Adding 3 and 11 gave 14 a double Lucas number. 1931 Sep 24 157 A°

+ 55

1986 Sep 11 095 A°

+ 15

2001 Sep 11 281 A°

Intervals gave an 11:3 ratio comprising two Lucas numbers. Adding 3 and 11 gave 14 a double Lucas number. 1986 Sep 11 095 A°

+ 12

1998 Aug 31 107 A°

+3

2001 Sep 11 281 A°

Intervals gave a ratio of 4:1 or two Lucas numbers. Adding 4 and 1 equals 5, a Fibonacci number. 1931 Sep 24 157 A°

+ 15

1946 Sep 03 091 A°

+9

1955 Sep 26 118 A°

Intervals gave a ratio of 5:3, which added together gave 8. All Fibonacci numbers.

Includes falls DJIA AOD falls => -4.50% since 1896. (a) The September 11 terrorist attack was taken as the day of maximum panic intensity for 2001. Abbreviations: E° is the degree on the ecliptical circle. A° is the angular degree between the Moon and Sun (lunar phase). AOD - The annual one day fall is taken as the biggest % one day fall in the year commencing March 1.


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

185


186

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

CHAPTER NINE

Psychology and Markets Articles in this chapter 188 Will You Win or Lose in the Financial Markets? Robert Newgrosh

190 Neuro Linguistic Programming and the Investor Antony Lehman

192 The Herd Instinct? Mark Jones

194 Cognitive Behavioural Psychology for Investment Analysis Herbert Labbie

Moving Averages and Trends


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

CHAPTER NINE INTRODUCTION

Michael Smyrk FSTA

Introduction Luck has played a big part in my business life. I happened into my first job (in the City, working for an Import/Export merchant that had a seat on the London Metal Exchange) because office hours started at 9:30am (compared to others that started at 9:00), so I had early experience of the principles of Futures trading. Later I was moved to the Sugar floor, running a book for the company. At that time, I (like everyone around me) was unaware of charting, but I then advanced to the London branch of a US Commission House, where my immediate boss (in New York) was the author of a recent book on Futures. He it was who in 1965 introduced me to charting, which at that time mainly involved trends and patterns. However, he also worked with Dick Donchian, who used Moving Averages to trade systematically - computers were a key tool. By this time, I had joined ACTA (the Association of Chart & Technical Analysts, which ultimately became the STA), and had started realising what a fascinating and worthwhile field Technical Analysis was. Psychology and the Markets covers a vast and varied area of knowledge, so it is not surprising that Market Technician has included many articles on the subject over the years. Choosing which to re-publish here has therefore proved a tricky task, but hopefully readers will find stimulation as well as knowledge to be gained. There are of course two aspects of the subject that could be covered - the psychology of the trader/investor in the face of the markets, and the psychology of the markets as influenced by the trader/investor. It has been written that “successful trading is 40% risk control and 60% self-control, and the risk control portion is one half money management and one half market analysis. So, market analysis is only around 20% of successful trading” [Van Tharp]. On that basis, we shall concentrate here on those articles that should help the trader/investor to approach the markets in the right frame of mind, aware of the risks and rewards involved. We start off with (almost) the oldest and (certainly) the most basic, published in the mid-1990s. Robert Newgrosh writes of ‘trying to beat’ the Financial Markets, “Winning” or “Losing”. He thus joins the ranks of those commentators who see the market as an opponent a not uncommon attitude, probably typical of the majority of participants. His short piece is therefore likely to be particularly useful. Then we go still further back in time, to MT 13 [March 1992]. Antony Lehmann takes a more purely psychological approach, with heavy reliance on Neuro Linguistic Programming as used by Van Tharp. Some readers may have an inbuilt distaste for this approach, with its focus on ‘modelling’, but the article itself should provide some threads that can usefully be followed.

Psychology and Markets

Systematic Trading

The recurring development of patterns is to some extent explained by ‘group’, or ‘herd’ behaviour, a topic that was covered in July 2009 [MT 65] by Mark Jones. This freelance journalist was intrigued by a study in which a dozen participants studied photos of scared, neutral and happy faces. The volunteers recognised fearful expressions a half-second quicker than other images. Bernard Baruch wrote, in a Foreword to the 1932 edition of “Popular Delusions & the Madness of Crowds” (published in 1841/1852 by the Rev Charles MacKay, a collection “of the most remarkable instances of ... moral epidemics ...”) that “All economic movements, by their very nature, are motivated by crowd psychology”; this statement is taken forward and amplified by Herb Labbie CMT [MT 37]. Herb, whose background was in Biophysics and the Law as well as Money Management, and whose newsletter was one of the few to advise taking protective action before the Crash of ’87, came over from Pittsburgh to speak to the STA in 1999; this article illustrates well his wide knowledge and enthusiasms.

187


188

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Will You Win or Lose in the Financial Markets? Robert Newgrosh

Article originally featured in Market Technician 20 (July 1994)

Methodology or Psychology?

How the Motivators Cause Problems

Millions of investors and traders have tried to beat the financial markets.

Let us now outline the key potential problem associated with each of the motivational factors.

Why do very few succeed? Why do the winners win and the rest underperform? Why is it that irrespective of investment methodology, wealth, intelligence or background, most investors end up with the same experience? Why does it take years for the winners to become successful? Why do most investors not perform as well as they think they will?

a) The desire to make money

All these questions share the same answer. Remarkably, most investors never discover it. Quite simply... Performance is not ultimately determined by your timing or selection method, but by your own psychology, emotions, impulses, disciplines and attitude. Investors are not beaten by the markets, they beat themselves through mismanaged thinking, mistaken beliefs and by falling into “mindtraps”.

What Motivates People to Invest in the markets? In order to gain an appreciation of the reasons why people make psychological and attitudinal mistakes in the markets, it is necessary to have an understanding of the motivation or the need to invest in the first place. The reason for this is that the driving forces behind an investor are potential traps when it comes to making decisions either at the time of entry or during the lifetime of an investment. Below is a list of possible motivating factors influencing a decision to participate in the financial markets. Most people will admit to the first, some of the other reasons may be present sub-consciously but would not normally be disclosed.

The first, and probably most obvious, of two primary motivators. No-one would enter into the markets if they expected to lose. This primary motivator causes major problems because the desire to make money from the markets means that investors can have tremendous difficulty in accepting and taking small losses due to the conflict with the motivator. An inability to cut losses while they are small is the main reason why the majority of investors eventually fail. b) The desire to make easy money A different motivator to a) due to the attitude towards effort. It is a basic error made by newcomers to the markets. From the outside it all looks pretty easy. Surely there are only two choices, thinks the novice: up or down. Get this right more often than not and we can give up the day job. This line of thinking severely underestimates the difficulty of the game. Easy money is a very rare commodity and gains from the markets are paid for one way or another, usually through years of hard work and initial losses. Trying to get something for nothing frequently produces nothing for something. Once the notion of easy money is dispelled, the investor switches from this motivator to a). c) To prove we are right/smarter than others The second primary motivator. The need to be right is the major reason why people can’t take small losses when they have the chance. They see a loss, no matter how small, as a blow to their ego and an admission that they got it wrong, thereby conflicting with the motivator. This, coupled with the desire to make money as explained above, can be a lethal combination and results in investors refusing to take small losses on adverse positions and eventually being forced to close with a much larger loss. d) The excitement, entertainment and action of the markets

a) The desire to make money. b) The desire to make easy money. c) To prove we are right/smarter than others. d) The excitement, entertainment and action of the markets. e) To gain a sense of independence and control. f) The thrill of winning. g) As a means of gambling. h) Everyone else is.

Anyone who enters the markets for these reasons is almost certain to fail. For the markets to provide this kind of stimulus would also mean a high level of stress due to constant participation, decision making and activity. For those who seek excitement, entertainment and action from their money, a trip to Disney World would be less stressful and ultimately a lot cheaper! e) To gain a sense of independence and control Most people, unless they run their own business, have their finances determined and controlled by someone else. In most cases this is their employer. For many people this financial straitjacket is a source of frustration. To be able to control one’s own


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

income is a very attractive idea and the markets would appear to offer this opportunity. However, money cannot be extracted from the market just like that. One has to accept a degree of risk and potential for loss, thereby negating a key benefit of a regulated income. f) The thrill of winning This is not the same as the desire to make money. The thrill of winning is a psychological benefit rather than a material benefit. It is interesting that the excitement of the occasional win is enough to keep most people playing a losing game, whether it be in the financial markets or elsewhere. g) As a means of gambling Some people use the markets as just another vehicle for gambling. Simply taking into account the costs of dealing and the tendency for markets to behave irrationally and illogically at times, there are better bets available elsewhere for the true gambler. Of course, markets are a game of skill where varying degrees of control are exerciseable, but the gambler does not see this. h) Everyone else is Some of the greatest crashes in history have occurred because people invested for this reason. You can not make money doing what everyone else is doing - not in the long run anyway. Indeed, if everyone else is invested, you are usually better off disinvesting. For some people, this list will mean that they are highly unlikely to succeed in the markets. They are simply unsuitable psychologically for one reason or another. For the rest, survival and success depends on the ability to understand your own motivators and weaknesses. In addition, success requires disciplines and attitudes which are often contrary to one’s impulses. The reason for this is that normal human thinking and behaviour clearly does not work in the markets as evidenced by the fact that the majority of people fail. Success, therefore, requires a different psychological approach. In particular, one has to be aware of the range of emotional forces that destabilise an investor once he or she begins to interact with the markets - hope and fear, joy and despair, anxiety and relief. This emotional rollercoaster, combined with elements of greed, bravado and the need to protect one’s ego, serve up a concoction of confusion and conflict for the psychologically uninitiated and are his or her eventual undoing. All investors need to be acutely aware that psychological factors mean that observing markets is one thing and interacting with them is something else. It is vital to appreciate that...

“THE GAME CHANGES WHEN YOU TRY TO PLAY IT!”

Robert Newgrosh is Managing Director of New Skills Ltd, a specialist training company. This article is an extract from his course ‘Will You Win or Lose?’

189


190

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Neuro Linguistic Programming and the Investor Antony Lehmann

Article originally featured in Market Technician 13 (March 1992)

When asked what is required in order to make money in the markets consistently, most professionals will mention three different areas which need to be addressed. The first is to have a tested set of rules or a system which will provide buy and sell signals in which the investor has faith. The second is to have an understanding of money management rules. The third is to have the emotional make-up which ensures that the first two can be put into practice. An understanding of the third area explains why two investors can be following the same buy and sell rules and have the same money management techniques yet one can make a fortune and the other lose everything. We all know that two technical analysts can look at the same chart and come up with two different recommendations. As a result we say that technical analysis is an Art not a Science. I remember well my frustration with this in my early trading. My background in physics had resulted in a need for a more analytical approach and consequently I spent much time reading anything that might help my understanding. That was how I came across the work of Dr. Van K. Tharp PhD who has used a branch of Cognitive Psychology called Neuro Linguistic Programming (NLP) to build a model of a Successful Trader. I sent off for his course and after 6 months of intense study felt it would be worthwhile to attend the 3 Day seminar in the Vienna Plaza last October. Dr Tharp was accompanied by Adrienne Toghraie, a Master Practitioner of NLP and Tom Basso, President of Trendstat who attended as a guest speaker in his capacity as a successful CTA. What has psychology got to do with trading is a common question put to Dr Tharp to which he answers everything. When he began his life’s work building a model of a “Supertrader”, the first step was to find what the hundreds of super traders he interviewed had in common. It soon became obvious that since they were all talking about their specific methodology that that had nothing to do with it. More important was their self discipline and their inner reactions to the markets. With this article I want to outline some of the points which made the biggest impression on me and convey why I believe more attention should be paid by investors to this subject. The content of the seminar was very much of the experimental type with lots of exercises resulting in some very profound changes to the personalities of some of us. Before going into more detail, some background about NLP would probably be useful. NLP itself is a model which is proving very accurate as evidenced by the success its practitioners are having in bringing about changes in peoples’ behaviour. It is extremely difficult doing justice to the subject in a short summary and I would recommend interested readers to the various books on the subject.

Basics We all experience our surroundings through our five senses: sight, sound, feelings, smell and taste. However a complication is that our conscious mind can only cope with a limited amount of input. Research has estimated this is no more than 7 (+/-2) items of information.* As the average number we are bombarded with is very much higher we need to filter the information to make it manageable. After filtration these experiences are re-presented to our brains and it is at this point we introduce Deletions, Distortions and Generalisations. As we all delete, distort and generalise differently this, in itself, is enough to explain why two investors using the same system could end up with different results. There are six filters which, in order of conscious awareness, are Meta Programmes, Values, Beliefs, Attitudes, Memories and Decisions. It may seem odd, at first, to describe the above as filters yet, after some thought, the logic becomes clear. A person with a certain set of values will not spend much time considering points of view which conflict with these hence the filtration. Likewise someone with a certain belief will look for evidence to support the belief and disregard (filter out) evidence conflicting with the belief. These filters, therefore, determine what information is retained as we make an internal representation of an event. It is our internal representaion which causes us to be in a certain state be it happy, sad, motivated, depressed, bullish or bearish. And of course it is our state that decides our behaviour be it action or inaction, buying or selling.

Meta Programmes These are the mind’s sorting patterns and are so deeply ingrained in us that often, when we are confronted with them, we respond as if they were obvious and wonder why “reality” is being questioned. They are best described by analogy with the Sort command in database software. Imagine a database file containing 100 records of name, age, house number, street, area, town and postcode. If a


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

person had these permanently sorted by ascending order of age in his mind he would see things and consequently act very differently to someone whose list was sorted in order of postcode, say. When a person is helped to shift his “sort” to include one of the other possibilities the difference can be profound.

with decisions is that they were often made unconsciously or when we were very young and then forgotten.

An important point to make here is there are no correct “sort patterns” or Meta Programmes but only useful and non-useful ones depending on what you are trying to accomplish.

Dr Tharp considers himself a coach and his work centres around helping the investor to change his or her filters to ones that may be more appropriate for success. One of the services he provides is The Investment Psychology Inventory which is a confidential questionnaire enabling the build up of a profile comparing the participant to other investors in the critical skills needed for success. The composite score is ranked against the scores of others who have taken the test.

Values Values are how we decide whether our actions are good or bad or right or wrong. Values are how we decide how we feel about our actions and they therefore provide the motivating force behind our actions. Values are those ideas in which we are willing to invest time, energy and resources to either achieve or avoid. They consequently function as filters. As well as all of us having different values, we will also have them arranged in a different hierarchy of importance. This is significant as it demonstrates one of the reasons why two investors can be following the same system and money management rules yet have different results. If one has a different set of values to the other he may abandon the system too soon or make simple mistakes in the execution. This latter trait is often the result of a conflict of values within the person. For example suppose he or she values the freedom that money will bring but also values security. Depending on the precise definition these can conflict and therefore result in subconscious self-sabotage of execution of the system. The resolution of these conflicts is one of the most elegant examples of the strength of NLP.

Beliefs Beliefs are convictions or acceptances that certain things are true or real. They are also generalisations about the state of the world. In this capacity therefore they act like on/off switches for our ability to do anything in the world. An investor with the appropriate beliefs for success will far outperform one with inappropriate ones. One of the more important elements of modelling excellence in investment (or in any field for that matter) is to discover the beliefs of the top investors and enable others to adopt the same.

Attitudes Attitudes are collections of values and beliefs around a certain topic. We are usually well aware of our attitudes and consequently they are difficult to change. It is as if we believe our attitudes reflect personalities which are carved in stone and we often tell people “Well that’s just the way I feel about that”.

Investment Psychology Inventory

One of the more complex meta programmes describes whether a person has a “Towards” or “Away From” strategy when it comes to motivation. Is a person attracted towards money or repelled away from poverty for example? This has many implications not least of which is the effectiveness of the stick or the carrot as a means of producing results. One important consequence for the investor is if he/she has both meta programmes, in other words is attracted towards the idea of making lots of money and at the same time wants to get away from poverty. In order that both wants may be satisfied the investor needs to create lots of little disasters along the way to success. If high drawdowns in equity is a problem for you then work in this area can achieve dramatic improvements. Let us suppose an investor has a hierarchy of values such that freedom is the most important priority and money, for the sake of argument, is fourth in order of importance. It is possible that a conflict between these two values exists. A part of the investor may want all the rewards that money can bring but another part may prefer the freedom of being undisciplined or not having to work in a systematic way. It is unlikely that the investor will achieve real success until this conflict is resolved. In order to whet your appetite, consider the investor who buys at the top of the bull market and holds on all the way to the bottom of the ensuing bear market in the hope of an upturn. Finally; unable to stand the despair any longer, he decides to sell only to watch the market recover. He resolves to buy a system and become more disciplined only to find that the system he chooses gives him a string of losses just at the point when he starts to use it. Finally he can stand it no more and decides to move to a better system just as the one he gives up begins a run of winning trades. What might the meta programmes, values, belief systems and decisions of this investor be?

Memories Memories deeply affect a person’s perception and personality and therefore act as a powerful filter. Some psychologists believe that as we get older our reactions in the present are reactions to collections of past memories. For example, some investors were responding to their memories of the 1973-75 bear market all the way through the bull market of the eighties and consequently never made as much money as was potentially possible.

Decisions The sixth filter is decisions we made in the past. These can affect our entire life. Decisions may create beliefs, values, attitudes and even life themes and as such affect our perceptions. The problem

Antony Lehmann has been trading futures since June 1989 and is in the process of setting up Tiger Futures Ltd.

191


192

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

The Herd Instinct? Mark Jones

Article originally featured in Market Technician 65 (July 2009)

If the City of London doubts that evidence for The Herd Instinct exists, it just needs to look out of the window. When the Millennium Bridge was opened in June 2000, it was heralded as an icon of futuristic urban design - then swiftly ridiculed when the structure began to wobble dangerously as the first pedestrians crossed over. The newspapers leapt with glee on the Wobbly Bridge story, and you can see why. Trendy designers build a footbridge not designed for hundreds of human feet: it’s a godsend of a story. But what was that design flaw? This. The bridge was designed to cope with the weight of thousands of individual, random footsteps. As they walked across, however, people began, unconsciously, to synchronise with one another’s walking patterns. It was the effects of this impromptu marching army that the designers failed to foresee. Let’s extrapolate: as individuals we may think we are going our own way, using our personal expertise and experience to make rational decisions. Unconsciously, we are marching in time with an army that has no goal, no mission - and which doesn’t even know it is an army. The phrase ‘herd instinct’ crops up time and time again to describe the panic selling that had brought so many markets and institutions low. We are talking about a group of trained, analytical, often well-educated financial services professionals. Yet they are accused of swinging with ‘irrational exuberance’ (in Alan

Greenspan’s words) to blind fear in the past difficult months. Behaviour is the buzzword. Economic and market analysis has moved from the spreadsheets, figuratively speaking, and onto the psychiatrist’s couch. The lexicon of the credit crunch is expanding correspondingly. The language of psychology and neuroscience is being recruited in the desperate search for answers and explanations. Many suspects have been fingered: sub-prime lending, the debt market, short-selling. The latest is not an economic phenomenon but a chemical. It’s called cortisol. The hormone - a form of steroid, in fact - is produced by the body during times of stress, anxiety and uncertainty. Testosterone promotes risk taking and - that key word in any market confidence. It is no cliché that a bull market produces gallons of the stuff on the trading floor, but an empirical fact. Cortisol acts differently. John Coates used to run a trading desk at Deutsche Bank. He is now a senior research fellow in neuroscience at Cambridge University. He describes the effects: “In small amounts, cortisol quickens reactions and sharpens the mind. But if elevated cortisol exposure lingers and becomes chronic, its effects are quite different. No longer sharpening the wits, it begins to hinder them and inhibits the urge to take risks”.


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Cortisol, he writes, is especially stimulated by conditions of novelty, uncertainty and uncontrollability - “conditions that describe a trader’s life during a crisis”. In a study, Coates took saliva samples from a group of traders. Cortisol levels rose by as much as 500 per cent during times of massive market volatility. The figure is bound to vary from individual to individual. But Coates believes cortisol may be a big factor in herd behaviour. Put simply, the stuff is highly contagious. “One animal’s surge in testosterone or cortisol can be spread throughout a herd and trigger aggression or panic”, he notes. You can see where he is going with the theory. “If traders become more risk-seeking during booms and more riskaverse during slumps, then the markets may be inherently unstable and traders may fail to seize on profitable opportunities”. Experiments have shown that anxiety is the easiest emotion to transmit between human animals. Think of the atmosphere on a plane when it goes through a period of extreme turbulence or a lift stalls between floors; how anxiety spreads as one stranger catches another’s eye. In a study of mass anxiety in the market for The Times, the writer John Naish cites a Vanderbilt University study in which a dozen participants studied photos of scared, neutral and happy faces. The volunteers recognised fearful expressions a half-second quicker than other images. That was the conscious reaction. Unconsciously, our response to anxiety is even more powerful. If you asked traders to say what drives their decision-making process, they are not likely to say “it’s my amygdala playing up again”. It may well be the truth, however. The amygdala is the brain’s anxiety centre and spots signs of fear in less than twohundredths of a second. As Naish puts it, “the sense of fear grows because we swap ever larger doses of worry with each other”. Adherents of the Efficient Market Hypothesis (EMH) - whose number have probably dwindled of late - believe that the market is structurally accurate; and that stocks will always find their proper level despite the efforts of individuals to find values through technical analysis, trend-spotting or exploiting observable investor psychology. There are different critiques of that theory. Warren Buffet sees himself as a living riposte: “I’d be a bum on the street with a tin cup if the markets were always efficient” he once said. A more academic rebuttal comes in the emerging discipline of socionomics. For three decades, former Merrill Lynch analyst Robert Prechter and psychologist Wayne Parke have been gathering data around social behaviour in times of uncertainty. In a timely paper published in June 2007, Prechter and psychologist Wayne Parker contend that “we behave one way when we are certain about economic value to ourselves; and very differently when we are unsure of others’ future financial valuations. In the first case, they approach value rationally and consciously, while in the second case they unconsciously herd and then rationalise their decisions.” The subsequent crisis in inter-bank lending and the vacillations of the LIBOR rate might be brought forward in evidence: has there been a time in history where banking professionals have been “more unsure of others’ future financial valuations”?

Psychology and Markets

Systematic Trading

economics writer of the 20th century, J. K. Galbraith, understood that extreme market phenomena cannot be explained by mathematical models and the close study of economic cycles. You have to look to the human factor. After the crash of 1929, many put the blame on the availability of cheap credit and historically low interest rates. Galbraith said that was “obvious nonsense”: rates were high in the late twenties compared to later years; and borrowing was just as easy in non-booms. Galbraith talked instead of a contagious euphoria that had driven the market. “Speculation on a large scale requires a pervasive sense of confidence and optimism and conviction that ordinary people were meant to be rich. People must also have faith in the good intentions and even in the benevolence of others, for it is by the agency of others that they will get rich... Such a feeling of trust is essential for a boom”. He goes on to describe a backlash which believers in the cortisol theory will recognise: “When people are cautious, questioning, misanthropic, suspicious, or mean, they are immune to speculative enthusiasms.” Robert Schiller, the foremost exponent of the behavioural school, echoed those sentiments in his book Irrational Exuberance written, presciently, three years before the latest crash: “The stockmarket has not come down to historical levels,” he wrote. “People still place too much confidence in the markets and have too strong a belief that paying attention to the gyrations in their investments will someday make them rich, and so they do not make conservative preparations for possible bad outcomes.” So can human beings override their complex hormonal and neurological impulses at times of immense stress - and avoid the contagion that spreads through the trading herd? John Coates writes that there is “some preliminary but tantalising research showing that certain people can adapt themselves into physiologically toughened individuals by displaying a muted cortisol response to stress.” He speculates, not entirely seriously, you suspect, that Olympic-style training regimes could be designed for traders. A former equities trading head had a simpler solution for traders who succumbed to a state the neuroscientists call “learned hopelessness”, where people freeze and lose faith in their own ability to spot an opportunity. “I sent them home,” he says. There is a branch of cognitive psychology devoted to the study of resilience: the ability of individuals to withstand stress and catastrophe. Stress can, in fact, be an ally as well as a danger. Studies suggests that those with complex, busy and indeed stressful lives are better able to cope with an unexpected shock such as the death of a relative. Physical health and stamina are assets too - so perhaps John Coates’s Olympic training camp is the right idea. Experience is a vital factor. Resilient individuals tend not to dwell on bad experiences: they learn and accept them. If nothing else, the credit crunch has shone a light on some unexpected quirks of the human mind.

Rational economics say that if something is on offer at historically low prices, bargain-hunters will rush in. Anyone who witnessed the dash to Selfridges and Marks and Spencer when they announced pre-Christmas 20%-off sales saw exactly that phenomenon at work. But Prechter contends that “the Law of Supply and Demand gets flipped on its head in the financial markets. As a stock’s price rises, demand for it tends to increase. When prices are cheap, few want to buy. This is exactly the opposite of what happens in the butcher’s shop or the shoe store.” Behavioural finance is a thriving discipline with a growing body of case study. But before the term was even coined the finest

Mark Jones is a freelance journalist. This article first appeared in Roar, the magazine for clients of Liontrust Asset Management PLC, Spring 2009.

193


194

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Cognitive Behaviorial Psychology for Investment Analysis and Management Herbert G. Labbie, CMT

Article originally featured in Market Technician 37 (February 2000)

This article is a summary of a talk given to the Society on 15th September 1999 Bernard Baruch was one of the greatest Wall Street speculators of the early part of this century. In 1931 he wrote: “All economic movements, by their very nature, are motivated by crowd psychology. Graphs and business ratios are, of course, indispensable in our grouping efforts to find dependable rules to guide us in our present world of alarms. Yet I never see a brilliant economic thesis expounding, as though they were geometrical theorems, the mathematics of price movements, that I do not recall Schiller’s dictum; ‘Anyone taken as an individual, is tolerably sensible and reasonable - as a member of a crowd, he at once becomes a blockhead,’ or Napoleon’s maxim about military masses: ‘In war, the moral is to the physical as 3 to 1.’ Without due recognition of crowdthinking (which often seems crowdmadness) our theories of economics leave much to be desired. It is a force wholly impalpable - perhaps little amenable to analysis and less to guidance - and yet, knowledge of it is necessary to right judgments on passing events.” Humphrey Neil is regarded as the fountainhead of sentiment and contrary opinion. He wrote in his classic The Art of Contrary Opinion: “This writer believes the great lack in economic and political studies is the failure to analyze human nature… How people think and how they act are too often overlooked when we are trying to project a future pattern. Human behavior is fully as important as, if not more important than, statistical behavior. I believe that the human figure, in other words, is fully as important as the mathematical figure in our calculations concerning economic trends and socio-political trends.”

Contrary opinion has not been confined to investments. Psychology and contrary opinion have been mentioned in conjunction in great literature for many centuries. In Europe in the seventeenth century, John Milton wrote in Paradise Lost: “The mind is its own place, and in itself Can make a Heaven of Hell, a Hell of Heaven” John Milton, Paradise Lost, Bk. 1 l. 254-5 (1667) Goethe wrote: “I find more and more that it is well to be on the side of the minority, since it is always the more intelligent.” Great American writers have made similar observations.

Richard Schabacker was a master technician whose work was acknowledged by Edwards and Magee as the basis for their Technical Analysis of Stock Trends. Schabacker wrote in his book, Stock Market Profits (1934): “...there is most definitely a broad field for the application of simple, elementary principles of psychology in security analysis and operation. In many respects, it appears to be a field quite as important as others on which much research has been done and on which many volumes have been written. And yet, this psychological aspect has been almost totally neglected in modern market analysis... Technical science is at bottom merely the attempt to analyze, through consideration of stock charts and trading data, the changes that are taking place in the activities of human beings...”

Mark Twain wrote in Huckleberry Finn, ”Hain’t we got all the tools in town on our side? and ain’t that a big enough majority in any town?” Henry David Thoreau in his much quoted comment said: “If a man does not keep pace with his companions, perhaps it is because he hears a different drummer. Let him step to the music which he hears, however measure or far away.” Writers in economics and business have commented on contrary opinion a long time ago. The renowned economist William Stanley Jevons (1835 -1882) noted: “As a general rule, it is foolish to do what others are doing, because there are almost sure to be too many people doing the same thing.”


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Charles Dow, of the classic Dow Theory, wrote a hundred years ago in an editorial in The Wall Street Journal, which he founded: “The elder Rothschilds are said to have acted on the principle that it is well to buy a property of known value when others wanted to sell and to sell when others wanted to buy. There is a great deal of sound wisdom in this. The public, as a whole, buys at the wrong time and sells at the wrong time.” The importance of psychology has been emphasized by a broad spectrum of writers. John Maynard Keynes wrote in his General Theory of Employment, Interest and Money (1936) that most investors’ decisions “can only be taken as a result of animal spirits-of a spontaneous urge to action rather than inaction, and not as the outcome of a weighted average of benefits multiplied by quantitative probabilities.” Keynes preceded the behaviorists by his picturesque musical chairs and beauty contest descriptions: “For it is, so to speak, a game of Snap, of Old Maid, of Musical Chairs - a pastime in which he is victor who says Snap neither too soon nor too late, who passes the Old Maid to his neighbour before the game is over, who secures a chair for himself when the music stops... Or, to change the metaphor slightly, professional investment may be likened to those newspaper competitions in which the competitors have to pick out the six prettiest faces from a hundred photographs, the prize being awarded to the competitor whose choice most nearly corresponds to the average preferences of the competitors as a whole; so that each competitor has to pick, not those faces which he himself finds prettiest, but those which he thinks likeliest to catch the fancy of the other competitors, all of whom are looking at the problem from the same point of view. It is not a case of choosing those which, to the best of one’s judgment, are really the prettiest, nor even those which average opinion genuinely thinks the prettiest. We have reached the third degree where we devote our intelligences to anticipating what average opinion expects the average opinion to be. And there are some, I believe, who practise the fourth, fifth and higher degrees.” Great technicians of the past dealt not only with psychology, but also with fundamentals. Two of the greatest technicians many decades ago spoke of the combination of technical analysis with fundamentals. Schabacker in Stock Market Profits wrote: “The reader must not get the impression that such fundamental factors are unimportant in trading, in analysis, in forecasting. They most decidedly are and they most decidedly should be taken into serious consideration when studying any stock or any stock chart for practical purposes. We merely assert that the stock chart itself has nothing to do with such fundamental factors. It concerns itself solely with the stock’s actual record in open-market trading.”

Psychology and Markets

Systematic Trading

Practically all of the outstanding technical students make a careful study, or at least are broadly conscious, of the fundamental situation.” The writings of these master technicians provide in our heritage as technicians a theoretical basis for renewed and increased discourse with the fundamentalists. The basic concepts of crowd psychology and contrary opinion were set forth by Humphrey Neil in the Theory of Contrary Opinion as follows: “Because a crowd does not think but acts on impulses, public opinions are frequently wrong. By the same token because a crowd is carried away by feeling, or sentiment, you will find the public participating enthusiastically in various manias after the mania has got well under momentum. This is illustrated in the stock market. The crowd - the public - will remain indifferent when prices are low and fluctuating but little. The public is attracted by activity and by the movement of prices. It is especially attracted by rising prices. ... Is the public wrong all the time? The answer is, decidedly, ‘No.’ The public is perhaps right more of the time than not; in stock-market parlance, the public is right during the trends but wrong at both ends! One can assert that the public is usually wrong at junctures of events and at terminals of trends. So, to be cynical, you might say, ‘Yes, the public is always wrong when it pays to be right - but is far from wrong in the meantime. It is to be noted that the use of contrary opinions will frequently result in one’s being rather too far ahead of events. A contrary opinion will seldom ‘time’ one’s conclusions accurately. If one relies on the Theory of Contrary Opinion for accurate timing of his decisions he frequently will be disappointed.” Our heritage as technicians with regard to crowd psychology and manias goes well back into the 19th century. Most prominent of the writers were Gustave LeBon and Charles Mackey. LeBon in The Crowd (1895) delineated the nature of crowds. Mackey in Extraordinary Popular Delusions and the Madness of Crowds described some of the great manias of history - the tulip mania, the Mississippi land bubble, and many others. From the great writers of literature over the centuries and from economists and technicians going back over a hundred years, we have writings encompassing psychology and contrary opinion. Starting slowly over a few decades ago, and now building in a crescendo, new voices of fundamentalists are speaking in stronger and stronger tones.

Harold M. Gartley in Profits in the Stock Market wrote: “the author chooses to take the stand that a neat combination of the fundamental and technical approaches forms not only the best plan of operation for the large investor, but is also the logical method for the trader of moderate means. Although it is true that a study of stock price trends includes a reflection of all the fundamental forces, and in addition, the psychological factors which, it has been argued, can substantially affect the correct timing of commitments, nevertheless there seems to be no justification for assuming the attitude of an ostrich, and blindly disregarding entirely the study of causes in favor of a sudy of effects.

In academia, as part of a revolt against the random walk and the efficient market hypothesis, professors at leading universities in the United States such as MIT, Harvard, the Wharton School of the University of Pennsylvania, and Stanford are researching, writing, and participating in conferences about their discipline that they label “Behavioral Finance.” Using elegant mathematics characteristic of academia, the psychology of markets and of participants are the focus of their work. Consequently, they are laboring in the same fields we are working in, though we use the labels sentiment and contrary opinion. The late professor Amos Tversky, at Stanford, - one of the earliest pioneers in the field summarised the behavioral finance point of view:

195


196

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

“Financial markets are influenced by many complex factors … last but not least, people’s reactions to and perceptions of risk … people do not always behave in accord with the classical rational model of economic decision making. The classical analysis assumes that people are perfectly consistent, satisfy criteria of coherence, and have unlimited computational power. The evidence, however, shows that human rationality is bounded by both emotional and cognitive factors.”

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

1. Cognitive Therapy is defined in terms of the Cognitive Model rather than the specific techniques employed. 2. The Cognitive Model stipulates that psychological disorders are characterized by dysfunctional thinking derived from dysfunctional beliefs. 3. Improvement results from the modification of dysfunctional thinking and durable improvement from modification of beliefs.

As a technician, I have sought ways to improve my skills in working with sentiment and contrary opinion in particular, and psychology as applied to investments in general.

Let us view the cognitive model at the beginning. The patient is confronted with a troubling situation/difficulty with spouse, problem at work, etc..

A simple idea occurred to me. The concepts and tools psychotherapists use in the treatment of patients would be helpful. Consequently, I took two courses in cognitive and behavioral psychology for health care professionals at the University of Pittsburgh School of Medicine.

The situation leads to automatic thoughts, emotions, and behavior - that is, physiological responses. Most importantly, the beliefs affect all of these.

Cognitive and behavioral therapy has become a major force in psychotherapy in the latter half of this century, just as psychoanalysis was dominant in the first half. Both fields of psychology were developed by men I can only describe as geniuses. These geniuses not only conceptualized with brilliance, they also had outstanding communication skills to write about their conceptions with great clarity and persuasiveness. For psychoanalysis it was Sigmund Freud. For cognitive and behavioral therapy it has been Aaron Beck of the University of Pennsylvania. Dr. Beck is a classically trained psychoanalyst, who observed he aided patients by helping them with their thinking. He proceeded to develop, write about and teach cognitive therapy.

The beliefs are on two levels. Intermediate beliefs - attitudes, rules and assumptions. These flow from the next level. - Core beliefs. For investment purposes, a famous quote of John Maynard Keynes summarizes the importance of core beliefs. “Practical men, who believe themselves to be quite exempt from any intellectual influences, are usually the slaves of some defunct economist ... it is ideas, not vested interests, which are dangerous for good or evil.” Core beliefs are the most central ideas about one’s self, other people, and the world. Negative core beliefs essentially fall into two broad categories - those associated with competency and those with lovability. The beliefs develop in childhood. The child interacts with others, and encounters a series of situations.

Cognitive therapy was developed by Aaron Beck in the early 1960’s as a present-oriented psychotherapy for solving current problems and modifying dysfunctional thinking and behavior.

For most of their lives, most people maintain relatively positive core beliefs. Negative core beliefs may surface only during times of psychological distress. Negative core beliefs are usually global, overgeneralized and absolute.

Though originally indicated for the treatment of depression, its use has been expanded to the treatment of a great variety of mental problems and illnesses. It aims to have the patient, in effect, become his/her own therapist. To do this, the therapeutic session are, to the extent possible, collaborative, with the therapist giving the patient the concepts and the working tools.

When a core belief is activated, the patient is easily able to process information that supports it, but often fails to recognize or distorts information that is contrary to it. Evidence that supports the core belief is readily processed and then overgeneralized.

A basic text in this field is Cognitive Therapy Basics and Beyond, written by Dr. Judith Beck, the daughter of Aaron Beck. The text is a major source of the materials for today, and is highly recommended reading for further details of the material presented. Judith Beck describes her work thus: “Cognitive therapy is based on the cognitive model, which hypothesizes that people’s emotions and behaviors are influenced by their perception of events. It is not a situation itself that determines what people feel but rather the way in which they construe a situation... So the way people feel is associated with the way in which they interpret and think about a situation. The situation itself does not directly determine how the feel; their emotional response is mediated by their perception of the situation.” Is this not really the same thought John Milton expressed hundreds of years ago by “the mind is in its own place, and in itself can make a heaven of hell and a hell of heaven”? And also, is this not really the same thought expressed by the Wall Street adage, it is not the news that is important, it is the markets’ response to the news. Dr. Lawrence Pacoe, the truly outstanding cognitive therapist and teacher, at the University of Pittsburgh, set forth basics of cognitive therapy as follows:

In the broad perspective, cognitive therapy operates on three levels. 1. The deepest level has to do with core beliefs. 2. Next, we have intermediate beliefs. 3. The most basic level is that of automatic thoughts. Let us see how it all fits together.

Cognitive Model of Psychological Therapy


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Early Childhood Learning is fine for patient work. For market work, we use such matters as what was studied for MBA, for PhD, favorite texts in technical analysis, methods of analysis, indicators, etc. Core or Basic Beliefs: They are absolute in nature, rigid global ideas. They are about self, world, future. Instead of Dr. Beck’s core belief categories of competency and lovability, for investment analysis purposes we use bullish or bearish. Under bullish - New paradigm, inflation and labor costs under control, technology impacting economy, globalization, etc, Under bearish - Too high P/Es, a bubble, cycles remain, labor costs will rise, Fed must tighten a lot, Phillips curve, etc. For investment purposes, let us modify the cognitive conceptualization diagram, as follows.

Psychology and Markets

Systematic Trading

Everything in extremes All or nothing thinking (also called black-and-white, polarized, or dichotomous thinking). You view a situation in only two categories instead of on a continuum, and thus reach only very bullish or very bearish conclusions. Tunnel vision. You only see the negative (positive) aspects of the situation: only see and consider the data that reinforces the all or nothing, very bearish or very bullish. ‘Should’ and ‘must’ statements (Also called imperatives). You have a precise, fixed idea of how you or others should behave and you overestimate how bad it is that these expectations are not met: the Fed must not raise rates, it will result in a severe bear market; P/Es are too high, so the market must fall. Emotional reasoning. You think something must be true because you “feel” (actually believe) it so strongly, ignoring or discounting evidence to the contrary. Emotions reinforce All or Nothing, Tunnel Vision and Imperatives

Cognitive Model of Investment Analysis

Reasoning Process The second broad category of thinking errors may be in the process. Disqualifying or discounting the positives (negatives): You unreasonably tell yourself that positive (negative) data do not count. If one acknowledges one’s “core belief” bias, be aware of the question of disregarding or minimizing contrary data. Magnification/minimization; When you evaluate yourself, another person or situation (market, industry, etc.) you unreasonably magnify the negative and/or minimize the positive. This is the other side of the coin of disqualifying or discounting the positive. Examples of magnification are panic selling at bottom and euphoric buying at top.

Intermediate Beliefs: Attitude, rules and assumptions Since I am bearish, I can not give buy recommendations. Assumption = market will fall and I will be right. Since I am bullish, I must give a lot of buy recommendations. The market will rise.

Mental Filter (also called selective abstraction): You pay undue attention to one (or a few) key negative (positive) details instead of seeing the whole picture. The above three are interlinked: Discounting (eliminating) on the one hand, magnifying on the other, filtering all work together, reinforcing core belief.

Conclusions

Automatic thought. These are the market’s response to news, data, and forces. In addition, they are shown by statements and articles by analysts and commentators in the various media.

The third broad category of thinking errors may be in the conclusion.

Cognitive therapy places great emphasis upon thinking. Thinking often falls into a consistent pattern of errors. In her book, Dr. Beck listed the most common errors of thinking. For investment analysis purposes, just about all of them are arranged in three categories:-

Catastrophizing (or optimizing). You predict the future without considering other, more likely outcomes: In the market, predicting very substantial bear (or bull) market. These predictions may flow from very strong core beliefs.

1. EVERYTHING IN EXTREMES 2. REASONING PROCESS 3. CONCLUSIONS

Overgeneralization. You make very sweeping negative (or positive) conclusions that go far beyond the current situation: In the market, related to catastrophizing or optimizing. This seems at times to be done for marketing purposes. Labelling. You put a fixed global label without considering the evidence that might more reasonably lead to less disastrous (or optimistic) conclusions: In the market, we have bull and bear labels. We should have a third label - A horse market. It is just horsing around.

197


198

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Spurious Correlation Spurious correlation is another problem in thinking. Two phenomena may correlate mathematically, but the one may not cause the other. In medicine, as a surgeon told me years ago, one must understand the physiological relations - the mechanisms. Similarly in investments. Let’s contrast the super bowl indicator, the hemline indicator, and the front cover indicators. The Super Bowl indicator has a high correlation with a good or a bad market year. Roughly, if a National League team, or a former American League team wins the Super Bowl, the stock market will rise. But common sense says there is no causal relationship between the two. It is a spurious correlation. The hemline indicator has tended to correlate with the markets. When hemlines are very high, the rise may be related to general optimism, which spills over to the market. When hemlines are very low, this may be reflected to general pessimism, which also spills over to the market. Now, the dictates of fashion and the need to change styles to stimulate sales may decrease sharply the value of this minor indicator. The magazine cover indicator correlates well with market tops and bottoms. A causal relationship may be due to the fact that by the time an economic trend has been persisting for enough time and with such intensity that editors decide the interest of its readers is so intense that a cover and long article on the subject is justified, the buying (or selling) of the investment community has just about exhausted itself. So the magazine cover indicator has a causal mechanism to support it. The past year has seen an interesting example of the magazine cover indicator. When the cover of magazines featured Viagra, Pfizer’s ant-impotence drug, it was just about exactly the top of Pfizer for the year so far. For treating patients, dysfunctional is the appropriate adjective. But we have to redefine terms for our work. For market and economic analysis, operative is the appropriate adjective. Important distinctions must be made here. Operative means simply contributing to the assessment of future probabilities, or deploying successfully investment funds. The contribution may range from very poorly to very well. An operative belief may be rational and not at all irrational. A core belief may be operative with poor results due to inadequate understanding of how technology and related developments are changing, and how the economy and markets work today, even though it is rationally based upon history and accepted learning. Even the Fed acknowledges problems and is working on understanding the present-day economy and markets. Operative does not imply, or necessarily mean, group thinking, or merging on mania. Poor results from operative thinking may be very independent - a contrary opinion, thus away from the group. Neil warned contrary opinion may be early. As analysts and market practitioners, our objective is not to treat or cure market analysts, practitioners, or governmental officials who impact, or can impact financial markets. Our job is to diagnose and interpret - to reach useful conclusions - better than average projections of probabilities, along with optimum strategy and tactics. We are in the prediction business. The job is to assess future probabilities. In that endeavor, we face a multitude of data and forces. The assessment of the probabilities involves thinking.

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Cognitive and behavioral psychology provide a structured framework for thinking. By the structured framework, thought processes that had been implicit become explicit. By using a framework that is explicit, the analytical and decision process can be sharpened. From logic, we are aware of the basic deduction and induction. Deduction from the general to the specific. Induction, from the specific to the general. Similarly, for the use of cognitive flows in investment work. From core beliefs move down to rules, assumptions, action recommendations and then to market behavior. Or, from market actions move upwards to core beliefs. Situation is broadly defined as the relevant environment. Broadly, the environment comprises unfolding data on the economy, labor conditions, and the monetary, including pronouncements and speeches by Fed officials, on the fiscal, which means congress and the White House, and even more broadly, the global environment, both economic and political. The environment includes broad forces - such as technology, with its seemingly accelerating growth, with its impact upon the economy and financial markets, not to mention industries and companies. Another important force is demographic. These are major social forces such as intensified media activity, the Internet, etc. The problem is information overload, too many factors and forces. The solution is to extract what is really the most important, the most relevant, from the data and the forces, and also what are the most overriding core beliefs that are presently involved. The concept of core beliefs is most useful in market analysis, and aids in integrating elements of standard technical analysis. For example, last summer, the Asian crisis began. TPS wrote on September 8 1998, with the DJIA at 7640.23, the following: “What a central banker should do in times of panic was spelled out over a hundred years ago by Walter Bagehot, in his classic, Lombard Street. He stated on pages 196-197: “...(the Bank of England) must in time of panic do what all other similar banks must do; that in time of panic it must advance freely and vigorously to the public out of its reserves...these advances, if they are to be made at all, should be made so as if possible to obtain the object for which they are made. The end is to stay the panic; and the advances should, if possible, stay the panic.” Awareness of Bagehot’s instruction as a core belief for central bankers led to our correct call on 8 September 1998 for containment and the Fed to lower rates by as much as 3/4 of a point. This forecast was made after analysing some of Alan Greenspan’s comments. For example: “...there is one important caveat to the notion that we live in a new economy, and that is human psychology. The same enthusiasms and fears that gripped our forebears, are, in every way, visible in the generations now actively participating in the American economy. “ “but judging the way prices behave in today’s markets compared with those of a century or more ago, one is hard pressed to find significant differences... As in the past, our advanced economy is primarily driven by how human psychology molds the value system that drives a competitive market economy. And that process is inextricably linked to human nature, which appears essentially immutable and, thus, anchors the future to the pasts.” The TPS concluded that “The Fed will soon lower interest rates.” The following refinement should have been made. When the FOMC lowers interest rates, it will not be by 0.25%. It will be by 0.5% or


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

0.75%. Also, it will be in co-ordination with leading European central banks. The action of the averages and of internal dynamics reveal a very intensely oversold market. Sentiment is so very pessimistic it is giving us a very powerful contrary opinion case for containment and higher prices. Alternatively stated, the stock market is primed to respond to news it perceives as positive. Multiple factors point to a V bottom. The first is the intensity of the oversold levels of both the averages and of internal dynamics. Pessimism is rampant, intensified by the media focusing upon and emphasizing the negatives. The pessimism is so rampant and intense we had a Bear on the cover of the Economist, giving us a classic bottom signal. Consistent with the extreme pessimism is heavy short selling. Note the very heavy odd lot short sales numbers. Upon good indication of a rebound, short covering will intensify the rise. Futures activity will intensify the rise, as it intensified the decline. Fear of “Missing the Move” will result in buying by money managers who sold out, lightened up, or were hoping for higher prices. Lastly, the “Buy the Dip” strategy will be reactivated, resulting in heavy buying by the public.

Psychology and Markets

Systematic Trading

The usefulness of using common sense to supersede a widely held core belief developed in the period following the crash. The widely held core belief at that time was that the adverse wealth effect from the crash would cause a severe recession. The economic impact of the crash was correctly diagnosed. In the 2 November 1987 issue (DJIA 1993.30) TPS wrote about the impact of the Crash upon consumer spending: “While a few yuppies may be forced by the market panic to sell their second Porsche at distress prices - or even delay buying their first one - these are not really widespread economic phenomena. More typically, the average consumer with his salary job - independent of Wall Street - will tend to shrug it off…” TPS would not apply the above reasoning now were the market to react very severely, or crash. The involvement of the American people in the stock market today is much greater than in 1987 with mutual funds, retirement plans, etc. Media coverage of the market, with CNBC, the Internet, etc. is so consuming it presents a grave, but indeterminate probable role in intensifying the emotional reaction to a severe market downmove. Greenspan has stated concern over market ‘overreaction’. A concept or core belief that worked with dramatic success in the past may not work now. The impact of present day forces and conditions upon the core belief must be assessed. Victor Hugo wrote “there are times people’s lives are like pebbles cast forth from a volcano”. To varying degrees, his aphorism applies to markets. Wars can be so powerful in their impact, they cast aside and change usual economic and technical factors. Wars must be intensely evaluated. In this situation a contrary opinion is not always appropriate. Majority opinion is often correct during a trend. TPS told clients to sell stocks on Friday, August 3 1990 (DJIA close Thursday 2864.60) the day after the Iraq invasion of Kuwait.

The crash on October 19, 1987 is also instructive for the integration of standard technical analysis with core beliefs to result in correct assessments of future probabilities and very useful strategic recommendations to portfolio managers. TPS stated in the 12 October 1987 letter (before the crash) (DJIA 2482.21): “...To sum it up, the market is headed lower. The technical condition of the market is so weak that the best strategy is to go to the maximum defensive position one’s operating procedure allows.” Applying cognitive concepts to the letter, core beliefs are evident. 1. 2. 3.

Central bank action for higher interest rates, especially if global, produces lower stock prices. Deteriorating technical conditions in the stock market in a monetary tightening environment will result in significantly lower stock prices. Index arbitrage has been an intensifier of market trends. Along with the deteriorating technical condition of the market and the tightening monetary environment, the probabilities for significantly lower stock prices were so increased that the wise strategy for money managers was to go to the maximum defensive position their operations allowed - i.e. sell all the common stocks possible.

Judgments that did not work can be as instructive as successful judgments. OPEC announced a deal to cut production, and oil stocks climbed. The widely held core belief was that historically OPEC members cheat, so the production cuts will not work. TPS agreed with the widely held belief, and thus delayed recommending oils. But the agreement this time was between heads of state, not just finance ministers. TPS should have given this factor more importance. Also, the sheer strong technical action of oil prices prior to the outbreak while stock prices were weak should have overridden the core belief that OPEC agreements do not work.

199


200

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Core beliefs on the nature of science has investment analytical implications. A good illustration is the uncertainty principal of Werner Heisenberg. He stated that to measure the position and the momentum of a particular particle, we must “see” the particle, and so we shine some light of wavelength on it. …the light when striking the particle could give up some or all of its momentum to the particle. Since we don’t know how much it gave up, as we don’t measure the photon’s properties, there is an uncertainty in the momentum of the particle.

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

for taxes and investment. The tax laws, investors’ preference, and emphasis upon performance renders historical dividend yields irrelevant. Greenspan’s brilliance was using new paradigm thinking to override traditional Phillips curve economic thinking. Indicators and clusters of indicators can change effectiveness over time.

The above helps to understand the uncertainty principle that the very act of observing can affect what is observed. Similarly in investments. The widespread deeply held core belief about an indicator can influence its operation. For example, the value of the 200-day moving average (MA) is a deeply held core belief. So deeply held, it affects the placing of both buy orders and of stops. Interestingly, the 200 day MA was developed in the search to determine the correct MA to preserve short positions from 1929 to the bottom in 1932. To go back to the market, a core belief can be so powerful in its operation that it can even prevail over rising earnings extending over years. The period from 1946 to 1949 saw a sidewise consolidating market extending over three years, despite continuously rising earnings.

Note how Total Equities (directly and indirectly held, market value) as percent of net worth has skyrocketed in the 1990s.

Applying cognitive concepts, two core beliefs were operating. 1. Raising earnings would send stock prices higher. 2. After a major war, a pronounced recession would follow. This second core belief prevailed over the first, resulting in the three-year sidewise stock market between 1946 and 1949. It had its historical precedence - the recession after the First World War. Core belief comparisons apply to industry analysis in diverse ways. A core belief is that earnings drive stock prices. The drugstore chains had enjoyed good earnings growth, and drugstores were a strong group. Then came the news of companies to sell drugs on the internet. The fear psychology from future internet competition drove the drugstore chains from among the strongest to being very weak. TPS has emphasised the new paradigm with regard to the economy and markets. Even P/Es are not too high, when adjusted

Another set of conflicting core beliefs is in regard to broad market action versus selectivity. When they raid the house, they take all the girls and the piano player, too. It is a market of stocks, not a stock market. These contradictory core beliefs impact recommendations and market action. The technical action rebuts or reinforces the core beliefs, with significant practical implications. Differing core beliefs prevail in regard to diversification. Broad diversification reduces risk.

Focussed portfolios have really less risk.

A classic view among fundamentalists is that broad diversification reduces risk. Market action and the better performance of relatively


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

concentrated funds suggest the classic view is not operative today. Economically, earnings growth is becoming narrow - primarily high tech - and competition is intensive and global. The changed economy of today reinforces the conclusion from market action that focussed portfolios have really less risk than broadly diversified ones. Another important set of conflicting core beliefs is growth versus value. TPS has emphasised growth over value. The nature of the unfolding economy, with technology in its broadest sense as the foundation has been providing earnings growth of significant size. Large companies with efficiencies of scale and the substantial capital for research, manufacturers, and marketing on a global scale, also reinforce the growth preference. In statistical studies on growth versus value, the starting date of the study can materially impact the results. Results differ when the start of the comparison is at 1970, or at the bottom in 1974. Large cap versus small cap are conflicting core beliefs. Market action has shown large cap stocks generally have outperformed small caps in recent years. Environmentally, the tremendous capital requirements for research, manufacturing, and marketing on a global scale favours large caps. Large cap companies have the resources, and have been buying, the small companies with niche and/or superior technologies. But of course, truly superb small cap picks can perform spectacularly.

Integration of cognitive tools 1.

Intensity scales are useful in cognitive therapy. For investments, their use can be manifold. The intensity of an economic cluster, or group of indicators, can be projected. For example on a line marked 10%, 20% … 90%, 100%. One prediction of the intensity can be made. Then predictions can be tested by market action. Intensity studies can aid in fine tuning indicators, economic data, and one’s analytic and predictive skills.

2.

Pie charts are another tool used by cognitive psychotherapists. The chart can be used to compare the relative percentage importance of various forces or economic data, or of indicators, or clusters of indicators. For example, % weight to sentiment, to action of averages, to internal dynamics, to fears of Fed action, of the dollar, etc. The pie chart can test the weightings versus the predictive effect versus the market action that results. Of course, pie charts and intensity charts can be used together.

Note the cognitive Model of Investment Analysis on page 5. Use the pie chart and/or the intensity scale in “Core Beliefs.” Set forth assumptions in the “Intermediate beliefs”, then place market action in “Behavior”. From market action, one moves up to assumptions and core beliefs - induction. From core beliefs, one moves down to assumptions and market action - deduction. These variations can be diagnostic - determining core beliefs from market action - or predictive, predicting market action from core beliefs and assumptions. Either way, one’s analytical skills can be improved. Psychology has always been a major force in the market, and will receive increased analytical attention - the Behavioral finance people will see to that. The comments of Professor DeBondt, a leading behavioralist, are

Psychology and Markets

Systematic Trading

of interest to us as technicians. He said: “Behavioral finance and technical analysis agree on the relevance of investor psychology for asset pricing. The two approaches differ, however, in that technical analysis is the arena of practitioners and has not really gone through the rigors of scientific testing. Behavioral finance is part of science. Its adherents start from fundamental behavioural axioms (such as loss aversion or overreaction to salient news) and ask whether the theory built on the axioms can explain the stylised facts around us. Finally, they perform empirical tests to check the theory’s predictive power out of sample. Thus, behavioral finance is much more rigorous than technical analysis. Cognitive and behavioural therapy has advanced the treatment of patients beyond classical psychoanalysis. Similarly, cognitive and behavioral therapy concepts and tools can as a practical matter advance our work as technicians in understanding the psychology of the markets, and of market participants, and even of ourselves as analysts and money managers. The advance will come from working with the concepts and tools, in applying them, in experimenting with them. Perhaps we should add the rigors of scientific testing to meet the challenges from the behavioural finance people. Finally, it is appropriate to recall the words of Claude Bernard, the founder of the science of physiology. “The truly scientific spirit, then, should make us modest and kindly. We really know very little, and we are all fallible when facing the immense difficulties present by investigation of natural phenomena.” One might very well say likewise for the investigation of investment phenomena.

201


202

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

CHAPTER TEN

Systematic Trading Articles in this chapter 204 Futures Trading - Keeping it simple Richard Adcock

206 Money Management Beyond Stops Malcolm Blazey

210 Systematic Trading: It is more about risk than you may imagine... Francesco Cavasino

213 Optimisation - From a Technical Analyst’s point of view Jeremy du Plessis

217 Analysts Make Predictions, Traders Make Money John Piper

Moving Averages and Trends


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

CHAPTER TEN INTRODUCTION

Malcolm Pryor MSTA

Introduction What is Systematic Trading? In essence it is trading using a set of rules covering all aspects of trade management from entry to exit. The rules can be mechanical or discretionary, and can be based on inputs from a wide range of disciplines, not just technical analysis. A full set of rules should cover, amongst other things, instrument selection, trade set up, position sizing, trade entry, exiting at a loss, exiting at a profit, risk management, and also portfolio management if multiple trades will be open. In some cases the effectiveness of the rules can be tested on prior data (back tested) and be refined and improved. But there are dangers in the testing process; too many input parameters or excessive optimisation can generate systems which only work on the past data (referred to as “curve-fitted”) and have zero predictive capability. Systematic Trading goes back many decades and is an important product of the technical analysis approach. Richard Wyckoff, a highly successful trader and exponent of point and figure charting, taught systems in the 1920s which are still taught today, drawing heavily on the concept of relative strength. H.M. Gartley, author of “Profits in the Stock Market” (1935) developed rules-based systems using momentum. His book contains some of the earliest references to moving averages. Richard Donchian, a key figure in the development of technical analysis in the mid20th century, and one of the earliest hedge fund managers, used and published several well-known systems, including his moving average crossover system employing a 5 day and a 20 day moving average, and his 4 week channel break out system, which subsequently inspired several famous traders oriented to technical analysis who are featured in Jack Schwager’s Market Wizard books (including Ed Seykota and Richard Dennis). The impact of position sizing on the performance of systems is often underestimated. Position sizing needs to be small enough for the long term edge of the system (if there is one) to be given the opportunity to materialise without the assigned funds being eroded beyond acceptable levels before that occurs. Two traders with different trading objectives will produce significantly different results using the same system and taking the same signals from the system, as a direct result of their differing position sizing algorithms. Another factor impacting the performance of trading systems is the psychology of the traders operating the systems. Again, two traders using the same trading system may get significantly different results due to mistakes being made, or the system not being followed for various psychological reasons well documented in behavioural finance research over the last three decades. The most mechanical systems become discretionary if the trader sometimes departs from the rules. A trading system has to fit the beliefs and objectives of

Psychology and Markets

Systematic Trading

the trader, and the trader has to have sufficient conviction in a system to stick with it when it goes into an inevitable drawdown. Not all trading systems work in all types of market (e.g. up, down, sideways, volatile, non-volatile) and the trader needs to have an understanding about which types of market suit the system best, and also benchmark expected performance levels for each market type. Some trading systems will degrade over time, particularly ones based on shorter time frames, therefore the trader needs to schedule reviews of the system at appropriate intervals, compare performance to benchmarks and determine the status of the system.

Selected Articles Many Market Technician articles have been written over the years on the subject of Systematic Trading, and related subjects such as position sizing and trader psychology. Richard Adcock’s 1997 article illustrates, without spelling out the detailed rules, how a fairly simple trading system can be created using multiple time frames and a range of techniques including trend lines, technical indicators and intra-day pivot points. Can a system that wins 90% of the time lose in the long run? Malcolm Blazey’s article from 1993 focuses initially on the key concept that system profitability depends on two numbers, win rate and the relative size of winning and losing trades. He then expands into the related areas of position sizing and stop placement. The next article discuss risk management. Francesco Cavasino’s article from 2007 focuses primarily on managing portfolio risk; a key thought underlying this article is that “Systematic Trading is as much about risk management and instrument selection as it is about designing a sensible and profitable strategy”. Should the parameters for a technical indicator being used in a trading system be optimised? The article from 2004 by Jeremy du Plessis recognises that optimisation in the wrong hands can be dangerous, but makes a strong case for optimisation of technical indicators, examining a series of tests conducted on the parameters for entry and exit signals produced by Welles Wilder’s RSI. This thought provoking article suggests that standard indicator settings may be far from optimal, and that the optimal settings may vary by instrument. The last article focuses on how Systematic Trading performance can be impacted by the behaviour and psychology of trader. John Piper’s 1994 article highlights the danger that a system signal may be overridden because the signal conflicts with other “analysis.” In effect the information provided by the additional “analysis” causes behaviour which potentially degrades the performance of the system.

203


204

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Futures Trading - Keeping it simple Richard Adcock

Article originally featured in Market Technician 29 (July 1997)

The shorter term nature of the futures market means volatility is high and trends are often very difficult to read. This results in a greater dependence on technical analysis, which can clearly be geared to a much shorter term perspective than fundamentals. However, I have found that taking a very simplistic view to the use of charts pays greater dividends than using some of the more complex systems. While these have their merits for longer term views, trends that are mostly driven on a minute by minute basis by locals are often better analysed by the use of pure momentum and sentiment indicators. It is also very easy to get blinkered when trading for the shorter term. Watching a 5 minute chart tick in front of you can give some interesting signals but significant turning points are easily missed as longer term support and resistance levels are not seen. Because of this, a longer term perspective must be taken, again using a simplistic approach.

When both weekly and monthly trends confirm each other, a particularly strong position is seen. With stochastics giving slightly earlier signals, we use these for timing, with the MACD used as confirmation. Having established the longer and medium term trends, as well as support and resistance points in the traditional way, the short term trading picture is consulted. As already indicated, locals hold the key to intra day trends and, as a result, sentiment is critical. Any change of trend will occur due to a change in a sentiment and again we have found a simple approach pays dividends. A 60 minute chart together with a 9 period slow stochastic is most useful.

While trendlines are used to give support and resistance levels, sentiment is gauged by using a simple 9 period stochastic indicator. This indicator highlights shifting sentiment, by signalling the lack of commitment to produce closes towards the highs of the month’s range. Sell signals over 80, mark clear tops in prices, opening the way for a prolonged corrective phase. Buy signals are generated when the signal line, having been under 20, crosses above the moving average. Buy signals also materialise when a turn up is seen. This follows a new low for the decline which develops after a non-confirmation by the indicator. This gives the direction of the longer term trend, allowing trading strategies to be structured against a shorter term view. After the monthly view, a weekly perspective is taken, using another useful trending indicator, the MACD. While its normal use is similar to the overbought / oversold signals of stochastics, I have found that crosses through the zero line give more reliable trend change signals.


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

This allows turning points to be seen, but from a much shorter term perspective, using the indicator in the same way as with the daily basis. Overbought readings are scored over 80, while readings under 20 signal oversold conditions. Using this 60 minute time period, it should be remembered that sentiment (often local) is the driving force behind short term price movement. Because of this, we have found again that trend and momentum studies offer the only way of analysing markets. This means that traditional pattern analysis is often of little or no use. Reversal patterns are built over a prolonged period of time, as the battle between buyers and sellers grows, highlighting the change of sentiment. Patterns built up over hours, rather than days or weeks, are not reliable and as such should not be used to initiate positions. What would appear to be classic head and shoulders reversals quickly failed and the original trend resumed.

Psychology and Markets

Systematic Trading

market condition is bullish with breaks of high 2 turning the market very bullish. However, if the two closes are similar, and it is likely the latest trading price can break that of a week ago, then a neutral condition is registered. For the market to be considered as bearish, the latest closing level must be some way under the corresponding close a week previously with little possibility of trading through; breaks of low 2 turn the market very bearish, signalling an acceleration to low 3. By using a combination of simple technical theories, in conjunction with pivot levels, we have found a very reliable way of analysing markets. By using these, a consistent performance is seen which will prove profitable over the longer term.

Contract

High

Low

Close

Pivot

High 1

Low 1

High 2

Low 2

Gilt

109/27

108/29

109/24

109/16

110/03

109/04

110/15

108/17

Bund

100.90

100.02

100.87

100.60

101.17

100.29

101.48

99.72

BTP

129/05

128/35

129.02

128.81

129.26

128.56

129.51

128.11

Contract

Mon

Tues

Wed

Thurs

Fri

Gilt

109.00

109/03

109/04

109/29

108-19

Bund

100.28

100.22

100.27

100.11

99.91

BTP

127.5

127.04

127.75

127.13

127.28

WEEKLY

Pivots When analysing the short term picture from a technical perspective, highs and lows are often scored for no apparent reason. Sometimes, this proves to be because locals are watching levels called pivots. These are based on a very simple calculation using the previous day’s high, low and close. As the chart opposite shows, the pivot, high 1, high 2, low 1 and low 2 are calculated very easily and can thus mark important short term turning points. While sentiment still determines market direction, a number of rules can usefully be followed when using pivots to structure trades. To determine market sentiment, we look at the latest closing level in relation to the closing level a week previously. If it is higher, the

PIVOT CALCULATION

PIVOT = (HIGH + LOW + CLOSE) /3

PRICE SWING = HIGH - LOW

HIGH 1 = (PIVOT - LOW) + PIVOT

Market Condition

Likely Trading Pattern

LOW 1 = (HIGH - PIVOT) - PIVOT

Very Bullish

Between high 1 and high 2 (Maybe to high 3 if high 2 is broken)

HIGH 2 = PIVOT + PRICE SWING

Bullish

Between pivot and high 1

Neutral

Either side of pivot

Bearish

Between pivot and low 1

Very Bearish

Between low 1 and low 2 (Maybe to low 3 if low 2 is broken)

205

LOW 2 = PIVOT - PRICE SWING

HIGH 3 = (HIGH 1 - HIGH 2) + HIGH 2

Richard Adcock is manager, Technical Research, HSBC Futures


206

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Money Management Beyond Stops Malcolm Blazey

Article originally featured in Market Technician 16 (April 1993)

This article is based on a talk that Malcolm Blazey gave at last year’s Annual Conference If you are on the mailing list of the junk mail brigade no doubt you will have seen the claims for trading systems with 85% or 95% accuracy. In some cases these claims are not far fetched but the question has to be asked “Why are they being sold if they are that good?” The answer is that they are often as good as they are claimed but this does not in itself guarantee a profit. A number of STA members, myself included, have developed neural networks (artifical intelligence systems) that trade the markets and have performed up to these levels of 90% accuracy. We are not selling these models and I can assure you that I am not writing this article on a portable computer while sitting on the beach in Barbados. The reason for this strange lapse on my part is that while these models are accurate they do not make any money. This may sound virtually impossible but we must look at the measurements of accuracy to get a truer picture of the situation. Many of the models produced, and especially those produced on neural networks, are looking for either the market to trade up or down on the following day. The quality of the move is not taken into account when deciding whether one particular day’s move is better than the next. For example let us take the situation where the model has predicted that the market will rise for three consecutive days and this prediction is in fact three individual predictions of one day rises. The first two days we see a rise of 5 points and 7 points in the FTSE and on the third day the market drops 15 points. The end result is a 3 points loss but the model is 66% accurate. When viewed on this basis, it is obvious that the system is not all that the headlines on the brochure claim. Systems may actually be profitable but often in the long term that profit is marginal. Where one of the real dangers comes from is when a trader is trading without a system but more importantly without a trading plan. Instead of having a “Claimed accuracy and return on capital” we now have a “Hoped for accuracy and return on capital.” The question is whether we should even try to match those claims of accuracy that seem essential for success. Where many traders have a problem, particularly technical analysts, is that they strive for accuracy where in fact they should be looking at efficiency. In economic terms we should be looking for the utility of each trade or string of trades. I am going to presume that the trader has an idea of where they are going to exit any one trade when it is entered, both on a profitable basis and on a stop loss basis. Let us take an example of a trader who is counter trading a strong downtrend in a market and realises that he must use a fairly close stop. They are willing to go long in the downtrend and expect to exit the trade with a £20,000 profit but would take a loss if the trade went £14,000 against them. This seems reasonable as they are not risking more than they can make and if they played this type of trade they would only need to average 2 out of 3 correct trades to net a profit of £26,000. The first question though is what is the expected probability of success on this type of trade. Counter trading trends is a tricky business. As we have a strong down trend and they are taking long positions let us presume that their true chance of making a profit on any one trade is 0.4 or 40%.

That is on average they will be successful in the long run only 40% of the time. Let us see the net result of the long term effect on their profit and loss account. Maximum Profit = MP Maximum Loss = ML Probability of taking profit before loss (success) = P Probability of failure = Q where Q = ( 1-P ) This is in a mutually exclusive environment where they will either take a profit or a loss but they will never stay with the position for ever with no further action. Average expected result on trades (R) = (MP x P) - (ML x Q) In this particular case we are looking at a maximum profit of £20,000 and a maximum loss of £14,000. We are therefore left with, R = (20,000 x 0.4) - (14,000 x 0.6) R = 8,000 - 8,400 R = -400 (a loss) The average return on this type of trade, presuming the probability of success is correct, would be a loss of £400. This would not mean that you would lose on every trade. You may have a string of profits, but in the long run this is how much you would lose per trade on average. How do I know the true level of probability on this type of trade? The answer is I do not. I can only estimate it but the point I am making is that many traders do not even estimate it, in fact they have no idea what it is. They just hope that the probability of success is high because they have already put the trade on. If the trader was making, over a reasonable period, several thousand trades of which 100 were of this type, there would be a net drain on his profit and loss account of about £40,000. They would be playing a negative expectancy game, and losses in the long run on this type of play are almost as sure as death and taxes. At this point we should look at the difference between positive and negative games of chance. (The word “game” will be used to relate to the individual plays but it can mean “trade” just as easily.)

Positive Games An easy example of a positive game is to spin a fair coin and if it falls with heads to the upside you receive £20 and if it falls with tails showing then you lose £10. Average result = R = (20 x 0.5) - (10 x 0.5) R = 10-5 The average result would be a £5 profit for each spin of the coin. With a starting capital of £100 you would have to be very unlucky to lose over a long period of time and the odds on losing in the first ten spins of the coin would be 1 in 1,024. (One over 0.5 to the tenth power).


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Negative Games One of the most common forms of negative games played is that of roulette. The “House” will pay out 35 to 1 on a single number.

Naturally the thrill comes from winning on these numbers and as this is infrequent the thrill is of course higher. Normally you do not lose a lot of money quickly. Great news - we may be bleeding to death but it’s slow! Let’s look at the numbers for a £10 bet. Maxiumum Profit (MP) = 350 Maximum Loss (ML) = 10 Possible winning numbers (out of 37 with the zero) = 1 Possible losing numbers = 36 R= R= R= R=

(MP x P) - (ML x Q) (350 x 1/37) - (10 x 36/37) (350 x 0.027027) - (10 x 0.972973) 9.45945 - 9.72973

The average play of roulette for a £10 bet costs you £0.27028. For a £10 bet this is not much to lose but it is steady, like a dripping tap. In America the casinos prefer to play roulette with an extra double zero. What’s one extra number when you have thirty seven already? The answer is it may be only one extra number but it means a lot to the casino. In fact it raises the average loss on a £10 bet to £0.525 from £0.27. On this basis it is easy to understand how one casino in New Jersey made nearly 33 million dollars in March 1992, letting the public play negative expectancy games. If you are playing a flat investment rate game (i.e. no reinvestment of profits) then the way you should play each type of game varies. In a positive game you should risk a little and risk it often. In a negative game you should risk as much as you can safely afford to lose and, strictly speaking, do it once. Taking this to its logical conclusion, the “ideal” mathematical evening, in a casino, is to put everything you can afford on one number and leave immediately after one spin of the wheel. Great fun for the winner but not exactly a stimulating night’s entertainment for the other 97% of the players.

Losing while playing a positive game While it is not difficult to imagine losing money when you are playing a negative game it becomes harder to believe that you could lose playing a positive game. Moving back to the original coin spin, the worst case scenario is to lose all your money in only ten bets, 1 chance in 1,024. Normally you would not expect to do this. If you were to lose, then the most likely way would appear to be that you lose after a period of betting rather than straight away. With normal distribution, the chances of losing after a period of playing a positive game are much less. If you were to attempt to

Psychology and Markets

Systematic Trading

double your money from an initial level of £100 to £200 then the odds against ruin increase from 1 in 1,024 to 1 in 59,049 with a two to one favourable payout. While the trader may not be aware of these exact relationships they will certainly realise that the game is not only going their way but is heavily stacked in their favour. The natural reaction is to increase the stakes. Instead of risking £10 per bet with a 2-1 payout in their favour let us increase the bet to £20, still with an original capital of £100. Where is the risk? If the first bet comes up then we will have a capital of £140 , a 40 % increase in capital from the original amount. If you lose then you will still have £80 to play with. Not so bad surely? So, let us look at the odds on losing now. This time, how are the odds on losing everything in one straight run? With a capital of £100 the chances of losing everything, without ever seeing a winner, drop from 1 chance in 1,024 to 1 in 32. The probability of ruin, while looking to double your money, changes from 1 in 59,049 to 1 in 243. The other factor that can turn a positive game into a losing game is randomness. While you may know that 70% of your trades are likely to be winners in the long run, you do not always know which ones they will be. This randomness is not in the sense that the market is responding in terms of random walk but more a case of the system is unable to measure and take into account minor fluctuations which are going to trigger adverse reactions. A good example is in the foreign exchange market where the data source at a five minute level contains too much bad data. A small discrepancy could change the position of a stoploss by just enough to trigger it one day, whereas another day with different data this would be avoided. It is possible that you could benefit from the randomness as well, in the short run. The danger then is that you over estimate the potential of your trading methods and increase the amount of risk you have. When the trading method’s results start to return to the mean, your losses will appear. This time with bigger positions and therefore greater downside to your trading account. If you are looking for the safest route to an increase in your capital then the message is clear. With a positive trading system or method, trade small and trade often. However, if your requirement is to maximise your capital, then you must change your style once again. In this case undertrading becomes almost as serious as overtrading. Before we look at trying to maximise profits let us have a look at the effect that stops have on your trading.

Stop Loss/Profit Ratios Let us consider the following three trading strategies and see which would bring us the most profit in the long run. A. 200 point maximum profit / 100 point maximum loss. B. 100 point maximum profit / 200 point maximum loss. C. 200 point maximum profit / 200 point maximum loss. At first glance it would seem obvious that strategy “A” would be the best and in certain respects it is. However, if we view the market on a Random Walk basis the picture changes somewhat. Working on the assumption that every discrete 5 point move in the market is a random move from the previous one, which is now the best strategy? The market will stagger up and down until finally it will reach the profit level or the stoploss. In this case it would appear that the best bet should still be strategy “A”. In actual fact there is no difference in the long run between any of the strategies, as a computer study of over 50,000 trades showed. The difference in the profit/stoploss ratios is directly offset by the probability of reaching the profit level before the stoploss. It is only when an element of trend is introduced that the ratio takes effect. If the stoploss is close then obviously a short term reaction will probably take you out of the longer term trend. The way to stop this happening is to make sure that there is technical support/ resistance between your current position level and your stoploss level. This will increase the probability of the original trend

207


208

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

resuming before your stop is executed. This is fairly obvious but it is amazing how many stops are set on a “money management” basis where the only management is the size of the individual loss taking no account of the probability of reaching the stop level.

Accuracy against Efficiency Earlier, I suggested that Technical Analysts often try to be too accurate. Many times this is not their fault as this situation is forced upon them by traders who feel that to make money you have to hit the absolute tops and bottoms. With some analytical methods it is possible to capture a good percentage of the highs and lows with outstanding accuracy. By outstanding I meaning within one or two ticks on the futures market and often with several weeks advanced notice. Presuming we have a method of analysis that can achieve this, will we automatically make a large profit? The answer is no. We again have the problem of randomness. The trader/analyst might be trying to predict a selling level which is accurate to within one tick of the actual high but the high is one tick lower than his order. End result no position. You could shade the orders and move the order inside the forecast but this downgrades the utility Of the trade. Even if the position is established, a good profit can only be achieved (looking for the bottom of the next downwave) if consecutive forecasts are correct. More than likely you will see a reasonably good profit disappear and turn into a case of crisis path management. Instead of looking for exact tops and bottoms perhaps it is worth looking for extreme areas to buy and sell with stops set further out. As mentioned above, this does of course lower the utility of the trades but if it improves the efficiency of entering a positive game then we can perhaps offset this with a different approach to money management. This leads us into the area of profit maximisation.

Profit maximisation

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

level of investment, the number of profits would have to equal the number of losses for it to be exact. Looking at a flat rate of investment (a £10 bet), starting with a capital of £100, flip the coin 100 times. You would expect to end up with £600 of which £500 would be profit [100 x £5 (average return)]. What would happen if you could choose how much you bet on each spin. In other words you had the opportunity to reinvest your profits or lower your risk at will. Obviously the lower the amount you bet the less you are likely to win but the more you bet the greater is the chance of losing all your money. Is there an optimum level of money to place as a bet The first stage of answering this question is to look at the Kelly Formula which relates to bets with just two discrete outcomes. f=

The amount that each bet should be (as a percentage of available capital). P = Probability of winning B = The ratio of the amount of a winning bet to a losing bet. f = [((B + 1) x P)-1)] / B f = [((2 + 1) x 0.5)-1)] / 2 f = (1.5-1) / 2 = 0.25 (25% of available capital) Running a random run of 100 spins the first result was: Winning spins = 52 Losing spins = 48 On the flat bet of £10 the resulting profit was £560 (£660 including original capital), close to the amount we would expect (Figure 1). Instead of betting the flat £10, increase (or decrease) the bet to 25% of available capital (this is an approximation, as with hindsight the true value is found to be closer to 29%). Over the same 100 bets the result, including the original capital of £100, is £144,439 (Figure 2).

Figure 1: No Reinvestment

The first choice that you have when trying to maximise profits is whether to reinvest your profits. By reinvesting profits incorrectly you can turn a trading method with a positive mathematical expectancy into a losing system. You cannot turn a negative game into a positive game by money management although you can win on a negative game in the short term. While you continue to trade these systems the order that the trades appear in will not affect the end result. I stress “while you continue to trade” as randomness could take your trading account below the minimum level to continue trading. This constant end result applies whether you reinvest profits or play a level stake game. You only have to remain in the (positive) game regardless of its order to see the same result. As we saw earlier though, the probability of ruin is geometrically increased if you raise the level stakes on a system with a fixed positive mathematical expectation (with a negative expectation you might as well roll over and die now, if you are looking for a long term profit). If we have such a massive increase in the probability of ruin why on earth should we wish to increase our risk? The answer to this is that we can also increase our profits geometrically while not increasing the probability of ruin to an unacceptable level. Ralph Vince, in his book “The Mathematics of Money Management” describes how you can do this by making the profit reinvestment rate a function of the efficency of the trading method, the available capital and the worst trading mistake we are likely to make. To take a simple example of this let’s return to the coin spin. In this case we know our capital is £100, and we know the efficiency as we have a two to one payout. The maximum possible loss will also be fixed (actually in this case a fixed percentage of our available capital). This percentage will be a close approximation of the true

Figure 2: 25% of Capital Reinvested

The increase in capital is so dramatic that it is almost impossible to read the results of the first 50 trades as they are “only” around the 5000 level. With this massive difference why aren’t we all using this system? The unpleasant side to this type of explosive growth in


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

capital is that the ride can be somewhat wild and drawdowns in equity very large, as the chart clearly shows. In Figure 3 we have the results of four different random runs and you can see from this how even a slight variation in randomness can make a large difference to the end result.

Figure 3: Variations due to randomness Random run

Total wins

Total losses

Profit with no re-investment

Profit with 25% re-investment

Maximum profit reached

1

52

48

660

144,439

180,345

2

55

45

750

1,155,516

1,155,516

3

49

51

570

18,054

50,039

4

47

53

510

4,513

11,120

This is all very well in theory but can we use this information in our trading? Yes, we can, by substituting Ralph Vince’s Optimal f calculation for the Kelly formula. The fixed loss is substituted by your largest trading loss which is the best estimation of your likely worst potential loss. The available capital is no problem to calculate but the real problem is estimating the efficiency of the trading system and drawing assumptions about repeating the performance. It must be emphasised that there must be a positive mathematical expectation on the trading method otherwise losses will just accelerate. The end result is that for any given amount of available capital you know how much risk you should have on the next trade. All this depends on knowing the probability of remaining within the same pattern of trading that has occured in the recent past. Ralph Vince’s book outlines a number of methods of calculating the probability of expected returns by using various statistical distributions. I am by no means qualified to argue with his assumptions and from the results I have seen, using computer simulations on actual results, there is a very strong argument to test his theories further in order to put them into practice. It should be said, however, that at least one highly qualified mathematician has cast doubts on just how accurate Ralph Vince’s assumptions are. This does not surprise me as we are discussing the probability of future events but I would not be put off by this totally. Unfortunately the calculations behind Optimal f are well beyond the scope of this article but the key features need to be explained. Vince argues that for any given trading method with a positive mathematical expectation there is a critical reinvestment rate of profits to create the maximum geometric growth. Any move away from this critical level will sharply downgrade the geometric growth of your capital. Naturally overtrading increases your risk but it also decreases your profits on a geometric basis. What is interesting is that Ralph Vince argues that the reduction in risk that you gain by lowering your reinvestment rate, below the critical level, is only linear whereas the reduction in profits is geometric. I personally find it hard to relate to a strictly linear reduction in risk when I look at the large downside change in the probability of ruin, however this is perhaps an academic rather than a practical argument. The fact is that computer simulations show that undertrading does show a high probability of sharply lower profits over a reasonable period of time. One of the big dangers when trading using Optimal f (the critical level) is the psychological strain of drawdowns. Not only can they be deep, they can also be long. Ralph Vince warns that when equity has initially peaked it may take 35% to 55% of the total trading period to surpass this equity peak. Many traders would not be able to handle this especially having seen large profits given back to the market but of course it is easier to handle if you know the risk in advance.

Money management has often been left at the level of “trade with a stoploss” which is really only the first step of controlling risk. While I have tried to show some of the problems that we face I have only scratched the surface of the subject of money management. In the near future I will continue the discussion of this subject with a further article on the subject of run theory and its relationship to the probability of ruin. In the meantime if you would like further reading material I would suggest the following:

The Mathematics of Money Management by Ralph Vince. Published by John Wiley (Finance).

The Commodity Futures Game Teweles, Harlow and Stone. Published by McGraw - Hill (for calculations on Utility and Probability of Ruin).

Malcolm Blazey is the Research Director of the Society and a Manager in Lloyds Bank’s Strategic Trading Unit.

209


210

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

Systematic Trading: It is more about risk than you may imagine… By Francesco Cavasino

Article originally featured in Market Technician 60 (December 2007)

Systematic Trading is one of the oldest hedge fund strategies, born at the beginning of the 1970s, when the first computers became available to satisfy the desire to study historical data using statistical tools. The scarcity of data and computational power kept these methodologies hidden for many years, until the 1990s, when the Windows revolution made data, software and computational power available to both professionals and private investors. In essence, “Systematic Trading” is an asset management method, based on time series analysis, aimed at identifying and exploiting repetitive price behaviour which cannot be considered as random from a statistical perspective. Systematic strategies can be directional or relative-value: meaning that one can buy or sell an asset outright or buy one asset versus another(s) to exploit the difference in relative performance between the two instruments. One simple example of a directional system might be the application of a slow and fast moving average crossover method to identify long or short directional trades in an individual stock. A simple example of a relative-value system might be the application of relative strength analysis to identify long and short positions amongst stocks in the same sector of a stock market. Analytical methods used to develop systematic strategies can derive from fundamental, technical or quantitative approaches. Quite often, the most sophisticated models try to blend different analytical methods in order to obtain a more effective investment process. Similarly, while systematic strategies may trade one asset class only, for example equities, they can also be applied across a more diversified universe of instruments ranging from interest rates, foreign exchange, bonds and commodities. Furthermore, many systematic approaches also seek to diversify across geographical regions and even time frames. Most importantly, Systematic Trading is as much about risk management and instrument selection as it is about designing a sensible and profitable strategy. For example, there is a huge amount of literature published on moving average calculation and optimisation methods, while risk management issues and instrument selection are often overlooked.

Instrument selection Harry M. Markowitz and his CAPM showed that positions which are highly positively correlated will increase the overall portfolio risk; this is why most professional systematic traders attempt to trade a high number of uncorrelated markets, ranging from soft commodities to equities and grains, from energy and foreign exchange to short-term interest rates. They are looking for lowly correlated instruments to aid their portfolio diversification to which they apply their trading models. As a result, a multi-asset class portfolio is likely to be more efficient than one focused exclusively

on one area. However, even the exercise of diversification is not easy. Correlations are not stable at all, they are much more volatile than returns and volatility. In particular market conditions - such as a flight to quality - correlations will jump, making the portfolio risk measures increase dramatically and unpredictably. Most importantly, the more a portfolio composition is based on the correlation estimates, as in relative value trading, the more the risk will be likely to jump when a sudden correlation breakdown appears. So many times, from a long term correlation point of view, a portfolio may seem to be well diversified. However, when an unexpected piece of information reaches the market, all positions may suddenly become correlated and portfolio risk increases. Frequently this is created by the positioning of investors, which may have one macro view expressed in many different ways. For instance, in the recent past, short volatility, long equities and long Latin American currencies, would all have reflected the same positive view on the global economy. These positions have since become even more correlated than that which a long term correlation matrix would suggest. Should an unexpected announcement, such as a weak payroll number, reach the market, it is very possible that traders of different asset classes will react in a similar way, pushing up the absolute value of correlations and portfolio risk will jump. Suddenly we could see the realization of significant profits or losses, which is sometimes referred to as “fat-tails” or Kurtosis. One way to anticipate and mitigate these issues is to analyse the data series of the instruments you would like to trade under specific market conditions. What many do is to evaluate how correlations changed in a peculiar historical period, for instance in a financial crisis. In this way, they are attempting to see how balanced their portfolios would have been in those specific market conditions. This is definitely a sensible way to evaluate the potential kurtosis of the portfolio. On the other hand, any such analysis is still only based on what happened in that particular period. As a result, such historical simulation is fraught with danger as subsequent shocks could be different, with unexpected correlation changes. It is very difficult to predict the shape of future financial market dislocations but at least one can do one’s homework on past events. Usually the best instruments to combine in a portfolio are those that have no direct economic relationship. For instance, the economic relationship between AUD/JPY and NY Coffee futures could be reasonably expected to be quite loose and should not be variable in case of a financial turmoil. Unfortunately with the globalization of finance it is becoming increasingly difficult to find non-correlated instruments. That said, introducing correlation stability analysis into an investment process, while selecting instruments to trade, will help in obtaining a more theoretically efficient portfolio and generate a better understanding of overall risk in times of both normal and abnormal trading conditions.


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Risk allocation After having gone through the process of instrument selection, strategy design and testing, several statistical tests such as stress testing, scenario and what-if analysis should be constructed in order to estimate the potential future losses and to calibrate risk according to the given mandate. In this business, the correct determination of the targeted risk is a step which is at least as important as strategy design. It is industry practice to express the sizing as a function of capital and be aware of the exposure limits. These limits can be expressed as gross and net exposures, beta exposure, leverage and so on. This way of looking at the portfolio is partially misleading from a risk perspective: being invested in Tesco for the 5% of the capital under management does not tell us much about the risk. The only thing measured is the exposure towards the stock, not its riskiness. Any sizing done as a function of the capital managed is calibrating risk only as a by-product. Sizing should be done as a function of the fluctuations of the asset traded i.e. its volatility. Investing 5% of capital in Vodafone has a different risk compared to 5% invested in BHP Billiton: the yearly volatility of Vodafone may well be around 25% while BHP’s could be as high as 40%. Without considering the impact of correlations, having positions rescaled as a function of volatility should generate a more diversified and efficient portfolio. The exercise of rescaling positions as a function of volatility can be done with a variety of indicators: from the standard deviation of historical returns to Average True Range (ATR) or by computing the Value at Risk or VaR, which measures the maximum loss that could be incurred in a set period of time at a certain level of probability. At a single position level, for the sake of accuracy ATR is probably

Psychology and Markets

Systematic Trading

the most effective measure because it also considers the intraday swings while the others consider just close-to-close variations. By calibrating the size as a function of the instrument variability, the strategy will continuously be adjusting its size maintaining the targeted risk. In this way one can manage markets’ volatility, without being driven by it. Targeting the correct amount of risk, given the risk limits associated to the mandate is critical. Let’s say there are two traders: Lewis and Fernando, running the same system called “Silver Arrow” on the same instruments and asset classes, with the same amount of capital and trading limits, over the same period. Can Lewis end up hitting his portfolio stop-loss while Fernando generates a profit by year end? Certainly he can. Lewis could target an average risk level too high for the given mandate and end up being stopped out before the realisation of the following run up. Meanwhile the other trader by keeping a lower targeted risk, does not trigger the stop loss and still has the chance to recover the losses generating a positive return for the period. So, even with an overall profitable strategy, a trader can trigger his stop loss if he is targeting a risk which is too high, given the mandate. In a way, risk budgeting is more important than the strategy itself and the above mentioned example demonstrates the implications of being too greedy. At portfolio level, risk metrics are expressed in a variety of ways: stop loss year-to-date, realised volatility and Value at Risk (VaR) are among the most common. Applied to a directional system, VaR and realised volatility can help determine the actual risk of the portfolio in normal market conditions. Financial literature describes many ways to compute these statistics - especially VaR - each one based on some underlying statistical assumptions. One of the most interesting VaR calculations is called “non-parametric”

211


212

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

or “historical”: it just computes the value associated to the n-th percentile of the historical returns distribution of a portfolio with no underlying assumption on the shape of the distribution. This allows the analyst to capture the overall instrument’s volatility, correlations and even the sudden changes in correlations which generate the fat-tails across the analyzed period. Computed on at least a couple of years of daily data, it should provide a fairly reliable measure of risk and its implementation is also quite straightforward. Realised annualised volatility is the standard deviation of the realised daily returns of a portfolio. It is a measure of the realised risk of the book and it is an industry standard. For directional portfolios it is a useful tool to determine the average risk taken by the manager. Realised annualised volatility and parametric VaR are statistically linked together. In a nutshell, under the assumptions of normality, stationarity of volatilities and correlations and independency of portfolio returns, an annualised volatility of 16.00% should correspond to a daily standard deviation of 1.00% and an average daily VaR @ 97.5% probability of 2.00% meaning that a book that has 16% annualised volatility should not lose more than 2% in one day at 97.5% probability. By taking advantage of the statistics associated with the normal distribution, we also know that events of +/- 2 standard deviations will happen over time with a probability of 2.5%. Assuming an Information Ratio 1 of 1.00 2, a volatility of 16% and the normality and independency of returns assumptions, a negative annual return should roughly be realised one year out of seven and a -16% annual return or more will be realised one year out of 40. Assuming we have a stop loss year to date of -16% with an information ratio of 1.00 and targeting a volatility of 16%, statistical simulations demonstrate the worse drawdowns could be in the region of 30% i.e. -2 standard deviations. Given the statistics that many strategies generate when back tested, these numbers sound much more worrying... or are they just more realistic than our over-optimistic expectations? Especially when considering that the hypothesis of normality, stationarity and independency are quite simplistic and not exactly the most conservative ones! Results can be exciting in backtesting but it is very difficult to move away from these figures. Too often, system designers focus their efforts on the historical results of their system without realising that those performance analytics are just a reflection of a limited data sample drawn from an unknown population. They should be followed as an indication - rather like a map in a treasure hunt, not as the output of a precise navigation system. In order to obtain a more realistic picture of risk, some stress testing should also be applied. Scenario and what-if analysis will help in formulating estimates for market conditions which did not happen in the historical dataset but that could still occur in the future. However, even stress test design is full of perils and designing a stress test is more of an art than a science. In essence it is about designing a market scenario which might never happen or may occur on a handful of occasions but which is also plausible from an economic point of view. One can increase volatilities, change correlations, create trends, manipulate the data in a million ways but defining where lies the fine line between the “possible” and the science fiction is not easy and it is completely arbitrary. Alternatively, it is also possible to let the computer randomly design a huge number of scenarios and then look at the aggregate results. In this case some of the outliers will be truly extreme and should be considered “cum grano salis”. However, even if some of those scenarios are meaningless, they will give a deeper understanding of the risks the portfolio will be running.

Conclusions This brief discussion is aimed at showing that portfolio composition and risk analysis play a critical part in Systematic Trading development. They are far more important than defining the optimal length of the RSI or the most effective method of calculating the moving average. Once the mandate and trading limits are defined and the Information Ratio estimated through the strategy and the time series, targeting the “right level of risk” becomes critical, just as in the Fernando and Lewis example. There are very few things we can control in trading - as in life generally. It is not possible to control the profitability of the strategy; one can only design it according to a given set of parameters, estimate the profitability on the historical data at disposal and run it. Furthermore, it is not possible to control volatilities and correlations either; one can only estimate themex-post and react to their sudden changes. However, what one can do is to size the bets using historical data, quantitative tools and some common sense, adjust sizing according to volatility changes and evaluate by how much one could potentially go under water... And believe me, that in itself is a fairly daunting task.

I

Information Ratio = Average Annual Return/ Annualised Volatility

2

In real life, for a directional systematic diversified portfolio an Information Ratio of 1.00 is an excellent result.


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

Optimisation - From a Technical Analyst’s point of view Jeremy du Plessis, FSTA

Article originally featured in Market Technician 50 (July 2004)

This is a summary of the talk given to the Society on 14th April, 2004 When the IBM PC arrived 23 or so years ago, many technical analysts thought that its power would unleash great riches. The ability to optimise and back test was thought to be the holy grail we had all been searching for. Sadly it was not the case and optimisation was abandoned by many in the 1980’s. A wise man once said that optimisation is just curve fitting and that optimisation simply allows you to forecast the past with 100% certainty. He was right. Optimisation takes a set of data and then works out what would be the best way to have analysed it with the benefit of hindsight, and if you apply those results to the same set of data you will have, by definition, fitted the curve. He also said that anyone stupid enough to optimise should optimise one section of data and then test the results of the optimisation on a different section and finally apply it to a third section. He was right about that too. His arguments are nothing new. Its been heard before and will be again. The problem is the wise man was a statistician, not a technical analyst, and that is the traditional way that a statistician would look at it. Statisticians do not understand the market or market charts. They see price data as a time series of numbers. They do not understand that the price data has ‘life’. They will only work on data that has no autocorrelation (autocorrelation describes a condition where data points in a time series are not independent of each other). If statisticians find autocorrelation in a data series they will use techniques such as first differences to try to eliminate it. They then work on the resultant lifeless data. Technical analysts on the other hand do understand the market and market charts. They understand the charts are created by price data and that the price data is created by human beings. They understand that these human market participants are subject to human emotions which affect the price

and consequently affect the charts. Technical analysts hope and believe that autocorrelation exists because if it did not and today’s price change is independent of tomorrow’s price, technical analysis would be worthless. Technicians realise that because prices, and consequently charts, are created by humans, they have human traits such as trends and pattern repetition. Technicians realise that each instrument has different characteristics and so understanding the movement of one, does not mean that another is understood. They understand that the characteristics of the price movement are different depending on whether the price is in an uptrend, downtrend or sideways trend. So, let’s look at optimisation from a technical analyst’s point of view. In doing

Chart 1

so, let’s stop to think for a moment what technical analysis is. Technical analysis the study of price through the use of charts. It’s the study of the past in the belief that it can tell us something about the future. It’s the understanding that patterns in price and indicator charts repeat. That is what chart reading is all about. Technical analysts look for ‘things’ in the chart that have proved reliable in the past - such has shapes, patterns, indicator movement and so on - on the premise that they will occur in the future and will therefore assist us in making decisions. If you look back and inspect a chart and notice that every time there is an inverted head and shoulders pattern, the price has a significant rise, or every time the RSI breaks through the 30 level and is rising it’s a good buy signal, isn’t this mental optimisation - looking at past history and applying what you have

213


214

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Pattern Recognition and Pattern Analysis

Moving Averages and Trends

find the combination that yielded the greatest profit.

Chart 2

Unfortunately it doesn’t stop there. We have made the assumption that buying and selling on a break of a level is the best way to read the RSI, but maybe it isn’t. Perhaps it would be better to buy on a break up of a level but to sell when the RSI crosses down through a moving average. But what moving average? We would have to test all moving average periods. Perhaps it would be better to buy when the RSI crosses the moving average and sell on a level. Perhaps it would be better to buy when the moving average of the RSI breaches a level. We will not know which RSI period is best and which is the best combination of entry and exit signals until we have tested all of them.

learnt to current data? Every time we look at a chart we are, in effect, mentally optimising. The problem with mental optimisation is that it takes a considerable amount of time and it is possible that you could make a mistake. So why not use mathematical optimisation instead? It’s quicker and far more accurate. To do this, you have to decide what you want to achieve. It’s a pretty good guess that profit is the motive, so you could take a section of data and then test a range of conditions which generate buy and sell signals and mathematically decide which set of signals gives the greatest profit. Let’s take RSI for example. When Welles Wilder developed the RSI 30 years ago, he decided on a 14 day period. He also decided that a buy would be indicated when the RSI broke above 30 and sell when it broke down below 70. Presumably he did some mental (perhaps even mathematical) optimisation in order to reach that conclusion. How else would he have arrived at those parameters? But do they apply to all instruments all the time? What if the 12 day RSI gave better results when breaking above 22 and below 75? How would you know? Manual inspection of every chart would be tedious and lead to errors. You could try it mathematically, by trying every RSI period from, say, 5 to 50. Then for each period, you could test buying at a break up of every level between 5 and, say, 95 stepping up 1 at a time. You could then test selling at a break down

from every level stepping down 1 at a time. It’s a complex test. For example, the optimisation would start with a 5 day RSI and then record a buy when the RSI breaks up through 5. Once it has done that, it would have to test a breakdown level starting with, say 95, then 94, 93 and so on, recording the gains or losses from each level. Then it would move to the next buy level and perform the same tests again, and so on. We would have to test every buy level and every sell level for every RSI period and eventually we would

Chart 3

Having done many of these tests, one thing becomes very apparent and that is, exit signals must be guaranteed. Welles Wilder suggested an exit when the RSI breaks down through the 70 level. It usually generates a very good exit signal, but what if, after giving a buy signal, the RSI never goes above 70 and the price falls. You have no way of exiting your trade. Chart 1 (on the previous page) shows the FTSE 100 index with a 14 day RSI using Wilder’s buy and sell criteria. It works fairly well until 2000 when a buy signal was generated because the RSI broke up through 30. The index then fell over 40% before generating an exit in 2004 when the RSI did finally go above 70 and break down through it. In fact during the bear trend of 2001/2002 the RSI fell below 30 and generated repeat buy signals which


Indicators and Momentum

Elliott Wave and Fibonacci

were not exited until 2004. Immediately the importance of guaranteed exit, no matter what the conditions are, is paramount. The only exit that can be guaranteed is a trailing stoploss. But knowing what percentage stoploss to use can only be done by optimisation. So an optimisation exercise could be performed that testing every single entry condition together with a range of percentage trailing stoploss exits. Chart 2 shows Wilder’s standard 30 entry signal with a 19.5% trailing stop giving the best exit. For a long term investor, a 19.5% stop loss is quite acceptable but not for a short-term trader. Simply reducing the maximum allowable stoploss to, say, 10% will not solve the problem, because patently the best stoploss is greater than 10%. There is another technique however which impacts on the stop loss percentage, that is signal delay. Signal delay means that trading on a signal is delayed by a predetermined period in order to reduce the chance of a whipsaw. The strategy is to wait a number of days after the signal and only act if, after waiting the required period, the signal is still in place. This avoids the 1 day spikes through the stoploss. Signal delays are designated (t+1), (t+2), (t+3) etc., where t is the day of the signal. By definition therefore it cannot be acted on until day (t+1) because by the time the signal is generated on day (t), the market has closed. (t+2) means do not act on day (t+1) and if at the close of day (t+1) the signal has not been cancelled, then act on day (t+2). Knowing what delay to impose is part of the optimisation too. It is to do with the characteristic, some would say, the volatility of the instrument. Taking the 14 day RSI again with a 30 level entry signal and testing various stoploss exits with various signal delays, the result is a 9% stoploss with a (t+3) signal delay as shown in Chart 3. Notice the spike through the stop loss in March 2003 which did not trigger a sell signal because the delay is (t+3). But what if there is a better entry point than a break up through 30? Going back and checking by optimisation will answer the question. The point is, you have absolutely no idea which has been the best entry method in the past without running an optimisation. Chart 4 shows an optimisation conducted on the FTSE 100, testing a range of RSI periods with a full range of break up levels together with a range of stoploss levels. The result is a 5 day RSI breaking up through 16 as the entry and an 8% trailing stoploss. Both signals have a (t+2) signal delay imposed.

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Mathematical optimisation is good. It does tell you things you would not otherwise know, but what about the second thing the wise man said about optimising and testing the result on different sections of data? An optimisation to find the best entry and exit signals could be run, but on what data? Technical analysts know that market conditions continually change and what worked last year in isolation may not work this year. This writer believes that it is a waste of time to randomly select a section of data, optimise it and then apply the results to another set of data. There is little chance of it being consistent. The market is a moving target, therefore it is important to take as much of the latest data into account when optimising, but it is important to re-optimise regularly as new data is added. Traditionalists (and statisticians) will throw up their hands in dismay but, if it is done sensibly and with a few rules, it does seem to produce the best ongoing results. The first thing to do, is to decide on one’s time horizon. There is no point finding which was the entry and exit that worked best over the last 20 years if you are a short-term trader. The result is a compromise which caters for all market conditions. If you are a long-term investor, then of course, the more data, the better. Short-term traders need to consider how much market information they require to assess a signal. This may be six months, one year or two years. The shorter the trading horizon, the less data should be used in the optimisation. This is because short-term traders require the optimisation to ‘squeeze’ as much profit out of the data Chart 4

Systematic Trading

as possible, which means going for the smaller trades. The next thing to consider is dealing costs or the ‘penalty’ for trading. If there was no penalty, the optimisation would consider every gain, no matter how small, as a profit. For longer term optimisation, something like 2% entry and 2% exit commissions should be applied. For shorter term, these can be reduced. Finally and most important, because the optimisation is conducted on the latest data, it is important to ignore any open trades in the result. This means that any open trade will be as a result of optimisations prior to the trade and is therefore not included in the total profit which determines the best technique. Chart 5 shows an optimised RSI of BG Group. The result is a six day RSI where entry is a break up through 20 and exit is a 9% trailing stoploss on the price. Both are subject to a (t+3) signal delay. Notice that the last trade is still open, which means the best entry and exit conditions were determined without taking any data to the right of the last entry signal into account. Notice that the last trade is still open, which means the best entry and exit conditions were determined without taking any data to the right of the last entry signal into account. The optimisations shown here apply to RSI only, but optimisation can be extended to any technical indicator and chart, including Point and Figure. It can be applied to all time frames. It can be applied to short trades as well as long trades and it is interesting to see that the best exit out of a long trade is not the best

215


216

Theory of Technical Analysis

Dow Theory, Wyckoff and Volume

Charting Types, P&F, Candlesticks

Chart 5 entry into a short trade. Optimisation is not the ‘holy grail’, but what it can do is tell you how you should be reading indicators. In doing thousands of optimisations, a number of valuable lessons have been learnt. The best exit is usually a different type of signal from the best entry. Exit signals need to be guaranteed or they may never be triggered. A trailing stoploss is one of the best and most reliable exit signals and imposing a signal delay to both entry and exit signals does improve the results considerably. Constant optimisation of the same indicator and instrument will adjust the results as the characteristics of the price change, but ignoring the last open trade ensures that the optimisation is not meaningless. Optimisation can be dangerous in the wrong hands but if used correctly and wisely, it can teach you a lot about the indicator.

Pattern Recognition and Pattern Analysis

Moving Averages and Trends


Indicators and Momentum

Elliott Wave and Fibonacci

Gann Analysis, Cycles and Forecasting

Psychology and Markets

Systematic Trading

Analysts Make Predictions, but Traders Make Money By John Piper

Article originally featured in Market Technician 19 (March 1994)

One of the points which became clear when giving our seminars, was the almost total lack of analytical input into the methods which we are putting forward for successful market trading. As I developed these systems this was always fairly clear to me, but it came as something of a shock, I think, to some of the delegates. Indeed this is a fairly revolutionary idea, perhaps even sacrilegious in some quarters. One can imagine the title of this article raising a few hackles at one of the monthly STA meetings true though it may be. The impact of this statement is that analysis is a red herring - just one more dead end in the road to trading success. The purpose of this article is to examine this proposition. I think the first point to make is that before we can dismiss analysis as a red herring we have to know what it has to offer. So the novice perhaps has no choice but to get to grips with the full range of analysis, be it technical, fundamental or both. After all it is only once he has mastered some of these techniques that he can judge whether they are useful or not. The second point to make is that clearly analysis is useful to some people. 90 per cent of traders may lose money (perhaps pouring scorn on the title of this piece) and many of those will be using some form of analysis - but of the winners some of those also clearly use analysis techniques. So I am not condemning analysis per se, indeed as I have written about it for some years readers may consider it somewhat absurd if I were to do so. However I have always stressed that I consider trading and trading skills to be of far greater importance - and in this I include reading market action. For example seeing and interpreting a failed breakthrough (or re-test) is not something I consider an analytical function but a function of trading. Clearly this is open to debate but this draws the distinction between the two for our purposes. Perhaps it would make the distinction clearer if it is borne in mind that the trader may have to act pretty quickly if he is going to benefit from a failed breakthrough, there is not usually enough time for an analyst to see the pattern and then contact the trader. However analysis can be the mother to the “view” - and the view can be fatal. This is because we must not allow ourselves to be swayed by our views, fixed views are invariably fatal - if not initially then over the course of time. So why is analysis the mother of the view? Because we have a tendency to believe what we want to believe. The more analysis we do the more we will be convinced of our original “view” - i.e. we may be achieving precisely nothing, indeed worse than nothing. Now we can overcome this problem by being particularly disciplined, but an enigma remains if first an oscillator is giving an over-bought reading, suggesting a possible sale, then any trend following indicator will be in “buy” mode - i.e. they will tend to always contradict themselves.

This article is not designed to cover “emotional” trading which is the root of most traders’ losses - you can hardly expect to make money if you have no system and act on impulse. But analysis can fall into the same camp and it can be more subtle. All those knobs and whistles, all that computer hardware and software, all that work - it must mean something, mustn’t it Yes! For most it means more losses. Also the facts of the matter are that a good system will out-perform a good trader. This also may be a somewhat controversial statement and it must also be considered something of a generalisation. But if you are a supertrader you won’t need a system - or you already have one. If not this statement is true. However let’s make this statement even more specific. I have one account which I am trading which was up 40 per cent at the end of September - however, if I had been trading the system outlined in our recently published “Trading Manual” then the gains would have been significantly higher. So personally I can say that methods utilising very little analysis (and perhaps that little bit itself is a negative) have outperformed my trading which involves a greater degree of analysis. Hardly conclusive I agree but significant to the extent that any trader who is not sure of the right direction to take should carefully consider the points in this article. Now you may say, “how can I trade the market without any analysis?”. Well we are not going to explain the details of the approaches outlined in our “Trading Manual” but both the IOP and the Options approach are “market driven”. This means that the market determines what positions are taken. So you don’t say, “my analysis says I should go long”, you say “the market has triggered me long”. This of course is the key - price is the most important feature of the market, it has no view, there is no difference of opinion, the price at any one time is (usually) fixed and certain. Our systems are based on that certainty.

John Piper is the editor of The Technical Trader, 76 Nunnery Lane, York Y02 IN. 0904-636407

217




The UK’s professional body for Technical Analysts. Founded in 1968. The oldest of its kind in the world. The Society of Technical Analysts Dean House Vernham Dean Andover SP11 0JZ tel: +44 (0) 20 7125 0038 info@technicalanalysts.com www.technicalanalysts.com

Copyright © The Society of Technical Analysts ISBN 978-1-3999-3292-9