the simple - the technical analyst...chicago board of trade’s “the market profile®” to its...

The Simple

Correction

Turning Japanese: A view of the Nikkei

Can you afford not to use

point-and-figure?

The pros & cons of technical analysis

www.technicalanalyst.co.uk

Advertisements - STA and Paritech

February 2004 THE TECHNICAL ANALYST 1

WELCOME

W elcome to The Technical Analyst - a new monthlymagazine devoted entirely to the subject of technicalanalysis and charting for traders and professional analysts.

We aim to cover the wide range of markets and techniques thatconstitute the world of technical analysis. Each issue willinclude in-depth features on analytical techniques, market viewsfrom renowned practitioners, the latest product news and inter-views with key market players.

Our first issue includes an academic perspective on technicalanalysis before going on to look at specific techniques and pat-terns, such as point-and-figure charting and the simple ABC cor-rection.

In the Market Views section, we look at the US dollar and USstocks and consider how much further the currency will fall or ifthe Dow Jones will follow. Other market views consider theNikkei and the British pound.

We hope to make The Technical Analyst a dynamic forum fordiscussion of technical analysis techniques and to keep youinformed about the latest developments in this area. We wel-come any comments or suggestions you may have to improveour publication.

Matthew ClementsEditor

2 THE TECHNICAL ANALYST February 2004

CONTENTS

The Pros and Cons of Technical Analysis

Richard L. Weissman, Oxford PrincetonUniversity

04

13

19

41

42

47

Product News

Subject Matters

The Technical AnalystTalks To...

Editor: Matthew Clements (MSTA) [email protected]

For subscription or advertising information please contact: Jim Biss ([email protected]) or complete the subscriptionform online at www.technicalanalyst.co.uk Subscription rates UK: £275.00 per annum Rest of world: £325.00 per annum

The Technical Analyst ISSN(1742-8718) is published by Clements Biss Economic Publications Limited 10-12 King Edward's Road, London E9 7SF Tel: +44 (0)20 8533 3025 Web: www.technicalanalyst.co.uk

Art, design and typesetting by all-Perception Ltd www.all-perception.com

Book Review

Commitments ofTraders Report

Training & Events Diary

The Complete Guide to Point-and-FigureCharting by Weber and Zieg


© 2004 Clements Biss Economic Publications Limited. All rights reserved. Neither this publication nor any part of it may be reproduced, storedin a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the priorpermission of Clements Biss Economic Publications Limited. While the publishers believe that all information contained in this publication was cor-rect at the time of going to press, they cannot accept liability for any errors or omissions that may appear or loss suffered directly or indirectly byany reader as a result of any advertisement, editorial, photographs or other material published in The Technical Analyst. No statement in this pub-lication is to be considered as a recommendation or solicitation to buy or sell securities or to provide investment, tax or legal advice. Readersshould be aware that this publication is not intended to replace the need to obtain professional advice in relation to any topic discussed.

Ending Diagonals on the Dow Jones

US Stocks, GDP Growth & a 40-year Chart

Turning Japanese

Outlook for the British Pound

Can You Afford Not to Use Point-and-Figure?

Sector Rotation is Alive and Well

Interpretation of Volume and Open Interest inthe US Treasury Bond Futures

The Simple ABC Correlation

A Test of Wyckoff

Trading Partial Declines and Rises

07

08

10

12

21

28

32

34

36

38

Market Views

Techniques

FEBRUARY 2004


Product News

UPDATED SCANNING

SOFTWARE RELEASED

BY PFSCAN

Pfscan has launched the latest version of its scanning software. According to the company’s chair-man, Dave Baker, “Pfscan 2 is still the only point-and-figure based scanning software on the market capable of locating over 20 differ-ent buy and sell signals.” The software also includes market timing tools, such as the ‘Bullish Percentage’ chart, as well as a backtesting module that assesses signal performance.

Mr Baker has also told The Techni-cal Analyst that a real-time Pfscan is under development and is likely to be eSignal compatible.

TESTERS REQUIRED FOR SNAPDRAGON RT

SnapDragon Systems will soon be releasing a new version of its real-time charting and technical analysis software, SnapDragon RT, for beta testing. SnapDragon RT is designed to work with a variety of datafeeds and also directly with trading platforms, thereby eliminating the need for a datafeed. The aim is to increase the product’s flexibility and ease-of-use. If you are interested in becoming a beta tester then please email Adam Hartley at SnapDragon Systems: [email protected]

MAN FINANCIAL TO OFFER TRANSPARENT

FX PRICING

As The Technical Analyst goes to press, Man Financial, a leading broker on the major securities and futures exchanges, is launching a new online FX platform that offers a unique form of price transparency. According to Man Financial, a choice price may even be displayed and trading clients “will benefit in the knowledge than no-one is reading them or skewing the bid/offer prices.” In return for this quote transparency, clients pay an agreed transaction fee. The system will take limit, market and stop orders. The minimum dealing amount will be $100,000.

eSIGNAL ADDS ADVANCED CHARTING TO ITS

ANALYTICS PLATFORM

eSignal, a division of Interactive Data Corporation, has announced the immediate availability of eSig-nal 7.5.

This upgrade to its market-analytics platform furnishes subscribers with new technical studies, including Kase StatWare, Bollinger Band Tool Kit and Jan Arps' Sigma Bands. Each of these studies provides proprietary indicators that automatically adapt to changing market conditions, increasing an investor's views in volatile market conditions.

The eSignal 7.5 update also offers several enhancements to the advanced charts feature. Most notable are the improved symbol and study overlays with better axis management, several marketpro-fileSM additions and the newly available indicator, Andrew's Pitch-fork.

Other additions to eSignal 7.5 include extended intraday data, added market depth, news from Dow Jones NewsPlus and AFX News, and full integration with eSignal Market Scanners.

Q-DATA TO ADD US

STOCKS AND

FUTURES TO ITS

DATA SERVICE

Q-Data has announced that US stocks and futures will be avail-able on its ‘End of Day’ service. Like the other data available on this service, the information will be presented in MetaStock format. This allows the data to be viewed in any charting pack-age that reads MetaStock format, such as MetaStock, OmniTrader and MTPredictor. In addition, a Market Scanner and Stop/Loss alert system have recently been added to the software. Both these facilities create watchlists that are refreshed immediately after updating for the end of day prices.

Product News


FutureSource has unveiled its

redesigned flagship website,

futuresource.com. The redesigned

site boasts a number of new

features, including new charting

functionality, market-specific

content, an all-new “Trader’s

Tools” section, an expanded

education section and additional

analysis from industry experts.

Charting solutions offered free

within futuresouce.com include 25

studies and indicators, and four

styles of customisable charts (bar,

line, candle and area).

MTPredictor is preparing for the

imminent release of its 4.0

series of programs – End-of-

Day 4.0 and Real-time 4.0.

The software is built upon a

unique treatment of the Elliott

Wave principle which, accord-

ing to MTPredictor, is entirely

geared to the business of trad-

ing. EOD 4.0 is engineered

specifically for position traders,

those taking trades from sev-

eral days to several months,

whereas RT 4.0 is for short-

term and day-traders. Both

programs incorporate modules

and functions to progress from

start to finish in the trading

sequence.

Steve Griffiths, founder and

NEW FUTURESOURCE.COM WEBSITE UNVEILED

“This site is more than just a rede-

sign. It offers users a totally new

experience when compared to the

old site,” said Jonathan Dean,

FutureSource vice president of

marketing and online sales. “Not

only is the look of the site com-

pletely different, but there are so

many new features, functionality

and technology utilised that it’s

really a completely new site from

the ground up.”

NEW MTPREDICTOR SOFTWARE IS IMMINENTQUOTESPEED

INCORPORATES THE

MARKET PROFILE®

QuoteSpeed has added the Chicago Board of Trade’s “The Market Profile®” to its version 4.04 software. The Market Profile®, which carries an addi-tional subscription charge for QuoteSpeed users, is an analytical decision support tool that uses time, price and volume data to reveal pricing patterns.

QuoteSpeed’s new addition

EFMTECH RENAMED

Stockcube PLC, the London-based technical analysis specialists, has rebranded its EFMTECH consultancy service for institutional investors as Stockcube Research.

The service continues to provide long/short strategies on international equities, based on momentum and volume studies.

developer of MTPredictor, says

“the software is designed exclu-

sively to enable traders in

stocks, currencies and com-

modities to achieve trading

excellence. This means being

able to identify, evaluate and

manage low risk/high return

trade set-ups consistently, time

after time.”

MTPredictor’s website is being

completely re-designed to sup-

port the new programs and a

494-page trading course is

available to guide users

through the principles and

routines of the software

programs.

Product News

DRESDNER KLEINWORT WASSERSTEIN

ENHANCES ONLINE FX PORTAL WITH

TRADERMADE CHARTING

Dresdner Kleinwort Wasserstein (DrKW), the investment bank of Dres-dner Bank AG, has introduced real-time charting functionality to its online foreign exchange trading platform, Piranha FX.

As a result, clients who use the portal to trade foreign exchange will benefit from increased functionality and will have the ability to display real-time FX charts ranging from one day to up to five years. The appli-cation, which will be provided by TraderMade International Limited, allows clients to access various studies including moving averages, trendlines and channels, and a wide range of oscillators.

Allan McKenzie, head of business development for FX at DrKW said, “The use of TraderMade charting is a logical extension to our increas-ingly comprehensive FX product suite. This is consistent with our aim of directly empowering our clients with tools and applications relevant to the trading process, via a single portal.”

SHARESCOPE

LAUNCHES REAL TIME

US AND INDEX DATA

Ionic Information Ltd has recently announced a new release of ShareScope Real Time. The new version adds live prices for US exchanges, US & European indices and FX markets to their original UK-focused offering.

Martin Stamp, Ionic Information’s managing director said, “More and more traders are looking to the US markets. The spreads are tighter, liquidity is greater, trading costs are lower and there is no stamp duty to pay.”

In March 1978 the Dow Jones made a major low near 740,which proved to be an important turning point in the histo-ry of the US stock market. At the time, the index was at thestart of a long and explosive bull run that would carry pricesto 11,750 in the next 22 years. Figure 1 shows the finalsubdivision in the Dow Jones from November 1977 toMarch 1978, which is an ending diagonal [1,2,3,4,5].

Figure 1.

Ending diagonals are a type of wedge pattern characterisedby two converging trendlines containing five waves. This isa perfect ending diagonal because each of the five wavestraces out an a-b-c pattern, a requirement by definition. Thefact that the ending diagonal occurs in the fifth wave of thedecline is confirmation that the downtrend is ending and areversal is imminent. At the time, Bob Prechter of ElliottWave International who first identified the pattern said"...the 740 area marks the point at which the 1977-1978correction, in term of Dow points, is exactly 0.618 times thelength of the entire bull market rise from 1974 to 1976."

On a longer-term figure, the decline from April 1977 toMarch 1978 was a five-wave sequence [1,2,3,4,5] of whichwave 5 was the ending diagonal. The entire decline lastedten months. Although not visible on the chart, this five-wavesequence was wave C of an A-B-C zigzag, which is a cor-

rective wave retracing a portion of the previous advancefrom 1974 to 1976.

Today we have a similar pattern unfolding on the S&P 500,but an inverted one (Figure 2). This time the final subdivi-sion occurs at the top of the market. As in 1977-1978, therally from the March lows is in five waves [1,2,3,4,5], ofwhich wave 5 is an ending diagonal [1,2,3,4,5]. Similarly,the current five-wave sequence has lasted nearly tenmonths. There is a striking similarity between Figure 2 andFigure 1. Although both patterns are inverted, their struc-ture is identical in that each wave inside the ending diago-nal is in an a-b-c formation.

Sometimes the last leg of the ending diagonal ends abovethe upper trendline as shown on Figure 2. This is a com-mon feature of ending diagonals that is referred to as athrow-over. Is this a case of history repeating itself? Weshall see, but what is clear is that the final subdivision in a-b-c appears to be complete, in which case the Decemberhigh marks the top of wave c of 5 of 5 and the end of theending diagonal. For this reason, odds of a trend reversalhave greatly increased.

Figure 2.

Thierry Laduguie is technical strategist at eYield.

Market Views


ENDING DIAGONALS by Thierry Laduguie

Strong gross domestic product growth in the thirdquarter (Q3) dominated US financial headlines formany weeks. "The economy is cooking." "This recov-ery finally has legs." "2004 promises to be a year wheneverybody wins." "The outlook for consumer spendingcould hardly be in doubt." Mind you, this was the ranting ofjust one article that appeared in Business Week in the firstweek of December - November and December includedmany just like it. Indeed, a lot of life for a single economicnumber, but maybe not surprising given how short memo-ries are in the press.

In Q1 of 2002, for example, the preliminary estimate ofGDP was 5.8%, yet fell to 1.4% one quarter later. Likewisein Q3 of 2002, when GDP showed 4%, only to drop back to1.4% in Q4. Facts such as these make it inconvenient todwell on the past. It's far more enticing for some commen-tators to consider the future, and go on with the inevitablediscussion of what it all means to stocks.

Virtually all the people who have these discussions believethe stock market plays dog-on-a-leash with news-as-mas-ter tugging at will on the other end. That belief reflects whatis probably the deepest ongoing financial myth of them all.Still, in the spirit of "what it all means," here's an example ofjust how many questions a single chart can answer:

• Does GDP growth lead the stock market?• Relative to GDP, when is the best time to get into and

out of the stock market?• In the three decades from 1970-2000, which one of

those decades showed the four largest jumps in quarterly GDP growth?

• Did GDP have mostly negative or positive quarters in the two years before the Great Bull Market started in 1982?

• Was quarterly GDP strongly negative or positive just before the stock market crash in October 1987?

• Which growth trend was clearly stronger in the 1990s: the stock market or GDP?

Figure 1 answers all these questions and more. It showsthat in terms of GDP, the best indicators of future econom-ic growth are actually periods of intense decline in the stockmarket. Consequently, the best time to expect such adecline in the stock market is after a large growth spurt inGDP.

Figure 1.

In fact, a study of the figure reveals that the two largestrises of 16.3% in 1978 and 11.6% in 1971 marked reversalsin quarterly growth that eventually led to two of the largestdownward spikes on record in 1975 (-5%) and 1980 (-7.9%). Both growth spikes also failed to lift the stock mar-ket out of a long-term bear market.

Another little-understood fact is the inherent weakness ofthe economic fundamentals during the stock market periodwe label Cycle V (1975-1999) when compared to the lastequivalent cycle, Cycle III (1942-1966). What's more,Cycle V’s relative economic weakness also holds true forthe first 15 quarters following the end of the cycles (labeled"follow through" on Figure 1).

On average, real GDP has grown at a 2.3% annual ratesince the bull market ended in January 2000, far slowerthan the average of 3.7% in the first 15 quarters after thepeak of Cycle III. Other key economic measures - industri-al production, capacity utilisation, unemployment, con-sumer debt and more - reflect an even weaker relative per-

STOCKS, GDP GROWTH, & A 40-YEAR CHART by Anna Troupe


Market Views

formance. From the beginning of 1966 through the thirdquarter of 1969, capacity utilisation fell from 90% to 86%.This time, the decline has been even steeper, from 82.8%to 74.8%. In the aftermath of Cycle III, industrial productionrose 17%. This time it declined 2%. Since the peak in 2000,unemployment has risen by more than half to 6.1%. Overthe same span after Cycle III, it declined to 3.7%. Thiskeeps Cycle V's "comparatively lackluster" track recordintact and intensifies the message of the fundamentals: theeconomy is far sicker than virtually anyone realises.

The persistence of the bullish consensus for stocks hasnow surpassed any prior period in history. For example, thepercentage of bullish respondents to the AmericanAssociation of Individual Investors weekly survey throughthe end of October surpassed 50% for 11 straight weeks totie the record set through the Dow Industrials' all-time high;then this record was smashed by six more 50%+ readingsthrough the first week of December.

On a 25-week basis, the level of bullishness among individ-ual investors has averaged 55.6%, well above the oldrecord of 52%, which was established right after the all-timepeaks in the first quarter of 2000. At 69.7% bullish, the lat-est figure is also the third highest on record. The secondhighest came in late June and the all-time high came onJanuary 6, 2000, a week before the Dow's peak. The

Investors Intelligence weekly survey shows that a majorityof investment advisors have now been bullish for 35straight weeks, which is also an all-time record; this partic-ular survey goes back to 1963.

The stability of bullish opinion reminds us of 1997, whenstock market corrections became virtually nonexistent. Atthat time we observed that a properly functioning chaoticsystem requires periodic disruption. We compared themarket's straight-ahead rhythm to a dangerously regularheart beat and cited research showing that a heart beat thatis "abnormally smooth and free of fractal variation" mayactually signal an impending heart attack. The observationproved accurate, as a financial heart attack hit in 2000-2002. The emergence of bullish expectations that are noweven more "abnormally smooth" than those of the all-timepeak is a sure sign that another, bigger onslaught liesdirectly ahead.

Anna Troupe is editor and writer at Elliott WaveInternational.

Market Views


"A properly functioning chaotic system requires periodic disruption … The market's straight-ahead rhythm may actually signal an

impending [financial] heart attack."

2003 will be known for the strong rebound in worldequities and the weakness of the US dollar, but whilstthe former may be simply a rebound in what shouldprove to be a longer-term bear market, the latter looks likea trend set to continue in 2004.

The main implications of a broadly weaker US Dollar arethat global "hot" funds are chasing a strong currency as wellas a strong stock market. It is here that Japan flashes up onthe radar as a potentially good investment for 2004 and per-haps longer.

My primary reason for being bullish on Japan is that thecountry entered into the bear market way ahead of the restof the globe and so I have been monitoring it for signs thatit would emerge first and establish a sustainable bullishtrend. As a contrary play, I have always liked Japan, thepinnacle of the bear market being when the SARS outbreakcaused many to avoid the country from both an investmentand physical point of view.

The technical set-up for the Nikkei shows how the maindownward trendline was breached in mid 2003 (Figure 1).

TURNING JAPANESE by Paul Rodriguez


Market Views

Figure 1.

This implied an interim change in trend. However, the keyfeature for this chart is the potential inverse head-and-shoulders reversal formation currently developing. Theimplication of this pattern is for a strong bull market forJapanese stocks on a 12-month time frame. If the marketcan maintain support at 9,300 in Q1 2004 then, after a peri-od of consolidation, targets up to 15,000 and possibly18,000 could be on the cards into 2005.

The chart of US dollar/Japanese yen (Figure 2) highlightsthe weakness of the US dollar. As 2003 draws to a close,the dollar continues to be shunned and $1.30 targets areapproaching fast against the euro and $1.80 against ster-ling. As we can see, dollar/yen has been in a very broadrange for the last 5 years and at present is attempting to

test the lower values of that range. Key support is at 101and is the next major target for this currency pair.

Yen gains twinned with the potential for Nikkei strengthmake Japan an attractive option for US investors. It couldalso encourage Japanese investors to pull money out of theUS and bring it home. For now the Nikkei's rally does notseem contingent on a rally in global stocks and I will bemonitoring the Nikkei's performance on a relative basis. Ifthe rally in the Dow Jones is limited to 10,500, as I am cur-rently anticipating, and 4500 for the FTSE, the technicalout-performance by Japan will be an important sign to lookout for in the coming months.

Paul Rodriguez is director of ThinkTrading.com.

Market Views


Figure 2.

The rally in the British pound has lasted for 2½ yearssince the low at 1.3688 in June 2001 when a conflu-ence of cycle lows met and reversed higher. This rallyhas exceeded expectations over the past 6 months, withthe strongest rally during the entire period being the moveup from 1.6570 to the current levels above 1.8000 in thelast 4 months. The question to be asked, therefore, is howmuch further this can go?

Figure 1 displays the current daily picture with cycles andElliott Wave count. The dominating influence appears to bethe declining longer term cycles. While these have notdeterred the uptrend, this should occur before long with acycle low due by the first half of March 2004.

We need therefore to establish a target area for the top ofthe current uptrend by looking at extensions of the pastwave structure. From these we can establish possible tar-gets:

If Wave (C) is equal to Wave (A), we arrive at a target of1.8486. If Wave 5 of Wave (C) is equal to 1.382 of the priorstructure, we arrive at a target of 1.8595. If Wave (iii) ofWave 5 is equal to 2.618 of Wave (i) then we arrive at a tar-get of 1.8285. Assuming this as a peak to Wave (iii) then

Wave (iv) should reach the 1.7770-1.7970 area. From thisretracement target we can arrive at target projections forWave (v) of between 1.8420-1.8825. (This would corre-spond with Wave 5).

The summary of these measurements suggests that some-time spanning the end of January we should see a move tothe 1.8425-1.8625 area that should be followed by a

decline into the first half ofMarch. There is a suggestionof a weekly cycle due aroundmid-summer and this maycause a deeper pullback.

While not shown, both week-ly and monthly charts are dis-playing bearish divergencesof price against momentum.In general these should notbe blindly followed but are anindication of a larger slowingof momentum. Breaks of keysupports and indications oftrend reversals should bemonitored.

Assuming the target areaholds, we would look for anyinitial retracement to drop tothe 1.7750-1.7950 area once

again with any penetration below here suggesting a deep-er fall to 1.6950-1.7050. We feel that the March low shouldbe around 1.7750-1.7950 and will be followed by somechoppy gains that should not establish a new high andeventually break down into the summer cycle low aroundthe lower support at 1.6950-1.7050.

Risk to this scenario is for the rally to extend beyond thehigher 1.8825 level. Should this be seen, further targets areat 1.8945 and one as high as 1.9585.

Ian Copsey is a consultant technical analyst. He is author of"Integrated Technical Analysis" (John Wiley & Sons, 1999)and provides daily commentary on the four major curren-cies for FX-Strategy.

OUTLOOK FOR THE BRITISH POUNDby Ian Copsey


Market Views

Figure 1.

It's dangerous for an academic to write on technicalanalysis. You can gain little credibility with practitioners.What can a professor know of the realities of the market-place? You can also lose a lot of credibility with other aca-demics. Isn't technical analysis closer to astrology than anyof the natural - or social - sciences?

My first encounter with technical analysis in the marketplace happened in 1982, when I was in Chicago learninghow to trade futures. I was amazed to find the bookstallscrammed with volumes on chart analysis and wave theory,and even more amazed that this stuff seemed to be the lan-guage of the trading floor. None of these ideas had fea-tured in the finance courses I had taken at university,except as objects of ridicule.

Our understanding of the world has moved on since then.Some of the extravagant claims of technical analysis haveindeed proved unfounded. But the attitude of academicshas also changed significantly, and some key tools of tech-nical analysis have been rigorously scrutinised and are nowalmost respectable as topics of conversation in the com-mon room.

"A correct forecast is due to our ownskill. A wrong forecast is due to someunexpected event which no one couldhave foreseen."

So here is my list of the pros and cons of technical analy-sis. There are five of each. Let's start with the cons.

Con 1. Modern Theory of Finance

Technical analysis appears to conflict with every assump-tion and prediction of the mainstream theory of financedeveloped since the 1950s by Nobel-prizewinning academ-ics like Harry Markowitz, James Tobin and William Sharpe.The Modern Theory of Finance assumes that players infinancial markets are well informed, smart and greedy. Thiscertainly squares with my experiences of CME traders. Asa result of their activities, you shouldn't be able to makemoney for nothing and certainly not by peering at the freelyavailable history of stock prices. So that $100 bill lying on

the floor or flashing on the screen can't really be there - if itwas, and there was no risk attached, someone wouldalready have picked it up. This theory has been hugelyinfluential across all activities in the financial services sec-tor. For the trader, its implication is that price changes can'tbe predictable, else there would be many $100-bill-typetrading opportunities. On the contrary, short-term pricemovements should follow a random walk, with changes inprice induced by unforecastable "news" about the underly-ing asset.

Con 2. Findings of Cognitive Psychology

In the past two decades, psychologists have catalogued anumber of biases which impact on forecasting and decisionmaking in complex environments. Two results in particularsuggest that technical analysts may be fooling themselveswhen they see patterns in prices, and claim success fortrading rules based on these patterns.

The first is the "illusion of control". Human nature abhorsrandomness. Presented with a series of random numbers,or a graph of a random walk, people try very hard to find alogic which is driving the numbers, and can often articulateunderlying rules when in reality none exist. In a tellingexperiment Fenton O'Creevy et. al. (2003) presented anumber of professional traders with a blank computerscreen on which a small dot was moving around. The sub-jects were told that there was a complex rule linking move-ments of the computer mouse and the dot on the screen.They were then asked to move the mouse, try to find therule, and control the dot so that it stayed in the top-rightquadrant of the screen. In reality, the mouse was not wiredup to the computer, and the movements of the dot were ran-dom. However, most subjects - and especially the less suc-cessful traders - happily reported that they were learning arule, and controlling the dot.

This links naturally to the second bias - overconfidence. Allof us have an exaggerated view of how competent we arein many tasks. We're almost all above-average drivers.Similarly, when we make forecasts, we know that there willbe a margin of error. But our subjective perception of thelikely error is generally much smaller than the objectivelymeasured error. Related to this, we are very bad at accept-ing feedback on our own performance. A correct forecast isdue to our own skill. A wrong forecast is due to some unex-pected event which no one could have foreseen.

THE PROS AND CONS OF TECHNICAL ANALYSIS:AN ACADEMIC PERSPECTIVE by Roy Batchelor


Subject Matters

Con 3. Bad Science

It is possible to avoid these biases, and assess technicaltrading systems objectively. But most books on technicalanalysis don't do this.

Suppose we wanted to evaluate a pattern-based rule forpredicting turning points. A good method would be to lookat a large sample of past instances when the patternoccurred, and an equally large number of instances whenthe pattern didn't occur, and then count the number of timesa turning point subsequently materialised. We would windup with a table like this:

where A, B, C and D are the number of occurrences. A suc-cessful rule would have larger values of A and D than wouldbe expected by chance, and there are statistical tests todetermine whether this is the case. Ideally, we would alsocheck whether the results are stable over time and in differ-ent markets, and we of course should look at the net prof-its and risk in these various contingencies.

Most technical analysis books report only events like A,when the pattern occurred and the price subsequentlychanged direction. Usually this is accompanied by chartsof events where the pattern is particularly easy to recogniseand the reversal is suitably dramatic. This may sell gee-whiz books, but it cuts no ice with academics, and it shouldnot cut any ice with the trading community either. I concede

that the alternative - a book full of contingency tables andarcane statistical tests - may not fly off the shelves. But it'sthe only way that we can tell whether our trading systemsreally work, or whether we are the victims of wishful think-ing and overconfidence.

Con 4. Inexplicable Weirdness

When something systematic happens in the world, it's bothhuman nature and good science to ask why. At the far sideof technical analysis there exist a number of techniques -Elliott Waves, Gann numbers and the like - which claim tohave predictive power, but offer no plausible reason why.We can all understand that the dollar might fall if the USruns a persistently large trade deficit and a persistently lowinterest rate policy. With a bit more economic theory we canprobably also understand why the dollar is liable to over-shoot its underlying equilibrium rate when this kind of cor-rection takes place. But why should it move to a level deter-mined by some Fibonacci ratio?

Some defenders of technical analysis would say that thiswill happen because many traders in the market useFibonacci retracement levels to determine their stop lossand limit orders. This idea that technical analysis worksbecause it is "self-fulfilling" received a serious knock fromresearch conducted over a decade ago by Helen Allen andMark Taylor (1992). They surveyed the FX forecasts of agroup of analysts week by week over an eight month peri-od. Every week, about half the analysts recommendedshort positions in the Dollar-Deutschmark rate, and abouthalf recommended long positions, and it was very unusualto find a large majority making the same directional fore-cast. This does not suggest that there is a single technicalview of the market, and indeed the rich variety of technicalindicators make it rather unlikely that this will happen.

Subject Matters


"There remains the suspicion that technical analysis is popular not because

it's successful, but because it's easy."

Con 5. Excessive Simplicity

The other problem that academics have with technical trad-ing rules is that many are very simple - moving averagecrossovers, and momentum, for example. At one level, wefeel these shouldn't work because they violate the efficientmarkets hypothesis outlined above. At a deeper level, weare uneasy that barriers to entry into the technical analysisprofession are so low. Anyone after a cursory reading of abasic text can be an instant expert in technical analysis, orat least use the vocabulary of technical analysis with someconfidence. On the other hand basic competence in theeconometrics of financial markets requires (a) genuinemathematical and statistical understanding, and (b) aboutfour years of university and postgraduate study. This mightjust be sour grapes on the part of the academics, who findafter all that effort they still can't forecast significantly betterthan anyone else. But there remains the suspicion thattechnical analysis is popular not because it's successful,but because it's easy.

That concludes the arguments against technical analysis.Here are my five reasons why we should all - academicsand non-academics alike - take technical analysis serious-ly.

"Combining the results of differentforecasting methods works better thantrying to pick a single superiormethod."

Pro 1. Modern Theory of Finance

Yes, you're right. This was also the first argument againsttechnical analysis. In fact all five Pros are the same as allfive Cons.

Early on in the development of the theory, SanfordGrossman and (Nobel Laureate) Joseph Stiglitz pointed outthat markets are only made efficient by the activities ofagents trading on forecasts. So you need technical (andfundamental) analysts to make the market efficient. To keep

these traders in the market, they must be compensated bysome acceptable return commensurate with the risk theybear and the costs of forecasting and trading. So a fully effi-cient market is, in theory impossible. You should with someeffort be able to make money even in a well functioningmarket. It won't be simple. And it won't be the megabuckspromised by the more extravagant newsletters. But it won'tbe zero.

The Modern Theory of Finance is only a model. Like thefrictionless model of Newtonian mechanics, it makesextreme assumptions, reaches extreme predictions, andwe should really use it only as a starting point for under-standing why the world doesn't quite work the way it pre-dicts. Let's take the assumption that players in the marketsare well informed, smart and greedy. This just doesn'tsquare with the reality of insider trading, amateur internetdealing, and governments that throw taxpayers' money intodefending indefensible exchange rate targets. The actionsof these players will put non-random patterns into prices,and there is no reason why technical tools can't be used toidentify these patterns.

Pro 2. Findings of Cognitive Psychology

Not all of the findings of psychologists are bad news fortechnical analysis. We have seen that people have biasesin the way they interpret noisy data. Daniel Kahnemann(another Nobel prizewinner), Amos Tversky and Paul Slovic(1982) have also shown that even with good data, peopledo not make decisions in a statistically optimal way. Insteadthey use "heuristics" or rules of thumb. Forecasts are oftenmade by simple extrapolation. In some situations peopleare conservative and fail to give enough weight to recentnews. Fundamental analysts, for example, are very slow torevise forecasts of company earnings especially in thewake of bad news. In other situations they place too muchcredence on news. For example, the fact that one retailshare has fallen in price may be taken as bad news for thewhole sector. People also exhibit "herding" behaviour, sowhen others buy, they buy too.

There are good reasons for some of the biases and heuris-tics uncovered by psychologists. Humans are made toevolve rather than to calculate. A degree of overconfidencemakes you more attractive to mates (and also potentialemployers). Going along with the crowd also has socialbenefits (and makes it less likely that you will be fired whenyour investments underperform). Indeed, some


Subject Matters

researchers have even suggested that in a changing andcomplex world the simple rules that people use in everydaydecisions may be at least as successful as well thought outstatistically optimal strategies (Gigerenzer et. al. 1999).

Whatever their merits, the fact that people act in this waymeans that patterns are liable to be injected into the timeseries of prices. This is the traditional argument made bytechnical analysts - that prices reflect market psychology.Nowadays, having lived through the tech stock bubble, it ishard for even the most slavish devotee of the random walkto argue against it. One result of this is that "BehaviouralFinance" - finance theory that starts with more realisticassumptions about how people act - is now an essentialpart of academic finance training.

Pro 3. Better Science

Another result of the new respectability of behaviouralfinance is that mainstream academics have begun to lookat the performance of technical trading rules in a scientificway. Brock, Lakonishok and Schleifer (1992) report excessprofits from simple moving average rules applied to a USstock index. But parallel studies in currency markets suchas Lee and Mathur (1994), and Olson (2003) have failed tofind consistently profitable moving average trading rules incurrency markets.

The validity of trading on breaks through support and resist-ance levels has been tested using filter rules for stocks bySweeney (1988), and for FX rates by Sweeney (1986) andLevich and Thomas (1993). They find that small filters(0.5%-1%) yield profits significantly greater than would beexpected by chance. Curcio, Goodhart, Guillaume andPayne (1997) show that FX trading on breaks through sup-port and resistance levels published on Reuters screens

can also be profitable, though their results vary from year toyear.

Academic studies of turning point formations are morerecent, with researchers using data mining methods toidentify and test the impact of turning point patterns.Examples are the studies of the head-and-shoulders pat-tern in stock prices and exchange rates by Carol Osler(1998) and Chang and Osler (1999), and the variety ofreversal patterns identified in stock prices by Lo, Mamayskyand Wang (2000). These studies show that specific pat-terns identified by chart analysts do occur around turningpoints more often than would be expected by chance. Butwhether this can be translated into profits is debatable. Loet. al. (2000) only note that the distribution of stock pricemovements changes after classic reversal patterns, but donot explore whether this could be used to trigger profitabletrades. Chang and Osler (1999) find little evidence that thehead-and-shoulders pattern can be exploited to make prof-its in major currency markets.

The evidence on the value of technical analysis is ambigu-ous (as is the evidence on more fundamental-basedapproaches to forecasting). But it is now seen as a propersubject for systematic investigation.

Pro 4. Explicable Weirdness

Some progress has also been made in understanding whyspecific patterns come about. For example, Osler (2000,2001) demonstrates that there is a high degree of agree-ment among analysts from different institutions aboutwhere support and resistance will be found, and that trans-actions-level data on exchange rates tend to cluster aroundpublished support and resistance levels.

Subject Matters


“Studies show that specific patterns identified bychart analysis do occur around turning points

more often than would be expected by chance.”

Paul DeGrauwe and Decupere (1992) document the clus-tering of exchange rates close to "round numbers" and sim-ilar barriers can be found in stock markets (Donaldson andKim, 1993). It is now recognised that this reflects more thanjust market psychology. Blake LeBaron (1996) andSzakramy and Mathur (1997) note that exchange ratesoften cluster around the limits of bands defended by centralbanks, as they intervene to offset the effects of market buy-ing or selling. And Osler (2001) shows that retail stop-lossand limit orders, and options exercise prices, often occur atround numbers, so it is rational for traders to expect pricemovements to be arrested at such levels.

All this doesn't exactly rationalise the Elliott Wave. But themore we investigate market microstructure, the more weunderstand why and how short term systematic movementscan be injected into an otherwise random price series.

Pro 5. Appropriate Simplicity

In 1982 Spiros Makridakis and associates (1982) publishedan influential study reporting the results of a forecastingcompetition. Experts in statistical time series forecastingmethods were sent 1001 business, economic and financialtime series, and asked to forecast future values for each.Since this initial study, there have been two more suchexercises - the results were reported in Makridakis andHibon (2000). The findings of the studies have beenremarkably robust and suggest there may be general prin-ciples that can inform any forecasting exercise. The goodnews is that best-practice technical analysis does conformto these principles.

One important rule is the KISS principle - "keep it simple butsophisticated". In the M1 and M3 competitions, very simplemodels such as exponential smoothing outperformed morecomplex models. So simplicity is seen as a virtue in fore-casting. Excessive complexity, which characterises manycurrently fashionable statistical models like vector autore-

gressions and neural networks, should be regarded withsuspicion.

A second important rule is that combining the results of dif-ferent forecasting methods works better than trying to picka single superior method. This follows straight from themodern theory of finance theory, which demonstrates thatthe risk (error variance) of a portfolio of assets (forecasts)must be lower than the average risk of the individual assets(forecast). All technical analysis texts emphasise theimportance of confirmation for trading decisions. It is notenough that a particular reversal pattern be present, forexample. There should be confirmation from another indi-cator such as a volume oscillator.

The academic studies cited above examine single technicalindicators in isolation. Because technical analysts use acombination of methods, these studies cannot really deter-mine whether technical analysis "works". With this in mind,I have been conducting at Cass a programme of researchon the performance of individual technical analysts them-selves rather than synthetic trading rules. Some earlyresults (Batchelor and Kwan, 2003) show that most ana-lysts in our sample do have significant expertise and canbeat the random walk and naïve forecasting rules. In manyfields - for example clinical diagnosis - models of expertdecisions have been shown to outperform the decisions ofthe experts themselves. Interestingly, we have found it veryhard to construct a model to replicate the performance ofour panel of technical analysts, even knowing the indicatorsthey claim to use. Construction of a convincing expert sys-tem for technical analysis remains a challenge for thefuture.

Roy Batchelor is HSBC Professor of Banking andInternational Finance at Cass Business School, City ofLondon. He teaches mainstream finance and economicscourses, and also a popular technical analysis course onthe school's MSc (Investment Management) programme.


Subject Matters

TTA: How is the forthcoming course designed specificallyfor energy market participants?

RW: The course covers various energy specific technicalanalysis issues such as contract rollover in exchange trad-ed energy futures, electronic "access" trading vs. pit tradingand how this affects charting.

TTA: Do different technical analysis techniques apply moreto some energy markets than others?

RW: TA techniques are applicable to all energy markets.However, trade execution is better suited to exchange trad-ed instruments due to their superior liquidity. Inability toexecute orders at price levels seen on charts can result intheoretical "paper" profits turning into realised losses.

TTA: What do you hope to achieve with the course?

RW: I hope to provide delegates with a comprehensiveoverview of classical and mathematical TA, and a solidintroduction to mechanical trading system development.The course will show how TA can aid in the developmentand refinement of risk management analysis and giveattendees a good introduction to behavioural technicalanalysis (technical analysis based on behavioral finance).

TTA: Does energy trading present any particular problemsfor the technical analyst?

RW: Liquidity or the lack thereof in the cash markets is a bigproblem for the technical analyst. For example, technicalanalysts trading the interbank FX markets can usually pin-point their entry/exit levels within 1-2 pips. By contrast,technicians trading illiquid markets like power or emissionscredits have tremendous difficulty assessing entry/exit lev-els on market and/or stop loss orders.

TTA: What do you consider to be the most important devel-opments in technical analysis recently?

RW: I see the most important development being the wide-spread proliferation of user friendly software solutions formechanical trading system development.

TTA: In your view, are there any specific software require-ments with regard to applying technical analysis to theenergy markets?

RW: Real-time datafeeds and the ability to backtestmechanical trading system results.

THE TECHNICAL ANALYST TALKS TO….


Richard L. Weissman

Richard Weissman is a trading consultant and faculty

member at The Oxford Princeton Programme in

Princeton, NJ. His course on the "Fundamentals of

Technical Analysis in the Energy Markets" will take

place in London on February 11th and 12th.

TTA: How has the expansion of energy trading in Europegreatly impacted on the demand for technical analysis?

RW: Increased volatility and a rise in trading volume hassensitised traders to the impact of market psychology ontheir cash and paper transactions. This has led to greateremphasis on technical tools and solutions.

TTA: Are there any indicators or analysis techniques thatapply specifically to the energy markets?

RW: No, I feel that the same indicators work in energy as inother markets - trend following indicators such as movingaverages and mean reversion indicators like BollingerBands are very popular and robust.

TTA: How does analysis of the energy markets in the USdiffer from Europe? Is Europe lagging behind the US in itsapplication of technical analysis to energy trading?

RW: I think that there is more widespread utilisation andacceptance of technical analysis in US energy markets thanin Europe but the knowledge gap is quickly narrowing.

TTA: What role can fundamental analysis of the energymarkets play for a technical analyst?

RW: Fundamentals are always an excellent complement totechnicals. For example, if during a bull market some bull-ish news fails to result in a market rally this could be a warn-ing sign of a change in trend. At the very least I would tight-en stop loss orders on all open long positions and probablywould exit long positions. Also, I would become much moreaggressive in placement of sell trigger levels (e.g. employ-ing shorter-term moving averages than normal).

Richard Weissman is a trading consultant and faculty mem-ber at The Oxford Princeton Programme in Princeton, NJ.Based in Florida, he may be reached [email protected].

Suggested reading:

Weissman, Richard [Forthcoming]. Mechanical TradingSystems: Pairing Trader Psychology With TechnicalAnalysis, John Wiley & Sons.


"… technicians trading illiquid markets likepower or emissions credits have tremendous

difficulty assessing entry/exit levels on market and/or stop loss orders."

Point-and-figure is a charting and trading system in one.As such, it offers significant advantages over other ana-lytical methods. Most notably it provides crystal clear sig-nals and unambiguous trendlines. In this article, we pres-ent an overview of how the technique works and why itshould give the advantages that it does.

Plotting a point-and-figure chart

Point-and-figure charts have only one scaled axis - price.Price moves are plotted onto a graph by dividing (or "dis-cretizing") the move into boxes. Rising prices are plotted asXs and falling prices as Os. Consequently, a chart is pro-duced that consists of columns of Xs (rising/bullish) and Os(falling/bearish), with each new column representing areversal.

A reversal is defined as a change in the direction of pricetrend, where the trend reversal fulfils a clear condition;namely, no new Xs or Os can be added to the current col-umn, but a certain amount of boxes can be filled in theopposite direction. Normally the threshold for the reversal isset at 3 boxes.

Let's say, for example, that the most recent or rightmostpart of our chart looks as follows:

The next price reads 98.00 100.50 94.75 99.00 (open highlow close). We have to check first if a new X can be addedto the last column. As the high (100.50) is above the num-ber of the next highest box (100) of the column we haveindeed to add a new X. Therefore no reversal has occurred

and the chart now looks like this:

If the next quote reads 98.00 98.50 96.75 97.00 (open highlow close) we can not add a new X to the column. So wehave to check if a reversal has occurred by checking if threenew O-boxes can be added in the opposite direction bycomparing the low with the box number that would allow usto add three new boxes, which is 97. Therefore a reversalhas occurred and we have to update the chart as follows:

The basic features of a point-and-figure chart

A point-and-figure chart represents the forces of supply anddemand with either long columns of Xs (rising prices) orlong columns of Os (falling prices). In the case where sup-

CAN YOU AFFORD NOT TO USE POINT-AND-FIGURE?An introduction to point-and-figure by Heinrich Weber and Kermit Zieg


Techniques

Techniques


Figure 1.

A description of the security, the trend reversal number (normally 3 boxes), the scaling and the time horizon. In this examplethe security is the EuroStoxx50 index, we are using a 3-box reversal number and a standard scale, and the chart covers theperiod from the end of 1991 to February 2003.

The box size scale on one or both sides of the chart, giving a number to each box. The space that could be filled with eitheran X or an O is called a box.

Alternating columns of Xs and Os. Individual Xs or Os may be substituted by a month code which indicates the beginning ofa new month. The codes are: 1=January, 2=February, 3=March, 4=April, 5=May, 6=June, 7=July, 8=August, 9= September,A=October, B=November, C=December. (When using intraday charts, which is much less common, other time intervals aremarked, such as hours or days).

If the chart spans several years, yearly vertical delimiters are also included.

Trendlines, either diagonally up at 45 degrees or diagonally down at 45 degrees, are often marked with a '+'.

In Figure 1, all the important elements of a point-and-figure chart are indicated. They include:

ply and demand are in a phase of contesting for suprema-cy - as in accumulations or distributions - we get short alter-nating columns of Xs and Os.

The essence: buy/sell signals

The great feature of point-and-figure is the occurrence ofrepeating, logical and easy to detect patterns that can beused as reliable trading signals.

The basic and most important signals consist of the simplebuy and the simple sell, which are double top or double bot-tom break-outs.

The simple buy signal represents the move towards newhighs, a retracement and a rise above the old high. Suchprice action, which is typical and frequent in securities mar-kets, attracts interest from market participants and oftenleads to short covering. Therefore demand increases andprices continue to rise.

The simple sell is a down move that stops, rebounds andthen turns down again. Because the rebound had no stay-ing power, the stock may be perceived as weak. This oftenleads to market participants with long positions liquidatingall or some of their positions, resulting in an increase in sup-ply. Short sellers will create additional downwards pres-sure. The down move continues to a point where the mar-ket participants feel that the stock is cheap and start buying.


Techniques

Table 1. Drawing trendlines

From the lower right-handcorner of a low O orexposed O at a 45 degreeangle from lower left toupper right when the charthas a rising bottom

From the lower right-handcorner of the lowestexposed O in a wall of twoor more exposed Os at a45 degree angle fromlower left to upper right

From the upper right-handcorner of the highestexposed X in a wall of twoor more exposed Xs at a45 degree angle downfrom upper left to lowerright

From the upper right-handcorner of the highest X ina column of abnormallyhigh Xs, down from upperleft to lower right at a 45degree angle.

Trendline is drawn …

Bullish support

Bullish resistance

Bearish support

Bearish resistance

A bottom formation is created and eventually a simple buysignal occurs.

In Figures 2 and 3, we have illustrated the eight most com-mon buy and sell signals. They are all based on the twosimple signals. B2 is a simple B1 buy with a rising bottom,B3 is a break-out from a triple top instead of a double top,B4 is an ascending triple top, B5 is a spread triple top, andB6, B7, B8 are created by breaking out of a bullish triangle(B6), bullish resistance line (B7) or a bearish resistance line(B8).

Based on the symmetry of point-and-figure, the sell signalsare exactly the opposite of the buy signals. Triple topstransform to triple bottoms, bullish resistance lines to bear-ish supports and so forth.

What is clear, though, from our statistical studies is that theones that matter most are the simple buy and simple sell,as they are the most frequent and are the base of the morecomplex signals. Some authors describe signals of a muchhigher order of complexity. We, however, don't considersignals to be important if they are so infrequent that one

never knows whether they are a singular constellation or apattern that can be used as a general trading rule.

Point-and-figure's powerful advantages

For us, the most powerful advantages of point-and-figure(ones that directly affect our proprietary trading activity) arethe clarity of the charts, the one-parameter optimization,and the money management and risk reduction strategiesthat it offers.

Clarity

The discretization of price results in either filled or emptyboxes - giving a buy signal, a sell signal or nothing - andtrendlines are drawn at 45-degree angles according to pre-cise rules (see Table 1). In a point-and-figure chart there isno room for fuzziness.

The 45-degree convention for drawing trendlines has beenshown to give signals that lead to the highest probability ofsuccessful trading. In contrast, how many times have we

Techniques


Figure 2. Buy signal summary

Figure 3. Sell signal summary

looked at bar chart trendlines and asked ourselves if thetrendline was correctly drawn. An uncertainty which younever have when using point-and-figure.

This can be seen in the following two charts. Figure 4shows a touched trendline on the Nasdaq volatility indexand Figure 5 shows a crossed trendline on LVMH.

Scaling (one dimensional optimization)

We use the 3-box reversal system. This means a reversalis drawn if three boxes in the opposite direction of the cur-rent column can be drawn. Choosing the 3-box reversal isnot unusual, but our intense study of the point-and-figuremethod over the last few decades suggests it is the optimalnumber.

Keeping the reversal number constant leaves us with onlyone parameter to adjust, namely the scale. The scale isalways indicated on the side of the charts and is the methodby which to discretize price action or, in other words, theprocedure for splitting price movement into boxes.

We usually use a quasi-logarithmic scale and we start ouroptimizations for different instruments with a specific initialvalue that is based on our experience. For stocks, forexample, we start the optimization with 2%, which meansthat each box covers 2% of the price move. For example,the box on a price level of 1000 has a size of 20, and on aprice level of 10 has a size of 0.2.

Through optimization, we choose a scale that will generatethe most optimal chart; one that tells the story of the stockand generates the most robust attributes, such as signalsand trendlines. Moreover, one-dimensional optimization isfast and, because there is only one parameter to change, itavoids over-fitting the data.

Money management and risk reduction

Point-and-figure includes a very efficient money manage-ment strategy where positions are increased on subse-quent trading signals of the same type. This pyramidingtechnique allows you to concentrate your bets on where theaction is. We have included one of our risky ones as an


Techniques

Figure 5. The break through the bearish resistance line is clearlyvisible and coincides with a buy signal.

Figure 4. Please note the two instances, marked in blue, wherethe bearish resistance was touched. The second instance isrecent and we are now watching out for whether volatility crossesthat trendline. If it does, we become buyers of volatility throughsome long straddles.

illustration (Figure 6), namely a double up on a short inAmazon - first short at the signal at 52, then increase of theshort position at 50.

Point-and-figure also gives clear instructions with regard towhere to place risk-reducing stops. The protective stop isset - as point-and-figure discipline requires - at the pointwhere the next reversal would occur, namely 55 (indicatedwith a green horizontal line).

As it turned out, the pyramiding during the run-up workedperfectly - increasing the positions at the subsequent buysignals (in grey) at 27, 37, 41, 44 and 52 and then gettingout on the sell at 52.

Heinrich Weber is a partner and risk manager of DeTraCo,an active Eurex member firm.

Kermit Zieg is a Full Professor of Finance and Investmentsat Florida Institute of Technology's National Capital Regioncampus in Alexandria, Virginia.

Techniques


Figure 6.


Techniques

Point-and-figure charting is easy to understand. It takes only a few minutes to become a pro-ficient point-and-figure chartist.

Point-and-figure charts are easy to maintain. You can maintain 50 to 100 charts daily in only20 minutes. You don't need a computer or an expensive subscription-based data service.

Point-and-figure has been rigorously tested. Numerous studies have been conducted overdecades, with the same results - point-and-figure charting is a highly profitable trading tech-nique.

Point-and-figure provides exact decision rules. You know exactly where to enter and exittrades. It also shows you where to place protective stop orders.

Point-and-figure is universal. Once you know the basic rules, you can chart every investmentmedium with the same technique.

Point-and-figure automatically concentrates the bulk of the portfolio in the "winners." Throughan exact pyramiding technique money is automatically accumulated in the trending positions.

Point-and-figure is straightforward to optimize, without risk of over-fitting. Using only a veryrestricted set of parameters to choose the optimal one, over-fitting of data is avoided.Optimization, though, is an advanced topic and needs basic computer knowledge.

Summary of point-and-figure's advantages

Since the end of the bull market in 2000,the markets have become much fairer.That is, some stocks go up and some godown. In 2000, the NASDAQ Composite wasdown 39.3% for the year, but while it was aterrible market for some key sectors, othermarket sectors realised record gains. Forinstance, if you invested in utility stocks orbank stocks, you had a great year. The DWAUtility Sector was up 41.6% and the DWABank Sector was up 22%.

Understanding how to benefit from sectorrotation will separate the good investors fromthe marginal ones. Needless to say, this is ahot topic within the investing world and it'slikely to become more important to theinvestor and advisor as the markets remainchallenging.

Sector bullish percents and the bell curve

Everyone with even an elementary under-standing of statistics recalls the bell curveconcept. In relation to the financial markets,when something you are evaluating is on theleft-hand side of the curve, it is consideredoversold. The right-hand side of the curve isconsidered overbought and the middle of thecurve is normal. At DWA we have applied thebell curve concept to the 40 sector bullishpercent indicators that we follow.

For those not familiar with the point-and-fig-ure methodology*, a bullish percent is con-structed by taking the percent of stocks with-in a particular sector that are on a point-and-figure buy signal. A point-and-figure buy sig-nal occurs when a column of Xs exceeds aprevious column of Xs, telling us that demandis in control of the stock. Within the method-ology, every stock is either on a buy signal ora sell signal. There is no grey area. Let's saythat a sector has 100 stocks in it (typically theminimum we require for a sector) and 50 ofthose stocks are on a point -and-figure buysignal. We would then say the BullishPercent reading for that sector is 50%. A ris-

SECTOR ROTATION IS ALIVE AND WELL By Tom Dorsey


Techniques

Figure 1. Dorsey, Wright Sector Bell Curve


Sector coding in the bell curve:

Red = Unfavoured

Green = Favoured

Yellow = Average

Capital Letters = Sector Bullish Percent Chart in Xs

Lowercase Letters = Sector Bullish Percent Chart in Os

Red Letters = Sector Bullish Percent Chart changed columns that day

AaA

ing column of Xs for a sector bullish percent means moreand more stocks within the sector are being controlled bydemand. A declining column of Os for a sector bullish per-cent means that more and more stocks within the sector arebeing controlled by supply.

Once we have calculated the bullish percent for each sec-tor, we then plot the aggregate of those sectors onto a bellcurve. The bell curve goes from 0% to 100% along the hor-izontal axis. The first four letters of each sector bullish per-cent are then plotted on the curve. This gives us a compos-ite picture of which sectors are oversold, which are over-bought and the general overall risk in the market.Sometimes, as in April 1998 and present, December 2003(Figure 1), the curve will begin to get very skewed to theright-hand side, indicating an overbought market.Sometimes, as in March 2003 (Figure 2) and September1998, the curve will get very skewed to the left-hand side,indicating an oversold market.

The bell curve picture of the sectors performance ofFebruary 2000 is unlike any other we've seen during thepast five years. Notice how there are almost two curves.Some sectors are huddled on the far left-hand side of thecurve while other sectors are huddled on the far right-handside of the curve. Basic statistics says that when what youare measuring goes to either extreme of the curve, it wouldeventually migrate back to the middle of the curve. Wellthat certainly happened with sectors in 2000. In fact, some

things just blew through the middle and ended up movingfrom the far right-side to the far left-side and vice versa.

Identifying favoured sectors

The bell curve does an excellent job in guiding us withrespect to sector rotation but we also have another tool thataids us tremendously in understanding sector rotationcalled "Favored Sector Status." In any type of market, thereare always some sectors that help carry the market andothers which are laggards. We've found the best way todetermine the leaders from the laggards is with the"Favored Sector Status." The idea behind this concept isthat a sector that is favored is likely to outperform the mar-ket, while a sector that is unfavored is likely to underper-form the market. In addition, utilizing the Favored SectorStatus keeps an investor or trader in for the bulk of themove instead of watching the sector continue to head thedirection you thought it would, yet you were too quick to pullthe trigger and take a profit. For a sector to reach FavoredSector Status it must have a positive reading in three of thefollowing four criteria:

1. Sector relative strength:

We use an in-house equal dollar-weighting index of stocksand then take a point-and-figure relative strength readingfor that index. The relative strength reading is calculated by

Techniques



dividing the price of the index by the S&P 500 Equal DollarWeighted Index and then multiplying by 100. Once thatreading is obtained, it is plotted on a point-and-figure chart.Every index has a relative strength chart. The strongest arethose whose relative strength charts are in a column of Xs,as that tells us the sector is outperforming the market as awhole.

2. Percent of stocks on a relative strength buy signal:

The point-and-figure relative strength buy and sell signalsare long-term in nature, lasting an average of two years.Within an industry group, we calculate the percentage ofstocks that comprise the sector whose relative strengthcharts are on buy signals. This percentage is then plottedon a grid from 0% to 100%. When it is in a column of Xsand rising, it means more stocks underlying that sector aregetting stronger versus the market on a long-term basis.

3. Percent of stocks in a column of Xs on their RScharts:

Within each industry group, we calculate the number ofstocks whose individual relative strength charts are outper-forming the market on a shorter-term basis. Though signalson average last two years or more, a change in columns onthe relative strength chart lasts generally six to eightmonths. This percentage is also plotted on a grid from 0%to 100% and when this indicator is in Xs and on a buy sig-nal, it suggests the sector is performing better than the mar-ket on a shorter-term basis.

4. Percent of stocks with a positive trend:

When a stock is trading above the bullish support line, wesay its main trend is positive. The basic tenet here is themore stocks that are in positive trends, the stronger thesector. The percent of stocks with positive trends is plottedon a grid from 0% to 100%. Like the other indicators, wewant this to be in a column of Xs and rising. A sector that is a market leader generally exhibits at leastthree of these four criteria as positive or in a column of Xs.

The Favored Sector Status groups are expected to outper-form the market averages while the Unfavored SectorStatus groups are expected to underperform the market.We also find that in many cases, a sector can remain in afavored or unfavored status for months. That makes thisstrategy appropriate not only for traders but also longer-term investors.

To illustrate our point, let's look at the DWA TelephoneSector. The Telephone Sector had been Unfavored formonths and finally on October 30th 2002, the telephonesector had three of the above outlined indicators move topositive. This changed the sector to Favored Status andsuggested investors should look at initiating positions in thesector. Now, there are a number of different ways that onemight gain exposure to the sector. You could buy individualstocks, Exchange Traded Funds (ETFs), buy calls onstocks, etc. The conclusion from the sector moving to afavored status is that the group should be a leader in mar-ket upmoves and hold up better than the broad market inany corrections. From October 30th 2002 to December19th 2003, the DWA Telephone sector is up 58.11% whilethe S&P 500 is up 22.23%. Those sector which areFavored are denoted in the sector bell curve by being shad-ed green.

With respect to portfolio implementation, the ideal situationfor a sector is to be at, or about, the middle of the bell curveand in Favored Status. These are the sectors which shouldbe concentrated on for new positions or core holdings.Conversely, those sectors which are Unfavored and havepoor field position, carry the most risk in the portfolio. Aninvestor who has the right instruments and tools to imple-ment a sector rotation strategy can successfully navigateany market, even the tough ones like 2000-2003.

Thomas Dorsey is a partner at Dorsey, Wright &Associates.

*Further reading: "Point & Figure Charting", ThomasDorsey, 2nd Edition". www.dorseywright.com.


Techniques

TraderMadePremier Global Technical Analysis Solutions

• Accurate Data

• Quick and Easy to use

• Wide choice of products to suit your exact needs

• Desktop and Thin Client Solutions

• Web Access for analysis on the move

• E-Commerce solutions for your customers

• Full user training and Technical Analysis Education Programmes

• TraderMade Workstation

o Advanced o Intermediate o Standard

• TraderMade Lite

• TraderMadeWeb

• TraderMade Charting

• On-line Charts

• TraderMade Quotes

Contact us today for more

information and a free trial…

Tradermade International Limited

8 Archers Court, 48 Masons Hill, Bromley, Kent. BR2 9J Tel: +44 (0) 20 8313 0992

www.tradermade.com [email protected]

Futures traders in the wheat pit at the Chicago Board ofTrade have been analyzing the interaction of price vs.volume and price vs. open interest for more than 150 years.The "general rule" to interpret the health of a price trend ona futures chart is well known by technicians:

Both volume and open interest shouldincrease as price moves in the

direction of the major price trend

The schematic diagram of the strongest bull and bear mar-ket is shown in Figure 1.

Interest rate traders want to know if this methodology isapplicable to the futures contracts they trade, for example,T-bond futures. The answer is yes. There is, however, anidiosyncracy specific to T-bond futures in a bull market.

A student of open interest, observing T-bond futures sincetheir inception in August of 1977, can make the followingobservation:

T-bond futures are driven by the perception of the potential

short hedgers

This means that the dominant price making force is theportfolio managers of the world that have exposure toincreasing interest rates - and specifically to rates at thelong end of the US dollar yield curve. If this contingent ofpotential users of the futures market perceive that rates willrise, they only have two choices:

1) Sell out long-term fixed instruments (cash bonds) fromtheir portfolio and move shorter in on the yield curve.

2) Hedge long fixed instruments in their portfolio with ashort sale in a future that is highly correlated with their cashmarket holdings. The T-bond futures traded on the ChicagoBoard of Trade is a very quantifiable hedge for numerousfixed instruments.

The result is that in a bear market in US T-bond futures,price moves lower on increasing open interest - as thepotential short hedgers become actual short hedgers.Open interest increasing in a price declining environment isthe classic definition of a healthy bear market, according tothe general rule.

The phenomenon just described is not easy to find in bearmarkets in the traditional agricultural commodity futures.Most bear markets in the physical commodity markets arecharacterised by long liquidation - meaning open interestdeclines along with price. Not so in the T-bond futures. Infact, a long term observer of T-bond open interest can statethat a bear market in the price of T-bond futures should notbe trusted unless it is accompanied by an increase in totalopen interest.

An example of a price downward move accompanied bypoor (not bearish) internal characteristics can be seen inFigure 2. Note the decline in total open interest in conjunc-

INTERPRETATION OF VOLUME AND OPEN INTERESTIN THE US TREASURY BOND FUTURES by Kenneth H. Shaleen


Techniques

Figure 1.

tion with the price decline between September 30 andOctober 16, 2003. The price downtrend in the March T-bond future was not to be trusted. It did not generate thedifference of opinion necessary to move the price consider-ably lower. It takes fuel to sustain a price trend. Without thisfuel, the trend is not likely to continue.

What about a bull market in T-bonds? If open interestdeclines during a price rally, does this condition fit the defi-nition of a poor price rally? In theory, yes. But if the T-bonds are really dominated by the perception of the poten-tial short hedgers, the actual short hedgers covering theirshorts (if they are the dominant force), will produce an openinterest decrease.

The 'smart money', the commercial users of T-bonds, tendto be unhedged if they think interest rates are moving lower.This phenomenon was first dramatically illustrated duringthe world equity market 'crash' back in 1987. T-bond futuresrocketed to the upside in price as the short hedgers cov-ered. After this undeniable short covering rally - price con-tinued up. Another leg up in the multi-year bull market in T-bonds had started.

Addendums:

1. Total volume (all contract months, all venues (electronic

+ open outcry, if any)) is combined and plotted on the chart.This combined information is available on the ChicagoBoard of Trade (CBOT) website.

2. Open interest in CBOT T-bond futures always includesboth electronic and open outcry dealings. All contractmonths are combined - and plotted.

3. A 'year-end' phenomenon exists in most financial futures.Total open interest tends to decline into calendar year endas both institutions and individuals lighten their exposure asthe tax deadline approaches. This can be seen with theopen interest decline on the March 2004 T-bond chart inFigure 2.

4. Looking ahead. If the Federal Reserve in the US beginsto shift toward an increase in interest rates, the smartmoney in T-bond futures will already have reacted by plac-ing short positions in T-bond futures - resulting in pricedown, open interest up.

Kenneth H. Shaleen is the founder and president ofChartwatch, an international research firm to the futuresindustry and author of ‘Volume and Open Interest’ (Probus,1991 & 1996).

Techniques


Figure 2 - Open interest can be regarded as "fuel"

Having been involved in the financial markets for over 17years there is one vital lesson that should never needto be re-taught. Against the backdrop of increasingly com-plex and sophisticated analysis techniques, the simpleapproach to analysis is sometimes all that is needed touncover excellent trading opportunities. Arguably, one ofthe simplest yet most powerful trade set-ups in a trader'sdaily routine should be the simple ABC correction.

Although this is not a new pattern, its profit potential hasbeen largely overlooked. Its usefulness in generating lowrisk/high return trade set-ups may have suffered because ofits very simplicity. Yet this pattern, when properly identified,can uncover potentially ideal trading opportunities bothquickly and decisively.

First, though, what is meant by a simple ABC correction? Inthe MTPredictor software, it is defined as a 3-swing correc-tion against the main trend, in which the third swingexceeds the price extreme of the first swing. (See Figure 1)

The chart shows an example on the US stock Yahoo. Itwas in a long, strong impulsive uptrend. This was followedby a simple 3-swing downward correction, in which the thirdswing exceeded the low of the first corrective swing down.This is a typical, simple ABC correction against the maintrend.

The crucial question is: Why is the identification of this typeof correction so important? The answer is that once the sim-ple ABC correction is complete, the main trend normallyresumes. Looking at the current example a few monthslater, it is clear that once the correction (See Figure 2) wascomplete, Yahoo resumed the main uptrend to reach newhighs.

As such, the end of this simple ABC correction representeda perfect place to enter the market to take advantage as themain trend resumed.

However, the main difficulty in the past has always beenidentifying where the simple ABC correction was most like-ly to end. This is resolved by looking at the price relation-ships of the three individual swings within the ABC correc-tion itself.

The most common relationship is where Wave C (3rdswing) is equal in price move to Wave A (1st swing).Although price clusters need to be assessed (using the keyratios of all the swings), the Wave C = 100% Wave A rela-tionship is by far the most important. Armed with this criti-cal knowledge, the trader can now anticipate the most like-ly price area for Wave C, and therefore the entire ABC cor-rection, to end. The price cluster in which the Wave C =100% Wave A relationship falls is called the typical Wave CWPT (Wave Price Target). (See Figure 3)

Figure 3 shows how this simple ABC correction ended pre-cisely in the typical Wave C WPT support zone. Knowing inadvance the most likely support area for the entire ABC cor-rection to end allows the trader to enter into a position witha very small initial (money) risk. This is the ideal point totake advantage as the main trend resumes, as it offers theprospect of a return well in excess of the initial risk neededto take the trade.

Next month, in the second part of the article, the importanceof the simple ABC correction's position will be covered. Itcan often be identified at the very start of what is frequent-ly the most powerful impulsive swing. In Elliott Wave terms,this equates to identifying the start of a Wave 3-type swing,which can be one of the most profitable and low risk tradesin a complete 5-wave sequence.

Steve Griffiths is the developer of the MTPredictor softwareprogram. www.mtpredictor.com


Techniques

THE SIMPLE ABC CORRECTION by Steve Griffiths

Techniques


Figure 3.

Figure 2.

Figure 1.

Wyckoff is a name gaining celebrity status in the worldof technical analysis as more and more practitionersbecome convinced of this powerful approach to trading. Somuch so, that it is due for prominent inclusion in theInternational Federation of Technical Analysts' (IFTA) forth-coming Body of Knowledge of Technical Analysis.

Yet despite its recent gain in popularity, the Wyckoff Methodis not a newcomer. It has withstood numerous tests sinceit was first put forward by Richard D. Wyckoff in the early1930's and has been the mainstay of a loyal band of tech-nical analysts (including, since 1990, those with a graduatecertificate in technical market analysis from the Golden

Gate University in San Francisco).

The theory behind the Wyckoff Method

The Wyckoff Method is a school of thought in technical mar-ket analysis. Although it is not a mechanical system per se,high reward/low risk opportunities can be routinely and sys-tematically based on what Wyckoff identified as three fun-damental laws:

1. The law of supply and demand - determines the pricedirection. It states that when demand is greater than sup-ply, prices will rise, and when supply is greater thandemand, prices will fall. Here the analyst studies the rela-tionship between supply and demand by monitoring priceand volume over time on a bar chart.

2. The law of cause and effect - provides an insight into theextent of the coming move up or down. The law postulatesthat in order to have an effect you must first have a cause,and that the effect will be in proportion to the cause. Thislaw's operation can be seen working as the force of accu-mulation or distribution within a trading range works itselfout in the subsequent move out of that trading range.Point-and-figure chart counts can be used to measure thiscause and project the extent of its effect.

3. The law of effort vs. results - provides an indication of aforthcoming change in trend. Divergencies between vol-ume and price often signal a change in the direction of theprice trend. The Wyckoff "Optimism vs. Pessimism" indexis an on-balanced-volume type indicator helpful for identify-ing accumulation vs. distribution, thereby gauging effort.

Applying the Wickoff Method to the US stock market

Figure 1 and 2 show the application of the three Wyckofflaws to US stocks during 2002-2003. Figure 1, a bar chart,shows the decline in price during 2001-02. An inverse

A TEST OF WYCKOFF by Henry Pruden


Techniques

Figure 1.

1. The law of supply and demand - used to determine DIRECTION of price

2. The law of cause and effect - used to estimate MAGNITUDE of price movement

3. The law of effort vs. result - used to signal CHANGE in price trend

head-and-shoulders base formed during 2002-2003 andmarked the start of a new bull market during March-June2003. The upward trend reversal defined by the law of sup-ply and demand, exhibited in the lower part of the chart,was presaged by the positive divergences signaled by theOptimism Pessimism (on-balanced-volume) Index. Theseexpressions of positive divergence in late 2002 and early2003 showed the law of effort (volume) versus results(price) in action. Those divergences reveal an exhaustionin supply and the rising dominance of demand or accumu-lation.

The bullish price trend during 2003 was confirmed by thesteeply rising OBV index. Accumulation during the tradingrange thus 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.

How far will this bull market rise? Wyckoff used the law ofcause and effect and the point-and-figure chart to answerthe question of "how far?" Using the inverse head-and-shoulders formation as the base from which to take a meas-urement of the "cause" built during the accumulation phase,

the point-and-figure chart (Figure 2) indicates 72 boxesbetween the right inverse-shoulder and the left inverse-shoulder. Each box has a value of 100 Dow points. Hence,the point-and-figure chart reveals a base of accumulationfor a potential rise of 7,200 points. When added to the lowof 7,200 the price projects upward to 14,400. Hence, theexpectation is for the Dow Industrials to continue to rise to14,400 before the onset of distribution and the commence-ment of the next bear market.

In summary, using the Wyckoff Method of technical analy-sis as the basis for forecasting, US equities are in a bullmarket with a potential to rise to Dow Jones 14,400. Theexpectation is that this powerful bull market in the DowJones Industrial Average will continue through 2004.

Henry (Hank) Pruden, Ph.D. is Nagel T. Miner Professor ofBusiness and Executive Director of the Institute forTechnical Market Analysis at the Golden Gate University,San Francisco, US.Charts courtesy of Wyckoff/Stock Market Iinstitute

Techniques


Figure 2.

This article discusses two formations, partialdeclines and partial rises, and how they canbe used to predict the breakout direction from achart pattern.

1. Partial declines

a) Broadening bottoms

Consider Figure 1, a chart of a broadening bottom. (That'smy term for what's usually called a broadening top, but Ihave found performance differences between tops and bot-toms). In a broadening bottom, prices enter the patternfrom the top (that is, after a downward price trend), but theycan exit in either direction. The pattern sports two trendlinesthat broaden out over time, one trendline slopes up, theother slopes down. The trendlines should touch at leasttwo minor highs on the top and two minor lows on the bot-tom with little white space in between (good side-to-sideprice crossings).

Once you have a valid pattern (a

the simple - the technical analyst...chicago board of trade’s “the market profile®” to its...

Documents