Market Technician No42

web site: www.sta-uk.org The Journal of the STA

1

The terrorist attacks on New York and Washington D.C.struck at the financial heart of the US. The biggest tragedy isobviously the cost in human lives and our sympathy goes outto all those who have been personally affected by thisappalling act of violence. The committee has already sent amessage of condolence to the MTA whose office (and library)were in the World Trade Centre. The administrator, ShelleyLebeck, and some others were in the building at the time butfortunately they managed to get out before it collapsed.

There is nothing the markets dislike more than uncertainty.And the terrorist attacks have generated a whole raft ofuncertainties. There has been the usual ‘flight to quality’although, with the US now essentially on a war footing,investors have been treating the dollar – usually one of thefirst bolt holes in times of trouble – rather circumspectly. Oneof the myriad of questions facing analysts as the markets tryto readjust to such a huge shock is whether this is just areaction to the present crisis or whether it is part of a largertopping-out move by the dollar.

As usual at this time of year, the Education Committee will berunning a series of courses on technical analysis. The first isan introductory course which aims to provide a generalunderstanding of the principles of technical analysis. The

foundation course goes into more detail about the differentmethods of analysis and is intended to prepare students forthe diploma course which starts in January. The nextopportunity to take either the STA’s diploma examination orDITA II will be on Thursday 15th November. Anyone wantingto sit these exams should note that there is a one-dayrevision session on Wednesday 24th October which pastcandidates have found extremely useful.

The Society’s web site is in the process of being given afacelift. It is being updated regularly and is intended to bethe first point of reference for anyone wanting informationabout meetings etc. If you have not looked at it recently, it isworth logging in.

Finally, to coincide with the IFTA Conference scheduled forLondon in October 2002, we would like to produce abumper issue of the Journal so if anyone has any ideas forarticles they would like to contribute, please let me know.

COPY DEADLINE FOR THE NEXT ISSUE31 JANUARY 2002

PUBLICATION OF THE NEXT ISSUEMARCH 2002

October 2001 ISSUE No. 42

MARKET TECHNICIAN

FOR YOUR DIARY

IN THIS ISSUE

Wednesday 14th November 2001 Robin Griffiths, Chief Technical Analyst, HSBC

Wednesday 5th December 2001 Christmas party, Michael Smyrk

N.B. The monthly meetings will take place at the Institute of Marine Engineers, 80 Coleman Street, London EC2 at 6.00 p.m.

Obituary Bernard Jones 2

M. Feeny Book Review 3

D. Watts Bytes and Pieces 3

D. Gramza Chronographics Trading Edge 4

M. Wignall Open Interest, Volume and PriceAnalysis: The Special Case of Markets with Inter-Exchange Mutual Offsetting Systems (MOS) 6

D. Furlonger Modelling Bull Market BehaviourPatterns using Interlaced Stochastics 9

E. Miller Analysing The Bear Trend in Coffee 13

P. Beuttell Great Bear Markets and Some Time Targets for a Major Low 14

A. Collins Using Marabuzo lines in TechnicalAnalysis 16

WHO TO CONTACT ON YOUR COMMITTEE

CHAIRMANAdam Sorab, Deutsche Bank Asset Management, 1 Appold Street, London EC2A 2UU

TREASURERVic Woodhouse. Tel: 020-8810 4500

PROGRAMME ORGANISATIONMark Tennyson d’Eyncourt. Tel: 020-8995 5998 (eves)

LIBRARY AND LIAISONMichael Feeny. Tel: 020-7786 1322The Barbican Library contains our collection. Michael buysnew books for it where appropriate, any suggestions for newbooks should be made to him.

EDUCATIONJohn Cameron. Tel: 01981-510210Clive Hale. Tel: 01628-471911George MacLean. Tel: 020-7312 7000

EXTERNAL RELATIONSAxel Rudolph. Tel: 020-7842 9494

IFTAAnne Whitby. Tel: 020-7636 6533

MARKETINGSimon Warren. Tel: 020-7656 2212Kevan Conlon. Tel: 020-7329 6333Tom Nagle. Tel: 020-7337 3787

MEMBERSHIPSimon Warren. Tel: 020-7656 2212Gerry Celaya. Tel: 020-7730 5316Barry Tarr. Tel: 020-7522 3626

REGIONAL CHAPTERSRobert Newgrosh. Tel: 0161-428 1069Murray Gunn. Tel: 0131-245 7885

SECRETARYMark Tennyson d’Eyncourt. Tel: 020-8995 5998 (eves)

STA JOURNALEditor, Deborah Owen, 108 Barnsbury Road, London N1 0ES

Please keep the articles coming in – the success of the Journaldepends on its authors, and we would like to thank all thosewho have supported us with their high standard of work. Theaim is to make the Journal a valuable showcase for members’research – as well as to inform and entertain readers.

The Society is not responsible for any material published inThe Market Technician and publication of any material orexpression of opinions does not necessarily imply that theSociety agrees with them. The Society is not authorised toconduct investment business and does not provideinvestment advice or recommendations.

Articles are published without responsibility on the part of theSociety, the editor or authors for loss occasioned by anyperson acting or refraining from action as a result of any viewexpressed therein.

Bernard Jones’s career spanned a period of enormous change in the City.Unlike many of his contemporaries who opted for gardening leave at theearliest opportunity, Bernard adapted to the changes (not always withoutcomplaint!) and was an immensely shrewd and perceptive interpreter ofmarket behaviour right up until his death. His contribution to the Society isprobably best summed up by the first issue of the Market Technician inwhich he contributed an article on ‘Masters of the Market’ – the first of aseries about the lives and methods of individuals who have concernedthemselves with economic and business cycles. He was also listed as speakerat one of the monthly meetings on the subject "1929 and after – will historyrepeat?" He would no doubt have enjoyed entering the debate on whetherthe present situation mirrors that of 1929. Bernard gave unstintingly of histime to the Society and even postponed an operation in order to honour hislast speaking engagement to the Society. In recognition of his outstandingcontribution, he was made a fellow of the Society in 1992. He was also afellow of the Royal Statistical Society.

The eldest of four children, Bernard was born in 1938 in Finsbury,London, though his earliest memories were from Nottingham where hespent the war years. After a few years with one of the chartered banks onleaving school, Bernard joined the Stock Exchange in 1959 as a bluebutton with Rowe Swan and went on to become the firm's youngest-everpartner. Bernard received the freedom of the City of London in 1991before retiring from the City in the mid nineties.

His retirement years, spent near Lewes in East Sussex, were devoted totrading, to his book collection on all aspects and types of markets and toentomology. A keen butterfly collector and breeder, Bernard worked hardall his life to protect key species from extinction. He was so thorough aresearcher and analyst on everything that he studied, that he often knewmore than his doctors did about the nature and treatment of his illness.After battling bravely against prostate cancer since 1995 Bernard died inthe early hours of 11th September 2001. The tragic events later that dayled to the suspension of trading in both London and New York and wouldundoubtedly have focused Bernard's keen mind on these new trends asthey developed. He is survived by his wife Vivienne, his son Alex anddaughter Melanie.

* * * *Dear Editor,

It is with great sadness that I recently learnt of the death of Bernard JonesFSTA. I remember him particularly well for his lectures to the Society. Hislecture and analysis of the 1929 crash (using Fibonacci analysis) in theearly 1990's especially impressed those of us who were new to the Societyand about to take the STA examination. More recently in his 8th March2000 lecture, he used his wealth of historical knowledge and experienceto compare the Radio Stocks of the 1929 crash with the TMT stocks oftoday and recognised that a similar fate awaited them. Bernard accuratelyforecasted the forthcoming bear market at the time the Nasdaq wasmaking new highs! (The actual high in the Nasdaq (Cash) was made some16 days later than his talk on 24th March 2000). How many sold all theirTMT stocks and took his advice?

Bernard became a friend and through a series of city lunches imparted agreat deal of wisdom. I particularly well remember his discourse on volumeanalysis (he was not a believer), his tale of late nights in the late 1970s,spent programming his then new Apple Computer with all sorts of newindicators and systems, (something many of us tend to repeat to this day!)and his discussion of his past commodity and stock market trades. Bernardkept a log of all his past trades and reasoning, thus he was able to backtrackover both his successful and unsuccessful trades. He always sought to learnfrom his past errors, so as to avoid repeating the same mistake twice.

I also remember Bernard as scholar of both history and mathematics, andhe was always able to place current market events within in the longer-term social and technical cycles. With common shared interests and a loveof good food and wine, we enjoyed many a good lunch together. Iunfortunately missed his Christmas 1999 gathering that I know manyattended and enjoyed.

Other friends tell of his great sense of humour and how he loved a practicaljoke. He was also very found of music and a long-standing Jazz devotee.He will be sadly missed by all those in the Society who were privileged tohave met him and who shared in his wisdom and convivial company.

David Watts MSTA.

2 MARKET TECHNICIAN Issue 42 – October 2001

Networking Bernard Jones 1938-2001

Issue 42 – October 2001 MARKET TECHNICIAN 3

“DUMB MONEY: ADVENTURES OF A DAY TRADER”By Joey Anuff and Gary Wolf (2001) London:

Methuen Publishing Ltd.

Every day, millions of people go online, day trading in the biggest casinoin the world: the stockmarket. A few get rich, a lot get burnt, everyonehopes for the big score. Dumb Money is the rollercoaster story of onewho’s been there and back, and emerged – enriched, enlightened andexhausted – to tell the tale.

So we are told, and Business Week says “if you’re thinking of becoming aday trader, start by picking up Dumb Money... a riveting diary of a year oftrading dangerously”.

I like this book: its breezy and often hilarious style is reminiscent of myfavourite book of all time, “The Money Game” by Adam Smith and itscolourful cast of unforgettable characters.

Like The Money Game, this riotous chronicle of trading and investment isbased on real life, and through all the carnage real-life lessons emerge…

Joey Anuff, the day trader in question, kicks off immediately by bootingpatriotism and morals out of the window: he is a day trader, not aninvestor, so “the only reason to own your share of American enterprisetoday is to sell that share to somebody who is just a little bit more eagerthan you to bet on its future”.

Well in fact that is just Keynes’ game of pass the parcel. I would simply addthat the whole art of the thing is to judge the exact extent of ‘that little bit’.

Warren Buffet read no further. Anuff says dump all notions you haveabout any stock, its brands, its fundamentals: if you are a day trader, theyhave nothing to do with you. The chart of a company’s stock “ifextended back far enough, tells a complicated story of its long-termrelationship with its workers, its customers, and its investors. Butremember – you are not a worker or a customer. Unless things go verywrong, you are not an investor, either. Nothing causes day traders moreamusement than evidence of a core group of believers married to thestocks they buy”.

That is the story of the book: You are not an owner. You are not aninvestor. You are only in it for the money. Is he, then, a technical analyst?He describes well the surging ocean of information facing traders, and thesleepness nights that result, the touts, hypesters and doom merchants,and yes, the summer of 1999 when day trading produced its first massmurderer. He contrasts all that with the coolness of technical traders, butto him that term includes value traders who trade on news items andfilter companies by market capitalisation.

He does, however, then contrast the psychology of a breakout upwardsthrough resistance with the view of value investors that the price rise ismadness. So despite the racy style this is a serious book: it has a glossaryof terms, and an extensive index, in which ‘technical traders’ has fourseparate mentions.

I don’t see that he did much actual serious long-term chart analysis,though he certainly read tons of research stuff from the brokers online.

But he clearly uses real-time charts as his window on the world: “Today Ihave six Level II windows set across three extra-large monitors, eachwindow dedicated to its own Nasdaq stock. The windows are about fourinches square, and below each I’ve placed a real-time chart showing stockprices at one-minute intervals. This produces a bizarre psychologicalillusion. Every time a stock ticks up, I have an almost physical convictionthat I knew in advance what is going to happen. Right now, waiting forthe next tick, I can feel the correct answer just at the edge of myconsciousness. I know it, but I can’t quite articulate it.”

That’s all fine and good, and I think we have all felt that to some degreeat some time or another. The problem of course is that the moment youdo articulate it, “it’s too late, because the number has already appearedon the screen”.

From there he reasons that it might be better to let the system paper-tradeonly: “with a few well publicised exceptions, Level II’s capacity to help youget scalped, taken out, whipsawed, and hacked to pieces is breathtaking”.

The whole book is really about the perils of day trading now that instantexecutions are possible. The catch is that with instant executions youhave to make the trading decisions for yourself.

And there are million of others out there in exactly the same predicament.Net net, my personal belief is that what market regulators call ‘systemicrisk’ is greater now than in the past.

This book, fun as it is, shows why.

Michael Feeny (The STA Library has purchased a copy of this book.)

Fibonacci Prediction.

For those that like to project future support and resistance zones, SteveGriffiths has just released his new software that uses the confluence offibonacci relationships to provide trade set-ups. His new softwareprogram "MT Predictor" plots the confluence of fibonacci relationships assupport or resistance lines on the chart. Many examples of the success ofthis method are provided at his web site.

Future updates to the program are promised that will include an ElliottWave analysis module and the addition of a market commentary module.

See: http://www.MTPredictor.com

* * * *

A Working Lunch?

Many technical magazines are going web based, "Technical Analysis ofStocks and Commodities" (TASC) the well known technical magazine,now has a web based service called "Working Money", with a month’sfree trial and online articles for lunchtime reading, it appears aninteresting addition.

Then there is "Traders Advantage" with additional articles and onlinearticles from TASC. In a similar format to "Working Money" it's asubscription-based service at $24.99 for one year.

Then for those who subscribe to Futures magazine it's now possible toreceive the full magazine in PDF format. Such that you can both readthen store all your Futures Magazines on a nice tidy computer disc (don'tforget the backup copy), instead of having piles of the magazines aroundthe house. This clearly looks like a trend; I will then no doubt morn forthose glossy covers and large-scale charts.

Now if I can only persuade "Active Trader" to have an Internet service, Imight be able to subscribe rather than be excluded from getting a copydue to an overseas mailing address.

See: http://www.working-money.comhttp://www.traders.comhttp://www.futuresmag.comhttp://www.activetrader.com

* * * *

On the future path of a Tennis Ball.

Andre Duka, Ph.D has released new software and an interesting web sitebased upon his research into Physics and Technical Analysis. On the sitehe claims that his approach can predict the path of a tennis ball in flightor the future path of a currency or commodity. With extensive articlessuch as "The General Theory of Evolution or Dukascopy" or "Motion andTime" then the mathematics to back up all this analysis, is one web site Iwill be glued to for some time.

Then there is a demo version of his price prediction software and a forumfor discussion. For those with a good maths background or those whodon't mind head scratching, it looks worthy of some effort.

See: http://www.dukascopy.com

* * * *

Place your Risk in a Box.

If your Risk management leaves a lot to be desired, you may wish to lookover RiskBox.com, which is a web based UK company that specialises intrading and risk management for exchange-traded derivatives. Thecompany has launched "RiskTrade", a web-based service to manage riskon derivative portfolios.

Interestingly the site provides market data on over 100,000 futures,options and underlying instruments (equity, forex and interest rates,energy and commodity indices) traded on most international exchanges.Then for derivative clients, such services as portfolio correlation, stop-lossmanagement and risk analysis.

You can trial the service for a 30-day period see http://www.riskbox.comto register.

Book Review Bytes and PiecesDavid Watts


Candle analysis has now been enhanced by a new charting methodcalled ChronoGraphics. This new technique displays candles ofmultiple time periods on the same chart and simplifies the process ofanalyzing how candles from different time periods fit together.

In order to truly appreciate the value of this revolutionary candledisplay, it is first necessary to understand two elements of candleanalysis that are brought together in ChronoGraphics – CandleAddition and Candle Development.

Candle Addition is the mechanical process of combining the open,close, highest high and lowest low of a series of candles to createone composite candle. Candle Addition sets the stage for CandleDevelopment, the comparison of shorter timeframe candles to thelonger timeframe composite candle they create. Candle Developmentallows the trader to take a look inside the longer term candle toidentify how, when and where the candle was created. Thisinvaluable process provides insights into the buying and selling forcesthat created the longer term candle, its strength or weakness, or eventhe beginning phases of a changing market. Figure 1 illustrates thesignificance of this analysis.

Sections A and B of Figure 1 show two very different, yet viable,creation scenarios of a large white-bodied weekly candle. By usingCandle Addition with both scenarios, the body of the weekly candleis created with Monday’s opening price and Friday’s closing price.Then the highest high and lowest low of the five daily candles createthe top and bottom price of the weekly candle.

In Section A, Candle Development analysis reveals a buyer-dominated trading session where daily white-bodied candles trend upwith higher highs and higher lows. The large white weekly compositecandle opens on its low and closes a little down from its high leavinga small shadow, which indicates possible profit-taking at the end ofthe week. In this scenario, upward continuation would be expectedand long positions would be maintained.

Candle Development analysis of Section B, however, reveals quite adifferent market with a strong seller presence represented by thelarge dark bodies of the daily candles. Each time the market attemptsto rally by gapping higher on the open, sellers drive prices lower,evidence that selling pressure is dominating the daily candles. Thesedaily candles finish the week with a bearish evening star pattern,further indication that the uptrend is losing momentum. In thisscenario, upward continuation would be doubtful. One course ofaction would be to tighten risk management sell stops to protectprofits and prepare for a change in trend.

In these scenarios, two different sets of daily candles created thesame white-bodied weekly candle. An examination of the weeklycandle alone did not tell the whole story. Only upon closerexamination of how and when the weekly candle was formed by thedaily candles, can its strength or weakness be determined.

This comparative analysis of independent daily and weekly candlecharts yields valuable information for the trader, but it can be

cumbersome. Take for example, a comparison of the Microsoft dailyand weekly candle charts in Figures 2 & 3. It is not easy to look atthese charts and quickly identify the relationship between a singleweekly candle and the daily candles that created it. EnterChronoGraphics, created by London-based Nicholas Burnton, theDirector of European technical analysis for Bridge InformationSystems. ChronoGraphics presents an overlay of multiple timeframecandles on the same candle chart.

Figure 4 illustrates a ChronoGraphic weekly candle chart created byusing an overlay of the candle for the second week in January,identified as “A” in Figure 3, and the five daily candles for this weekidentified as “B” in Figure 2. The open, close, high and low prices ofthe ChronoGraphic weekly candle remain the same as the weeklycandle in “A”, however, the ChronoGraphic candle is oriented so itshigh is placed on the highest high of the five daily candles and itslow rests on the lowest low of the five daily candles. This placementcreates a ChronoGraphic weekly candle with a forward slope.

ChronoGraphics Trading EdgeBy Daniel Gramza

Figure 2

Source: Bridge Information Systems

Figure 3


Figure 4

Source: Captial Management, Inc.

Figure 1

Source: Captial Management, Inc.


In ChronoGraphics, the amount of slope of the longer timeframecandle is determined by when the highest high and lowest low weremade during that timeframe and the difference between thesereference prices. If the longer timeframe low occurs before the longertimeframe high, the ChronoGraphics candle slopes forward and istypically considered bullish. If the high occurs before the low, thelonger-term candle slopes backward and is typically consideredbearish. These added visual clues present a picture of when the highand low were made and which occurred first – additional importantmarket information for the trader.

An analysis of the Microsoft stock ChronoGraphics candle chart inFigure 5 “A” illustrates the wealth of information a ChronoGraphicschart holds. The weekly ChronoGraphic candle “A” is a bullishhammer. This pattern for the second week in December indicatesthat buyers are entering the market. The daily data that created theweekly ChronoGraphic candle shows that buyers entered the marketin a big way on Thursday of that week creating the large bullishengulfing line candle “A4”. Follow-through occurred on Friday withanother white bodied candle “A5”. The dramatic entry of buyers onThursday and Friday created the long lower shadow of theChronoGraphic weekly candle hammer pattern.

The following week “B” began with a daily doji candle “B1”. Thispattern indicates a slow down of the buying order flow that enteredthe market the previous Thursday and Friday. Even though week “B”is a dark-bodied candle which indicates that sellers are entering themarket, the small bodies and large shadows of the daily candles“B2”, “B3” and “B4” indicate that sellers did not have the sameenergy going into the end of the year as they demonstrated inprevious weekly candles. This implies a potential last push by sellersand a transition to a bullish trend. Short positions could be coveredand long positions initiated if price moves above Friday’s high. Toconfirm a possible change in trend, it is critical that a long position isprofitable by the end of the week the trade is entered.

The dominant large white-bodied ChronoGraphic weekly and dailycandles for weeks “D” and “E” indicate a healthy uptrend until week“F”. The smaller body and shallow angle of this weekly ChronoGraphiccandle indicate a slowing down of the longer-term trend. After theentry of sellers on Thursday with the dark-bodied candle “F4”, Friday’slarge white-bodied daily candle “F5” shows that buyers made a valianteffort to enter the market on a lower open. However, the small white-bodied candle on Monday “G1” is a bearish corner pattern andindicates a loss of momentum for the Friday buyers. Price movementbelow this candle occurred the next day with “G2”, and a closing pricebelow the reference candle “F5”, which occurred with “H5”, weresignals to exit long positions and initiate a short position.

Strong follow-through for a short position did not occur thefollowing week, as evidenced by the small dark-bodiedChronoGraphic weekly candle “I” and the small-bodied daily candles“I1 through I5”, with shadows at their tops and bottoms. Theformations and price action that created the weekly and daily candlesreveal a market that is struggling as it tries to move to lower prices.This is not a good sign for continuation lower. The candlecharacteristics mentioned indicate that neither the buyers nor sellersare in control. Although a short position would be profitable, sellersshould be cautious and prepare for a possible trend change by exitingtheir position with a close above candle “I2”.

An alternative to the ChronoGraphic display that includes daily pricedata is shown in Figure 6. This is the same weekly chart as Figure 5,but without the daily candle data. This format provides a clear pictureof the ChronoGraphic weekly candle relationships.

The March US Treasury bond futures contract in figure 7 providesanother example of a weekly ChronoGraphic candle chart. Themarket uptrend begins with a ChronoGraphic weekly doji candle“A”. The stage for this weekly candle was set when the daily dojicandle “A2” indicated that selling pressure was slowing down, andthe bullish engulfing line “A3” revealed that buyers were entering atthose price levels. At this point, short positions should be exited anda long-term buying opportunity could be initiated if the price movesabove the weekly doji high. This trade must be profitable by the endof the week the trade was entered. The long position initiated onTuesday “B2” of the following week would remain in place with thehigher close of “B5” on Friday of that week. The size of the whitebodies and the angle of the weekly ChronoGraphic candles for theensuing bullish trend indicate that buyers are in control until week“C”, when, the smaller body and slower angle of the bearishhanging man pattern imply that the trend is slowing down. A closebelow the low of this weekly candle would indicate a longer termchange in trend. Long positions should be closed and short positionscould be initiated. If a short position is entered, it must be profitableby the end of that week. The possibility of this change in trend, afterthe market makes one more attempt at new high prices with weeklycandle “E”, occurs three weeks later with the large dark-bodiedweekly ChronoGraphic candle “F”. The range, body size and angleof this weekly ChronoGraphic candle, and the consistent selling ofthe daily candles reveal that sellers may be in control and short termpositions should be maintained.

Knowing how, when and where a candle is created is usefulinformation for any trader. Add information on market momentum,market strength, when and how a high and low occurred for a longerterm candle and which occurred first, and you have a powerfultrading tool for trending or sideways markets. ChronoGraphicssimplifies the analysis of candle development by blending severalcandle analysis tools into one visual candle chart display.

Daniel Gramza is president of Gramza Capital Management, Inc.. Heis a trader and teaches market related courses to institutions andprivate traders worldwide.This article is an excerpt from a new book‘The Handbook of Japanese Candle Trading Strategies’ due out laterthis year.

Figure 5


Figure 6


Figure 7



Part 3 MOS Induced Distortions1

From Internal to External Dynamics

The exploration of the internal dynamics of the non-MOS or stand-alone market conducted via Figures 1-4 with Tables 3-4 and of theMOS process via Figures 5-9 with Tables 5-7 has been restricted, forsimplicity, to one exchange – that of the CME2. It has also been foronly one type of visible activity – that of a Chicago based participantcreating transient commitment in the CME and having it transferredto Singapore for subsequent retention.

What follows is an examination of the ability of the MOS activity tointroduce distortions in exchange statistics, as perceived by anoutsider – the analyst. This is done through a detailed examination ofwhat can be termed the external dynamics of the market process.The full scope of a MOS-able exchange’s visible activities are morecomplex than described so far.

The full picture is presented in Table 8 to enable the non-MOS orstand-alone market to be contrasted with that of a MOS market. Theimpact on the generation of the exchanges’ statistics can then beexamined in order to explore how MOS induced distortions mayappear. These, as will be shown, are capable of invalidating thetraditional rules of analysis, as summarised in Table 1 “ConventionalWisdom Regarding Open Interest, Volume And Price Changes”. Forbrevity, only two of the rules will be selected, namely those of Case 1and 2 in order to examine the consequences of the MOS induceddistortions.

Table 8 lists the eight possible sets of activity that can involve theCME and SGX exchanges. They include two cases of conventional orstand-alone modes of market activity and six involving MOS activityinvolving both exchanges. For the record, the MOS activity that wasexplored in detail earlier, via its internal dynamics, is represented byCase 3.

Non-MOS External Dynamics

If each exchange is viewed individually, there are no less than seventypes of external dynamics that contribute to each exchanges’ marketstatistics. For the CME, these are Cases 1 and 3-8 inclusive, and forthe SGX, Cases 2-8 inclusive. Non-MOS or stand-alone exchangeactivity, on which the conventional wisdom of Price, Open Interest

and Volume (POI&V) analysis is based and which was summarised inTable 1, can now be compared with a MOS-able exchange.

The stand-alone exchange’s type of market activity is representedthrough either Case 1 or 2 in Table 8. For simplicity, only Case 1, thatof the CME, is explored, however the same logic also applies to Case 2of the SGX. The CME is shown as operating as a stand-alone entity byvirtue of the initiation and retention of positions being limited to localactivity. Thus each and every transaction has an impact on theformation of the auction and a “footprint” is captured within the intra-day time and sales record. This is in terms of the price levels,[Price/Time], at which the transactions that took place, and the totalnumber of contracts traded i.e. volume [∆ Vol.]. In addition the openinterest [∆ OI] figure published at the end of the trading day will reflectwhether the transactions contributed a net reduction or increase in thenumber of open contracts in circulation. This information is containedin the column entitled “Contribution To CME Statistics” in Table 8.

MOS External Dynamics

The MOS situation can now be explored. Again, for simplicity, theCME is used, but the discussion that follows is equally applicable tothe SGX. If the remaining entries in the “Contribution to CMEStatistics” column are examined it is apparent that there are twotypes of Time/Price, ∆ Vol., ∆ OI profiles. One involves contributionsto all three variables. These are Cases 4, 6 and 8. The other involvescontributions to only two of the variables, namely, Price/Time andVolume, with no contribution to Open Interest. These arerepresented by Cases 3, 5 and 7.

The first type of profile, represented by Cases 4, 6 and 8 is significantbecause the Time/Price, ∆ Vol., ∆ OI profiles are the same as that ofCase 1. It is thus apparent that MOS activity is, under certaincircumstances, capable of generating a Time/Price, ∆ Vol., ∆ OI profilethat is indistinguishable from that generated by stand-alone, or non-MOS market activity. Or for that matter, a MOS-able market in whichno MOS transactions are occurring. These three MOS modes involve:(a) positions being initiated in the CME by SGX members and beingretained in the CME – Case 4, (b) positions being initiated in the SGXby SGX members and being transferred to the CME for retention –Case 6, and (c) positions being initiated in the SGX by CME membersbut being transferred back to the CME for retention – Case 8.

A fundamental, but subtle assumption underlying conventionalapproaches to POI&V analysis is that the inter-relationships of the

Open Interest, Volume and Price Analysis: The Special Case ofMarkets with Inter-Exchange Mutual Offsetting Systems (MOS)

By Dr Michael Wignall MSTA

CASE#

1 � CME CME � � � � � CME

2 � SGX SGX � � SGX � � �

3 � CME CME � � � � SGX � � �

4 � CME SGX � � � � � CME � �

5 � CME SGX � � � � CME � � �

6 � SGX SGX � � � � � CME � �

7 � SGX CME � � � � SGX � � �

8 � SGX CME � � � � � SGX � �

MARKETTRANSACTION MODE

NORMAL (non – MOS)

MOS

EXCHANGEINCURRINGINITIALACTIVITY

DOMICILE OFPRO- ACTIVEPARTICIPANTS

PRO- ACTIVE PARTICIPANTSACTIONS

CONTRIBUTION TO CME STATISTICS

DOMICILE OFRE-ACTIVEPARTICIPANTS

PRICE/TIME

∆OI

∆Vol.

CONTRIBUTION TOSGX STATISTICS

PRICE/TIME

∆OI

∆Vol.

INITIATE RETAIN TRANSFER TO

SGX CME

Table 8: External Dynamics Of Normal [non-MOS] Versus MOS Activities And Their Impact On Market Statistics


changes in the three sets of variables are a direct reflection of theresolution of the interaction of the activities of a market’sparticipants. These activities are: (a) buying to establish commitment[Flat to Long], or to offset existing commitment [Short to Flat], or (b)selling to establish commitment [Flat to Short], or to offset existingcommitment [Long to Flat]. As such the intra-day auction profile interms of the price/time [tick price] with volume at tick, and the inter-day auction summary profile in terms of the Open, High, Low, Closeprices with total volume and net cumulative open interest, provideuseful clues as to (a) the how of the way things took place, togetherwith (b) the what that has taken place, respectively.

As such the statistics, when compared, day on day, can suggest whatthe market participants are likely to do based on what they havealready done, and thus an intimation can be gleaned as to theconsequential likely impact of their subsequent actions on theformation of future price levels. The rational underpinning this is thatonce a participant holds commitment (s)he is an integral part of thatmarket and thus shifts in price level, which may be eitheradvantageous or disadvantageous, will catalyse subsequentbehaviour. This is reasonable – given a stand-alone market.

MOS Induced Distortions

There is a problem however, in that it can be expected that the rangeof motivations of the participants undertaking the MOS operationsrepresented by Cases 4, 6 and 8 will be of a different nature to thoseparticipating in a stand-alone market. The presence of the MOSfacility enables participants who already hold positions in their“local” market to also hold opposite positions in the same contract inthe “foreign” market. Thus arbitrage in the same delivery month ispossible. This, of course, is impossible in a contract trading in thenormal stand-alone exchange situation, where arbitrage can only beconducted in the same contract by taking opposing positions indifferent delivery months. In the case of arbitrage in a stand-alonemarket situation, the participants’ motivations for interaction with,say, the nearest to delivery month will be influenced by the pricebehaviour of the first deferred delivery month together with the spotmarket. However, in the case of a MOS-able contract, a participantwill, in addition, be capable of being influenced by the pricebehaviour of the nearest to delivery month, the first deferred deliverymonth and the spot market at the “foreign” exchange. Thus noamount of analysis by a speculator of the Time/Price, ∆ Vol., ∆ OIprofile, of a MOS-able contract, at his/her national exchange, as itevolves, is likely to provide anywhere near as reliable information ascould be gleaned from a non-MOS-able contract, which has half thecomplexity in terms of factors capable of influencing participantinteractions with the market being analysed.

Unfortunately a further, and more disturbing complication arises fromthe MOS activity when Cases 3, 5 and 7 are examined. Once again,for simplicity, only the CME situation is explored. Each of these casesis characterised by having no open interest included in theTime/Price, ∆ Vol. profile. This is because the nature of the MOSprocess means that the CME is involved in results in the commitmentthat comes into existence not residing in the exchange. For example,Case 3 represents the situation where a CME member initiatescommitment in the CME but has it transferred to Singapore. In Case5 an SGX member initiates commitment in the CME and has ittransferred back to Singapore, whilst in Case 7 a CME memberinitiates commitment in the SGX and has it retained there. In eachcase, the exchange which incurs the initial activity only retains theTime/Price, ∆ Vol. elements of the full profile, as evidence of theMOS process that has been undertaken. The profiles generated ineach of these three cases could well be dubbed the “ghosts ofdeparted quantities”3. The commitment thus formed during theseprocesses ends up as open interest in the other exchange i.e. theSGX. This is apparent from the full Time/Price, ∆ Vol., ∆ OI profilesthat appear there in each of the Cases 3, 5 and 7.

This outcome has significant implications for conventional POI&Vanalysis. This is because in each of these cases it would be possiblefor CME inter-day statistics to show increases in volume but noexpected changes [positive or negative] in open interest, stemmingnot from the trading activity normally found in a stand-aloneexchange, but from MOS activity of the type represented by Cases 3,

5 and 7. In a stand-alone exchange, increases in volume with nosignificant change in open interest indicates an unusual situation. Thisis that (a) the rate of the arrival of buy and sell orders is increasingand also that (b) the rate at which new commitment is being initiatedis close to being the same as the rate at which pre-existingcommitment is being terminated. This can be equated to thesituation where the open interest is historically high but unchangingover a period of a few days, and volume is behaving similarly,together with rising prices. This would be Case 1 in Table 1 which ina conventional market suggests that any pre-existing rally trend inprices would likely continue. However, this could well be a spuriousjudgement due to the significant MOS activity of Cases 3, 5 and 7,as the statistics would have been artificially inflated by the ghosts ofthe departed quantities.

The opposite type of MOS activity, that of Cases 4, 6 and 8, whichgenerate full Time/Price, ∆ Vol., ∆ OI profiles, would similarly becapable of generating spurious results if conventional wisdom were tobe followed. This is because an increase in open interest accompaniedby an increase in volume, represented by Case 2 of Table 1, could bedue to MOS activity artificially inflating the statistics, rather than the“normal” activity expected in stand-alone markets.

Interestingly, the MOS process contains an anomaly that has abearing on the way it is capable of distorting the statistics on whichconventional POI&V analysis relies. It is that commitment is onlypermitted to be initiated or brought into being, through MOS trades.Once that MOS – produced commitment exists it can only beterminated through normal or non-MOS activity in the exchange inwhich it resides. It cannot be terminated by invoking the MOSprocess a second time, in the opposite direction. This means that thepotential distortions that are induced in the statistics as far as openinterest is concerned will be restricted to artificial increases in levels.MOS activity cannot artificially reduce existing levels. This appears tobe the only consolation that a MOS-able market can make availableto the perceptive analyst.

It should now be apparent that Table 1, representing as it does, theconventional wisdom, can no longer be appropriate in all marketsituations. The changes in open interest and transactional volume thatoccur in markets that have a MOS facility may well have nothing todo with the similar types of changes that occur in non-MOS markets.This will occur when inter-exchange transfers of commitment takeplace as part of the MOS process being invoked by participants.

The Need For Discrimination

The situation regarding which markets are MOS-able and which arenot, is unfortunately not straightforward. The problem is that marketswhich are currently MOS-able may not have possessed the facilitythroughout the whole of their earlier history. For example, thefollowing CME futures markets only became MOS-able withSingapore as follows:

3 month Eurodollar September 1984Euroyen March 1996Euroyen Libor April 1999Japanese Government Bond January 1999

A variation of this problem is that the opposite situation may alsoexist i.e. markets that are not currently MOS-able may well havebeen at some point in their past. CME currency futures contracts area case in point as some of them went through a period when theywere MOS-able, but they subsequently reverted to singleconventional stand-alone markets.

CONCLUSION

It is apparent that MOS-able markets represent a challenge to theperceived wisdom relating to analysis based on futures Open Interest,Volume and Price. The established rules, which have evolved fromscrutinising the traditional stand-alone, non-MOS markets can nolonger be relied upon. It would therefore be prudent for analysts whoincorporate open interest and volume with price changes into theirmodus operandi, in order to provide commentaries on current marketactivity, to ascertain whether their market of interest has a MOSfacility or not, before adopting traditional analytical approaches.


Trading system developers, performance testers and ranking agencieswould likewise be advised to adopt this precautionary approachwhen choosing market histories against which to assess systems thatare known to have open interest, volume and price changes factoredinto their logic. Unfortunately, this group have to face an additionalcomplication, as they typically “look back” over appreciable timeperiods of market history that can stretch to years. It would besensible for them to identify the periods within the historical dataseries when a particular contract was MOS-able and when it wasnot. This would allow separate scrutiny of the two types of periodand thus allow the opportunity for any anomalous theoretical resultsto be discovered and reconciled.

Market globalisation continues, ever onwards, fuelled by the growingfashion for exchanges to combine and/or de-mutualize, so as tofunction as commercial profit-seeking, shareholding operations.Increased inter-exchange competition is leading to joint venturesbetween exchanges across time zones, in order to make their existingfinancial products more accessible. This is in order to increase tradingvolume, and revenues from exchange fees, which ultimately translateinto dividends for shareholders. Inter-linking exchanges via MOS orequivalent systems is an attractive option in this context, and itwould not be surprising if more financial futures contracts becomeMOS-able as time goes by. This would result in progressively lessjustification for analysts/trading system developers to uncriticallyapply standard rules to current markets, simply on the basis that theyare “well established”, or “tried and trusted” in the analysiscommunity. The outcome could prove embarrassing.

Notes (Part 3)

1. For Parts 1 and 2 see the Market Technician, Nov. 2000, IssueNo. 39 and April 2001, Issue No. 40, respectively.

2. For Figures 1-4 and Tables 3-4 see Part 1, and for Figures 5-9and Tables 5-7 see Part 2.

3. Apologies to those who remember Bishop Berkeley’s criticism in1734 regarding shortcomings in the calculus developed byNewton and Leibniz.

Sources And References

Chicago Mercantile Exchange (CME), Interest Rate Marketing Dept.to Wignall, Jan. 24, 2000 (e-mail)

CME, Trade Processing Unit to Wignall, Jan. 24, Feb. 11, 22, March01, Aug. 28, 29 & 30, 2000 (e-mail)

CME, CME/SIMEX [sic] Mutual Offset System; The World’s MostSuccessful Trading Link, CME, n.d., Chicago, IL.,http://www.cme.com/market/interest/ mos.html

Hadady, R. E., ‘Using open interest to analyse price trends’, Futures,Vol. 16, No. 7, pp. 59-61, 1987.

Kallard, Thomas, Fortune-Building [sic] Commodity Spreads For The 90s, Windsor Books, Brightwaters, NY., 1991.

Miner, Robert, The WD Gann Trading Techniques Home StudyCourse, Vol. 1, Text, Gann/Elliott Educators, Tucson, (Feb.) 1990.

Shaleen, Kenneth H., Volume And Open Interest: Cutting Edge TradingStrategies In The Futures Markets, McGraw-Hill, London, 1991.

Singapore Exchange Derivatives Trading Ltd, (SGX-DT), Singapore Exchange, n.d., Derivatives Tradinghttp://info.sgx.com/SGXWeb_DT.nsf/DOCNAME/SGXDT_MAIN?OpenDocument.

SGX-DT, Contract Specifications of SGX Eurodollar Futures,Singapore Exchange, n.d.,http://info.sgx.com/SGXWeb_DT.nsf/DOCNAME/SGX_Eurodollar_Futures

SGX-DT, Contract Specifications of SGX Euroyen LIBOR Futures,Singapore Exchange, n.d.,http://info.sgx.com/SGXWeb_DT.nsf/DOCNAME/SGX_Euroyen_LIBOR_Futures

SGX-DT, Contract Specifications of SGX, 10-Yr JapaneseGovernment Bond (SGX JGB) Futures, Singapore Exchange, n.d.,http://info.sgx.com/SGXWeb_DT.nsf/DOCNAME/SGX_10Yr_Japanese_Government_Bond_SGX_JGB_Futures

Teweles, Richard J. & Jones, Frank J., The Futures Game: Who Wins,Who Loses, Why, 2nd. ed., McGraw-Hill, New York, NY., 1987

Wignall, M., “Mutual offsetting systems – a problem for technicalanalysis?”, Market Technician, Issue 38, pp. 10-11, (July) 2000

Wignall, M., “Open interest, volume and price analysis: the specialcase of markets with inter-exchange mutual offsetting systems(MOS)”, Part 1, Market Technician, Issue 39, pp. 12-15, (Nov.)2000 & Part 2, , Market Technician, Issue 40, pp. 7-9, (April) 2001

Wignall to CME Currency & Interest Rate Marketing Dept. Jan. 24,2000 (e-mail)

Wignall to CME Interest Rate Marketing Dept. Jan. 25, 2000 (e-mail)

Wignall to CME Trade Processing Unit, Feb. 18, March 01, Aug. 28,29 & 30, 2000 (e-mail)

Acknowledgements: The assistance of various CME and SGX-DTdepartments is gratefully acknowledged. Any errors of interpretationor omission regarding the MOS logic inferred from the dialoguesbetween the exchanges and the author, remain the latter’sresponsibility. URLs valid as at September 03, 2001.

Michael Wignall (who, contrary to conventional wisdom, is seekingto develop the “Holy Grail” of computer based trading systems forexchange traded derivatives markets) welcomes comments on thisarticle, via the Journal’s Letters section or tel. 01923 450681, or e-mail at [email protected]

©Copyright Michael Wignall 2001

Pre-Diploma Courses in TechnicalAnalysis at South Bank University

The Society of Technical Analysts Ltd (STA) EducationCommittee is running a pre-diploma courses in TechnicalAnalysis at the South Bank University.

Technical analysis has become an important part of mostinvestment house activity. The STA is the professional bodyassociated with technical analysis. Their courses are taught onlyat the South Bank University and all teaching is by STAmembers. The courses are considered suitable for the annualPIA Continuous Professional Development Programme.

The following courses are available:

Introductory Course in Technical Analysis30 October – 20 November 2001

The course provides an introduction to technical analysis. Noprevious knowledge is required and there are no formal entryqualifications. Delegates attending the Introductory Course willbe given the option to move onto the Foundation Course aftertwo weeks and at the end of the four weeks. Please note thepenalty charged for such a switch made after the initial twoweeks. The course runs for four Tuesday evenings.

Foundation Course in Technical Analysis30 October – 4 December 2001

The foundation course prepares students for the DiplomaCourse in technical analysis starting in January. An examinationis offered after the Foundation Course, and will be held inJanuary 2001. The course runs for six Tuesday evenings.

Both courses are held from 6.00pm – 9.00pm in the CastleLecture Theatre within the Abbey Conference Centre at theSouth Bank University’s London Road building (a short walkfrom the Elephant & Castle underground station).

If you would like further information please contact Katie Abberton on 07000 710207.

STA Administrative Services,Dean House, Vernham Dean, Hampshire SP11 0LA.

Tel: 07000 710207 Fax: 07000 701208Email: [email protected]


This paper summarizes the development of a group of modelsspecifically aimed to capture “Buying the Dips” behaviour in theFTSE100 bull market and through this seeks to show the value ofInterlacing Stochastics.

The Existence of Buying the Dips Behaviour inthe FTSE100

During the last prolonged bull market, buying the Dips in theFTSE100 was a recognized, even publicized concept over many years.

Some support for the concept can be seen in Chart 1 (FTSE100 fromOctober 1996 to November 1997) which shows the results of amodel where the entry signals were triggered as soon as the indexclose dipped by more than 150 points from a closing high, and eachexit signal was an arbitrary stop when the high had been recovered.As can be seen this simplistic approach captured five successfulBuying the Dips opportunities, although the inadequacy of it as asystem is shown in October 1997, when, after the entry signal wasgiven, the index continued to fall by several hundred points.

So while the concept of Buying the Dips can be shown to be naive itdoes appear to have been evident at times outside bearish periods.Based on this observation, an investigation was commenced in 1998to see if a Buying the Dips model could be developed that might besuitable to be combined with an appropriate options or futures-basedmoney management trading system.

The Approach Taken for the Development ofthe Model

The key requirements for a Dip Recovery are a Dip which can bemeasured using a range indicator and a Recovery which can bemeasured by a momentum indicator. Thus Stochastics which canincorporate both elements of measurement seemed an appropriatetechnique to form the basis of the model.

Initial testing of Stochastics using daily data did not capture morethan a few of the opportunities. Also the relative infrequency ofthe data, when trying to capture comparatively small percentagemoves in the index, meant the signals (both entry and closing)were often late.

As a result hourly data (10 datum points per day including the openand close) were used to build the following models. However, thedata collected was not in OHLC format and so the absolute High and

Low for a period could not be included in the model. This may notstrictly follow published rules for the use of Stochastics (see John J.Murphy – Technical Analysis of the Futures Markets 1986) but it wasconsidered that for the purposes of this study Stochastic indicatorscould still be used.

The first stage of the models’ development was a series ofoptimization tests undertaken for a range of Stochastic indicatorswith signals being generated by Stochastic K/D crossovers and themovement of Stochastics through levels.

Nevertheless, no single Stochastics on their own were found toproduce a model which was considered to be suitable. So an attemptwas made to enhance their performance.

The first attempted improvement was to introduce combinedindicator criteria. e.g. only allowing a faster Stochastic signal to beaccepted if slower Stochastics or other indicators were at specifiedlevels. These combined models often reduced the number of lossmaking signals but this was usually offset by an even greaterreduction in high gain signals.

Therefore, it was decided to start again with a different approach tosee if a more sensitive Stochastic indicator could be found. The aimbeing to capture additional information and through this betteridentify the initial characteristics of successful price recovery patterns.

The technique applied to achieve this greater sensitivity was to use asa signal the cut of a faster Stochastic’s K or D line through anotherslower Stochastic’s K or D line when they were between specifiedlevels i.e. to interlace two separate Stochastics.

An example of interlaced Stochastics is shown in the chart of ModelA1 which was developed at the end of 1998.The long entry signal isgiven by the cross of the faster Stochastic’s D line over the slowerStochastic’s D line within specified limits. The closing signal is givenby the faster Stochastic’s K line crossing below the slower Stochastic’sK line above another limit.

Comparing the Interlaced Stochastics Modelwith a Single Stochastic

To give an example of how the performance of this InterlacedStochastics Model compares with a single Stochastic’s normal K/Dcrossover signal a Comparative Report for Model A1 is set out below.This summarizes the performance of both Model A1 and a singleStochastic which, in order to be comparable, uses K and D periodsbased upon Model A1’s fast and slow Stochastics.

Modelling Bull Market Behaviour Patterns usingInterlaced Stochastics

By D. J. Furlonger

Chart 1

Model A1


The period of assessment is from 6th June 1997 to 1st August 2001(i.e. 4 years and 2 months) and thus includes data both before andafter the model was developed.

The first impression is that during the comparative period there were60 more trades with the single Stochastic than the InterlacedStochastics. However, only 26 of those extra trades were winningtrades (i.e. where the gain in index points was greater than zero beforecosts) while 34 of the extra trades were loss making. Consequently theInterlaced Stochastics Model has a higher winning ratio at 77% thanthe single Stochastic which has a winning ratio of only 62%.

Just as important, the average 57 point gain of the InterlacedStochastics Model was 78% higher than the average gain for thecomparative single Stochastic. In addition the Interlaced Stochasticsperformed better in terms of the number of consecutive winningtrades, the fewer number of consecutive losing trades and the lowersize of the maximum recorded open trade loss (i.e. a fall below thelevel of a trading entry signal while the trade is open).

This suggests that the additional information captured over the testperiod by the Interlaced Stochastics resulted in an importantenhancement of modelling performance.

Refinements to the Interlaced StochasticsModel

Although the Interlaced Stochastics Model A1 outperformed the singleStochastic it was considered that it still gave too many poor signals.

The starting point to try to improve Model A1 was to go through themodel’s trading history to see if there were any regular bearishfeatures about the FTSE100 at the time of the failed signals. This wasinvestigated using moving averages, momentum indicators andfurther Stochastics.

Then selected indicator conditions were combined into the Model inan attempt to filter out those bearish factors. As previouslyexperienced, each additional condition applied to the model reducedthe number of winning trades as well as the number of losing trades,and this can of course undermine the overall performance of themodel. Therefore, to reduce this problem, it was decided that ratherthan incorporate all conditions in to a single refined model, twoseparate refined models would be developed that complimented each other.

MODEL A2

The first refined model, which was developed by the end of 1999,uses the same Interlaced Stochastics signal as Model A1 but withdifferent level settings and also combines a third Stochastic conditionand an RSI condition. See the chart of Model A2.

The history of this model over the period June 1997 to Aug 2001 isset out in the Summary Results and History of Model A2. The fullhistory is given in order to show not only each trade’s gain but alsoits trading period and maximum recorded Open Trade Loss (OTL), inview of the influence these factors have on trading viability.

Model A2COMPARATIVE REPORT FOR MODEL A1

(performance in FTSE100 points before costs)

Model Comparable SingleTest from 06/06/97 to 01/08/01 A1 Stochastic Model

Total number of trades (all long) 70 130Total number of winning trades 54 80(i.e. where the points gain was greater than zero)Total number of losing trades 16 50Percentage of total trades where points gain greater than zero 77% 62%Average points gain per trade 57.4 32.2Most consecutive winning trades 15 9Most consecutive losing trades 3 6Max open trade loss from opening level in a single trade -683.6001 -840.80(not OHLC data)

SUMMARY RESULTS AND HISTORY OF MODEL A2(performance measurement in FTSE100 points before costs)

Test from 06/06/97 to 01/08/01 (1050 trading days)

Total number of winning trades (i.e. where points gain was greater than zero) 28Average FTSE100 points gain per winning trade 110.69Average length of winning trade (trading days) 3

Total number of losing trades 2Average FTSE100 points loss per losing trade -51.35Average length of losing trade (trading days) 7

Trading History

Index@ Index @ TradingTrade Open date signal Close date signal days open Gain MOTL

1 12/9/97 4869.0 17/9/97 5010.0 3.3 141.0 -33.40

2 28/10/97 4502.6 30/10/97 4805.5 1.4 302.9 0.00

3 28/4/98 5794.3 1/5/98 6010.3 3.6 216.0 -9.70

4 1/6/98 5837.9 4/6/98 5873.6 2.3 35.7 -24.40

5 15/6/98 5715.7 18/6/98 5798.8 2.6 83.1 -16.50

6 24/7/98 5879.6 31/7/98 5890.9 4.4 11.3 -71.50

7 1/9/98 5199.5 8/9/98 5335.4 4.7 135.9 -99.90

8 22/9/98 5023.4 24/9/98 5241.6 2.4 218.2 0.00

9 2/10/98 4750.4 7/10/98 4781.4 2.7 31.0 -101.70

10 10/2/99 5766.3 12/2/98 5942.4 1.9 176.1 0.00

11 28/5/99 6171.0 2/6/99 6248.0 1.3 77.00 0.00

12 30/6/99 6348.8 2/7/99 6484.2 2.3 135.40 -33.40

13 26/7/99 6116.9 28/7/99 6276.3 2.1 159.40 0.00

14 11/8/99 6038.8 13/8/99 6129.0 1.9 90.20 -32.20

15 3/9/99 6223.0 7/9/99 6365.1 1.8 142.10 -6.20

16 15/9/99 6067.7 28/9/99 6060.5 8.4 -7.20 -179.00

17 18/10/99 5889.3 21/10/99 5973.4 3.4 84.10 -85.20

18 7/1/00 6480.0 11/1/00 6570.4 2.1 90.40 -6.30

19 21/1/00 6354.8 28/1/00 6414.8 4.6 60.00 -80.70

20 17/4/00 5994.6 25/4/00 6187.4 3.2 192.80 -39.60

21 11/5/00 6155.3 15/5/00 6244.3 1.6 89.00 0.00

22 22/5/00 6070.1 26/5/00 6173.4 3.7 103.30 -49.40

23 30/6/00 6290.2 4/7/00 6438.9 2.3 148.70 -9.10

24 12/10/00 6144.8 17/10/00 6220.1 3.2 75.30 -80.20

25 23/11/00 6237.3 27/11/00 6374.7 2.3 137.40 0.00

26 12/2/01 6197.7 13/2/01 6228.5 1.7 30.80 -5.50

27 14/3/01 5627.7 28/3/01 5662.2 9.3 34.50 -323.20

28 31/5/01 5785.6 5/6/01 5850.4 2.9 64.80 -11.00

29 15/6/01 5745.6 22/6/01 5650.1 5.5 -95.50 -120.20

30 6/7/01 5479.2 16/7/01 5512.1 5.3 32.90 -88.80

Totals 98.2 2996.6

Notes:

1) All trades are Long and entry and exit points are intra day.2) Gain is measured in index points before costs.3) MOTL is the maximum open trade loss measured from each trade’s entry point based

on the data used for the model which does not include highs and lows.


1) Although as expected there are fewer trading opportunities thanin Model A1, there are still 28 winning trades. 22 of the gainsexceed 50 points and of those 13 exceed 100 points.

2) The percentage of winning trades (i.e. results greater than zero) isthus 93% (28/30).

3) The average win for all trades has improved to approximately 100points and an average 110 points per winning trade.

4) The average winning trades are relatively fast, completing in justover 3 days.

5) The average OTL of winning trades is only -43 with the maximumOTL of a single trade reduced (in comparison to Model A1) to -323.

6) The results of Model A2’s signals since the beginning of 2001have declined. This would support the view that Buying the Dipsbehaviour has diminished in the current bear market.

MODEL A3

The second refined model, see the chart of Model A3, was not fullydeveloped until mid 2000. It is similar to Model A2 but the indicatorrules are set to capture a slightly different Buying the Dips recoverybehaviour.

The history of the model is shown in the Summary Results andHistory of Model A3.

It has only 13 results over the 4 year test period and thus must betreated with more caution but:

1) Of the 13 results there were 10 trades signalled where the gainswere in excess of 50 points and of those 8 had gains in excess of100 points. There were no signals with losing points.

2) Thus the model has 100% winning ratio.

3) The average win is 127 points per trade.

4) The average winning trade period is approximately 5 days.

5) The average OTL is approximately -44 points and the maximumOTL is reduced to -213 points.

6) Although in late April 2000 and again in late March 2001 themaximum OTL exceeded -200, all the signals manage to show point gains and this model’s last recorded trade signal, in April 2001, identified a 213 point rise in under 9 days withno OTL.

An Either/Or Combination Based on the TwoRefined Interlaced Models

The latest stage of the development attempts to see if a furtherenhancement can be achieved using an Either/Or system i.e. theopening purchase is triggered either by model A2 or A3 but with ashared closing signal.

The history of this model is set out in the Summary Results andHistory of Model A4.

Interestingly, the use of the same closing signal for both the A2 andA3 opening signals has resulted in greater volatility. This might makeModel A4 more suitable for combining with an options based moneymanagement trading system.

1) The number of winning trades over the period was 33, with 26 ofthe trades having gains in excess of 100 points and 17 of thosehad gains in excess of 150 points.

2) The percentage of winning trades is 89% (33/37).

3) The average gain for all trades is increased to 126 points, with anaverage winning trade of 168 points.

4) The average winning trade is completed within approximately 6 days.

5) As expected, given the greater volatility of Model A4, the averageOTL of winning trades is higher at -56 points and the maximumOTL on the largest loss making trade was -894 points. However,as referred to before it might be possible using call options tocontrol exposure to these falls whilst benefiting from the highergains of winning trades.

6) Finally, and subject to a large caveat that these assessments are allbefore costs, it is worthy of note that Model A4 achieved a gain of4674 points in only 37 trades and while being in the market lessthan 282 trading days. In comparison a Buy & Hold strategy overthe period (a total of 1050 trading days) only achieved a gain of902 points. In fact even at its peak the Buy & Hold strategy onlyachieved a gain of 2305 points.

Model A3



Total number of winning trades (i.e. where points gain was greater than zero) 13

Average FTSE100 points gain per winning trade 127.17

Average length of winning trade (trading days) 5

Total number of losing trades 0

Average FTSE100 points loss per losing trade N/A

Average length of losing trade (trading days) N/A

Trading History

Index@ Index @ Trading

Trade Open date signal Close date signal days open Gain MOTL

1 24/9/97 5056.3 29/9/97 5206.0 2.6 149.7 -0.2

2 27/10/98 5273.3 28/10/98 5283.7 0.7 10.4 0.0

3 30/9/99 5990.3 7/10/99 6203.8 5.6 213.5 -19.6

4 28/10/99 6132.7 1/11/99 6236.8 1.6 104.1 0.0

5 25/2/00 6136.7 3/3/00 6438.5 4.7 301.8 -59.0

6 28/4/00 6285.4 16/5/00 6318.7 11.5 33.3 -205.2

7 6/7/00 6412.3 12/7/00 6518.5 4.3 106.2 -12.0

8 4/8/00 6356.9 16/8/00 6532.0 8.6 175.1 0.0

9 27/9/00 6270.2 2/10/00 6329.7 3.6 59.5 -54.7

10 8/12/00 6272.6 12/12/00 6342.1 2.1 69.5 0.0

11 26/1/01 6283.2 30/1/01 6309.5 1.6 26.3 -2.6

12 30/3/01 5577.8 11/4/01 5767.8 8.3 190.0 -212.4

13 17/4/01 5761.1 30/4/01 5974.9 8.9 213.8 0.0

Totals 64.1 1653.2

Notes:

1) All trades are Long and entry and exit points are intra day.

2) Gain is measured in index points before costs.

3) MOTL is the maximum open trade loss measured from each trade’s entry point based



CONCLUSIONS

1) Buying the Dips behaviour patterns

This study quantitatively supports the view that Buying the Dipsbehaviour was evident in the FTSE100 from at least 1997 until therecent phases of the bear market.

It also shows that it is possible to produce models that capture asignificant part of that behaviour.

2) Interlaced Stochastics Models

Although this is only one case study it does suggest that whenStochastics are relevant, but in their singular form are found to beinsufficient to forecast market behaviour, modelling performance canbe enhanced through the use of Interlaced Stochastics.

MetaStock charts courtesy of Equis International. Reference: John J. Murphy – Technical Analysis of the Futures Markets 1986. All rights reserved. (03/08/01)



Total number of winning trades (i.e. where points gain was greater than zero) 33Average FTSE100 points gain per winning trade 168.68Average length of winning trade (trading days) 6

Total number of losing trades 4Average FTSE100 points loss per losing trade -223.10Average length of losing trade (trading days) 21

Trading History

Index@ Index @ TradingTrade Open date signal Close date signal days open Gain MOTL

1 12/9/97 4869.0 17/9/97 5010.0 3.3 141.0 -33.42 24/9/97 5056.3 29/9/97 5206.0 2.6 149.7 -0.23 28/10/97 4502.6 4/11/97 4887.2 4.3 384.6 0.04 28/4/98 5794.3 1/5/98 6010.3 3.6 216.0 -9.75 1/6/98 5837.9 9/6/98 6035.4 5.1 197.5 -24.46 15/6/98 5715.7 29/6/98 5888.4 9.4 172.7 -26.07 24/7/98 5879.6 19/8/98 5689.3 17.9 -190.3 -518.98 1/9/98 5199.5 16/9/98 5304.3 10.9 104.8 -199.29 22/9/98 5023.4 24/9/98 5241.6 2.4 218.2 0.010 2/10/98 4750.4 13/10/98 4980.7 6.3 230.3 -131.711 27/10/98 5273.3 28/10/98 5283.7 0.7 10.4 0.012 10/2/99 5766.3 12/2/99 5942.4 1.9 176.1 0.013 28/5/99 6171.0 2/6/99 6302.2 2.2 131.2 0.014 30/6/99 6348.8 6/7/99 6579.5 3.9 230.7 -33.415 26/7/99 6116.9 16/8/99 6237.7 14.8 120.8 -140.716 3/9/99 6223.0 7/9/99 6365.1 1.8 142.1 -6.217 15/9/99 6067.7 7/10/99 6203.8 15.9 136.1 -179.018 18/10/99 5889.3 1/11/99 6236.8 10.2 347.5 -85.219 7/1/00 6480.0 17/1/00 6669.5 6.8 189.5 -6.320 21/1/00 6354.8 18/2/00 6226.6 20 -128.2 -356.021 25/2/00 6136.7 3/3/00 6438.5 4.7 301.8 -59.022 17/4/00 5994.6 26/4/00 6323.9 4.2 329.3 -39.623 28/4/00 6285.4 16/5/00 6318.7 11.5 33.3 -205.224 22/5/00 6070.1 31/5/00 6373.7 5.7 303.6 -49.425 30/6/00 6290.2 12/7/00 6518.5 8.7 228.3 -9.126 4/8/00 6356.9 16/8/00 6532.0 8.6 175.1 0.027 27/9/00 6270.2 2/10/00 6329.7 3.6 59.5 -54.728 12/10/00 6144.8 23/10/00 6279.3 7.5 134.5 -80.229 23/11/00 6237.3 6/12/00 6277.6 8.7 40.3 -151.430 8/12/00 6272.6 12/12/00 6342.1 2.1 69.5 0.031 26/1/01 6283.2 30/1/01 6309.5 1.6 26.3 -2.632 12/2/01 6197.7 28/3/01 5662.2 31.9 -535.5 -893.233 30/3/01 5577.8 11/4/01 5767.8 8.3 190.0 -212.434 17/4/01 5761.1 30/4/01 5974.9 8.9 213.8 0.035 31/5/01 5785.6 6/6/01 5914.7 3.8 129.1 -11.036 15/6/01 5745.6 3/7/01 5707.2 12 -38.4 -209.737 6/7/01 5479.1 16/7/01 5512.1 5.3 32.9 -88.8

Totals 281.1 4674.1

Notes:

1) All trades are Long and entry and exit points are intra day.2) Gain is measured in index points before costs.3) MOTL is the maximum open trade loss measured from each trade’s entry point based


ANY QUERIES

For any queries about joining the Society, attending one of the STA courses on technical analysis or taking the

diploma examination, please contact:

STA Administration Services (Katie Abberton)P.O. Box 2, Skipton BD23 6YH.

Tel: 07000 710207 Fax: 07000 710208 www.sta-uk.org

For information about advertising in the journal, please contact:

Deborah Owen108 Barnsbury Road, London N1 OES.

Tel/Fax: 020-7278 7220

STA DIPLOMA COURSE15 January – 26 March 2002

This course aims to prepare students for the Diploma examinationon Friday 26 April 2002. It runs for 11 Tuesday evenings between6.00pm and 9.00pm and also includes the Revision Day (9th April2001) – including Report Writing, and Diploma Exam (3 hours).

The Course is expected to cover:

1. Bar charts. Gaps. Islands, key reversals. Defining priceobjectives from gaps and patterns on bar chars. Arithmeticversus logarithmic scales.

2. Moving averages – arithmetic, weighted, and exponential.Centred, non-centred and advanced. Single, double andmultiple moving average crossovers. Moving envelopes,including Bollinger Bands.

3. Candle charts and candle patterns.

4. Point and figure charts. Construction, scale, box reversal,objective counting. Advantages and disadvantages comparedto other types of chart.

5. Dow Theory.

6. Chart patterns, e.g. triangles, flags pennants, diamonds,broadening patterns (megaphones), wedges.

7. Reversal patterns and how to identify/anticipate them.Rounding tops and bottoms, head and shoulders, spikes,double/treble/multiple tops and bottoms.

8 Trend. How to draw correct short, medium and long-termtrendlines. Trend channels. Return lines and internaltrendlines. Unconventional but useful trendlines.Acceleration. Speed Lines. Trend characteristics.

9. Consolidation – how and why it occurs. Breakouts and howto recognise them.

10. Corrections: when and how far.

11. Support and resistance. The various chart points and facetsthat can act as such.

12. Basic elements of Gann theory.

13. Basic elements of Elliott wave theory.

14. Fibonacci series, fan lines, arcs and time zones.

15. Cycles. AMplitude, length, phase, harmonicity, synchronicity,left and right translation. Detrending.

16. Relative performance and how to interpret relative strengthcharts.

17. Momentum indicators and oscillators including:Rate of change – Welles Wilder’s RSI – Stochastic (%K & D)Moving Average Convergence Divergence (MACD) &MACD histogramDirectional Movement Indicator – Parabolics – CommodityChannel Index

18. Volume signals and indicators, including On-Balance Volume,Volume Accumulator etc. Open Interest.

19. Breadth indicators.

20. Sentiment indicators and contrary opinion.

21. Market Profile (TM).

22. Investor psychology – Individual and group.


Commodity markets have offered adifferent asset class for funds that aretrying to diversify away from equitymarket related returns over the lastyear.

The coffee, cocoa and sugar futures,traded both at LIFFE and at the CSCEin NY (now part of the NYBOT) arefairly mature and liquid contractswhich present many tradingopportunities. While the contractspecifications are different betweenLIFFE and the CSCE, the broad markettrends are the same in the respectivecommodities.

Last week the Nov01 LIFFE RobustaCoffee contract posted a low below$350 for the first time, while CSCEDec01 Arabica Coffee posted a newlow at 44.60c per pound, just belowthe 45.25 record from April 1975.The market is taking its cue fromfundamental factors such as thebreakdown in the producersagreement to try and curtail exports.But to the average man on the street,it would seem that coffee pricesshould be soaring given the “coffeecafes” that grow overnight along every street in every town inthe UK these days.

While new lows should attract new sellers and old bears shouldbe heartened (as every Gann or trend following trader will tellyou, new lows should be sold on) from our Elliott waveperspective there is a decent chance that the bear trend in coffeeprices could “grind” to a halt over the coming weeks. The CSCE

contract has been in the throes of a bearish wave C correctionsince May 1997 and the current decline appears to be the finalfifth wave of wave 5 of C, trading lower in a channel on thedaily chart. Current channel support comes in around 41.00 cper pound. Many of the wave equality targets between 1 and 5and 3 and 5 have already been met, while extensions from Atend to come in below zero which seems too bearish, even in thismarket.

However, adding wave 1 and 3, andprojecting 61.8% of that from thetop of 4 gives 43.54c as a likelytarget, with the full equalityextension noted at 34.30c. Wewould not be surprised to see somecoffee users and funds either bottomfishing or actually taking somemoney off the table on approachesto these levels, with a long termbasing pattern likely to form over thecoming months. Confirmation ofany turn in the trend can bemeasured using traditional technicaltools, such as trendline resistance.The channel resistance line comes inaround 48.80 c per pound currentlyand would need to be regained totake some of the pressure off thedownside. Additional resistancewould be expected to be found atthe top of wave 4 at 58.50c on asustained upmove.

Elizabeth Miller and Gerry Celaya are Head of Research and

Chief Strategist of Redtower Research. Charts courtesy of CQG

Analysing the Bear Trend in CoffeeBy E. Miller and G. Celaya


If you do not believe that the last two years in stock markets haveparallels with the late 1920’s in the US, the Nifty Fifty approach ofthe early 1970’s, the panic and greed which preceded the peak inGold in early 1980, or the Bubble phase in Japan in the late 1980’s,then this article will be of little interest. Its historic analysis is basedon the presumption that the technology mania of 1999/2000 willhave repercussions for global stock markets much like these earlierperiods, when investors and speculators made the same mistake ofchasing a fashionable trend, and lost sight of value. Just as marketsovershot on the upside in 2000, there is now a risk that they willovershoot on the downside. This piece of analysis was drafted beforethe tragic events of 11th September, but they can only have servedto exacerbate the situation.

There have been few Great Bear Markets in the last 75 years if youexclude “emerging” indices. The four mentioned above are worthexamining because of both the lengthy advances and the investorpsychology which preceded their tops. Whilst four is not a greatstatistical sample, the contention of this analysis is that these beartrends all lasted a fairly similar amount of time, and this may beuseful in helping pin-point the eventual low of the current trend. Theproposition is not so much that this will turn out to be a Great BearMarket, or that a low will occur in a given time zone, but that ifmarkets are still dropping at a particular point in time, investorsshould be aware of historical precedent and be looking for indicationsthat a low may be occurring.

The table below sets out the extent and duration of some of the20th Century’s worst bear markets. In fact two of them (Gold andJapan) are arguably secular, and still in progress. This table includesonly their initial phase.

The US bear market of 1973/74 is included because it needs dealingwith before it can be excluded. It is not one of the Great BearMarkets. It did not see losses greater than 50%, or last the 900-1,000 days of the others. (Similarly the UK lost “only” 50% in theearly 1930’s).

On this occasion, the 1973/74 US experience is interesting though.Figure 1 shows it with 2000/01 superimposed. The similarities maywell be spurious as the 1929 comparisons were after the 1987 Crash,but this chart is worth keeping an eye on all the same.

It is worth pointing out that all five of the declines in the tableapproached one of these Fibonacci-related decimals:.50/.618/.7236/.854. Coincidence?

If the time periods for Great Bear Markets are added to the date ofthe 2000 high in the S&P 500, the time window which results is:

Great Bear Markets and Some Time Targets for aMajor Low

By Peter Beuttell

Index Closing Closing % High Low Duration Peak Low Decline Date Date in Days

1. S&P 500 * 31.83 4.41 86 6.9.29 8.7.32 1,0352. S&P 500 120.71 62.28 48 5.1.73 3.10.74 6633. UK All-Share 228.18 61.92 73 1.5.72 13.12.74 9574. Gold 835 296 65 1.8.80 21.6.82 8925. Topix 2,885 1,102 62 18.12.89 18.8.92 973

* Week ending figures/dates

Figure 1

Figure 2

Figure 3

Figure 4


24th March 2000 + 663 days = 16th January 2002 (included for interest)

+ 892 days = 2nd September 2002 + 957 days = 6th November 2002 + 973 days = 22nd November 2002 + 1,035 days = 23rd January 2003

Clearly Autumn 2002 will be the period to watch.

Two techniques provide some support for this time zone as apotential target.

The first is the “Cycle of Pi” phenomenon introduced to us by anMTS Research client in the mid-1990’s. Expressed mathematically, Piequals 3.1415 to four decimals places. The Cycle of Pi lasts 3,141days, with important market peaks occuring at intervals of 1/4 of thecycle. The last cycle began with the 1st April 1994 low in the US.Figure 6 illustrates how the NYSE Composite’s peaks in 1996, 1998and 2000 occurred within a very few days of these Cycle sub-divisions.

The “Cycle of Pi” is due to bottom on 6th November 2002 – one ofthe dates above. This is almost certainly a coincidence, but we thinkthat this is worth highlighting because of the Cycle’s success inidentifying these previous NYSE peaks. The Cycle was one day out inMay 1996, spot on in July 1998 and two days early in September2000.

Finally, we finish with a technique recently published in TechnicalAnalysis of Stocks and Commodities by Mohab Nabil. If this is aGreat Bear Market, then it is quite possible that the pattern in theUK’s All-Share Index will develop into a Head-and-Shoulders top.Mohab’s technique involves using the cyclic forces which create theHead-and-Shoulders pattern to help pin-point the time periodduring which the top pattern’s target is likely to be met. Very simply,the time zone for target achievement is calculated by measuring thetime taken between the point at which the first shoulder began toform and the top of the Head. That period is then projected from

the top of theHead into thefuture. In thisinstance, theHead is a DoubleTop, so we haveused its mid-point. Theresulting timezone occursaround 1stOctober 2002(figure 7).

Since a similarpattern may beforming in theNYSE Composite,it cannot beignored in thiscontext. Figure 8illustrates, thepattern timetarget suggestedis in the secondhalf of 2003. No

confirmationhere, but it willbe interesting tosee whether theUS produces aJapan-typetrading range,where a secondlow occurs in2003.

This techniquedoes not alwayswork, and is notexact, but giventhat Time is themost difficultaspect oftechnical analysis,it is a welcomeaddition to thetechnicalarmoury, andadds a degree ofconfidence to theforegoing.

SUMMARY AND CONCLUSION:

Events of the past two years, in particular the appearance of atwo-year top pattern in the S&P 500 and the devastation amongst NASDAQ stocks, provide the right background for aGreat Bear Market to occur. Sadly, proof that one is occurring willonly begin to appear once a retracement from the all-time highsexceeds 50%.

Great Bear Markets on average run for about 964 days. If theaverage were to occur on this occasion, it would suggest a lowaround 13th November 2002, but with leeway of about two months either side. Such a wide time window would require astaggered entry strategy, but Fibonacci levels of .618 etc. could beused to refine this. It goes almost without saying that we would love to discover with hindsight that this piece of analysis was a wasteof time!

Copyright 2001 MTS Research Ltd. Graphs courtesy of Datastream

Figure 5

Figure 7

Figure 8

Figure 6


Using Marabuzo lines in Technical AnalysisBy A. Collins

What is a Marabuzo line?

A Marabuzo is the body of a candlestick formation; the differencebetween the open and the close. The Marabuzo line is a line drawnbetween those two points. In other words, a 50% retracement. As acandlestick chart is only a bar chart with the open and close filled in,the Marabuzo line could be drawn using a bar chart just as well, butfor graphic purposes it is easier when applied to candlesticks.

All of you will be familiar with retracements and in particular therelevance of 50%. The major difference here is that the 50%retracement is applied not to the high and low of a movement but tothe open and close of a time period. A horizontal line is drawnbetween the two and is extended to the right (into the future) whereit will act as a support or resistance point. The second difference isthat the support/resistance is applicable to the close of the periodbeing used, not necessarily during the period.

Obviously a 50% line such as this is not drawn on every candlestickbody but is used only on those judged to be greater than the norm.While it is a subjective measurement, I have found using anopen/close difference roughly 50% greater than the averageproduces the best results. For example; Eur/Usd daily charts for 2001give an open/close difference of approx 53 points. Therefore adifference of 75 or greater will be of immediate interest.

There is no hard and fast rule that can be applied here. An evaluationof the chart being studied and the time period being used willdetermine the relevant numbers. Even that result will have to bereappraised to reflect different market conditions. Sideways,consolidatory trading will create less sizeable movements, therefore arelevant Marabuzo will, naturally, be of a smaller size than in volatilemarkets where all candlesticks are larger. Let us be clear; this style ofanalysis, like all Technical Analysis, is not a science but a reasonedapplication of principle.

One thing will become obvious while looking at the followingexamples and by trying these concepts out in practice yourself; thelarger the Marabuzo, the greater the importance that the resultant50% level is likely to have on subsequent movements.

How to use the 50% Marabuzo Line

The primary use of this line is as a support or resistance line. Picturea downward Marabuzo; the close 100 points lower than the open.The Marabuzo line will be drawn 50 points below the open (50points above the close). The high and the low are not irrelevant asthey may help to form a recognisable candle pattern but, initially,they are ignored. As it is likely that the next period’s open will beclose to the previous close, the Marabuzo line is likely to act asinitial resistance. Like all such lines it is only potential resistance.That resistance has to be tested and proved. Once that is done theline is likely to show continued importance. The same principle isapplied to supports.

Figure 1 shows a gradual up move suddenly accelerated with two largeMarabuzos acting as confirmation. The first 50% retracement was

relevant for a very short period but upside exhaustion after the secondled to a period of consolidation that was protected by the second 50%line, which provided a platform for the next upward break.

The second example shows the Marabuzo line as resistance ratherthan supports.

This pair, after breaking through the 55 hr displaced moving average,declined in an almost step-like fashion, with periods of consolidationconstrained on the topside by the 50% lines of the larger Marabuzos.The first two marked lines held without any serious tests. The thirdsurvived a number of powerful attempts until finally, after a doublebottom, the Marabuzo line was broken. It is significant that after thatbreak, the line provided a strong platform for the powerful correctivemove.

These Marabuzo lines are applicable to all time periods, of course,and although the actual number of points required to make arelevant Marabuzo will differ, the ratio will be broadly similar.

Figure 3 is a weekly Cable chart. The first, higher, Marabuzo lineprotected the rally begun three weeks later, capping it tightly for threefurther weeks before the decline continued. After the longer-termbase was created the bounce saw two Marabuzos, the latter of whichprovided the second line, this time acting as support. Three strongtests found that level too tough to crack and lead to an eventualupside extension which then found resistance at the, still relevant, firstMarabuzo line.

CONCLUSION

It can be seen from the above examples, and from putting this simplemethodology into practice yourself, that although using 50%Marabuzo lines provides new support and resistance levels, theapplication of these levels is exactly the same as in those arrived at byother, more traditional means. Like all such levels they must alwaysbe regarded as potential influences until tested and proved. However,I think you will find that regular use of these signals will prove aninvaluable, additional tool and one that will not easily be discarded.

Figure 1

Figure 2

Figure 3

EUR/USD Hourly

Usd/Chf Daily

Cable Weekly

Documents

Market Technician No42