deka investment - the technical analyst · 2016-11-23 · at deka bank, discusses his methodical...

6002

jan

/feb

.technicalanalyst.co.uk wwwThe publication for trading and investment professionals

DEKA INVESTMENT

US Yield Curve 2005 ReviewHow the stock market

rules performed

Trading the trendA strategy for

individual stocksWill inversion

mean recession?

A rigorous approach to market timing

More and more Bloomberg Professional users are running Updata Technical Analyst to unleash the power of their Bloomberg data

BloombergUpdata for

For more details: type API <GO>on your BloombergGo to 3rd Party anddownload the PDF about the Updata system

To arrange a Trial or DemonstrationCall: 020 8874 4747Email: [email protected] Download your trial at:www.updata.co.uk/bloomberg

Some reasons why you should join them

All Trade Marks are the property of their respective owners

Familiar Bloomberg codes for fast ticker entry

Full Windows look and feel with multiple desktops

High speed scans for any technical criteria

Hundreds of TA techniques and tools covered

Draw or add notes, labels, logos and images

Automatic creation of Index Constituent Lists

Create your own portfolio watch lists

Market Maps of any list of instruments

Chart Bloomberg CIX Indices

Best Point and Figure charts in the world

Risk Reward Ratios and pricetargets for any trade

Chart your own Excel data,Metastock and other formats

Schedule Automatic Chart Runs

Optimise and Test any Indicator

Optimised Trailing Stops

High quality printing and page setup

Report Writer for documents and presentations

Sophisticated technical analysisalerts with screen pop ups, SMS messaging or email

© 2006 Clements Biss Economic Publications Limited. All rights reserved. Neither this publication nor any partof it may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic,mechanical, photocopying, recording or otherwise, without the prior permission of Clements Biss EconomicPublications Limited. While the publisher believes that all information contained in this publication was correctat the time of going to press, they cannot accept liability for any errors or omissions that may appear or loss suffered directly or indirectly by any reader as a result of any advertisement, editorial, photographs or othermaterial published in The Technical Analyst. No statement in this publication 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 inrelation to any topic discussed.

CONTENTS 1 FEATURES

InterviewGuido Stiel, research and portfolio manager

at Deka Bank, discusses his methodicalapproach to portfolio management.

Trend Following StrategiesUS hedge fund, Blackstar, looks at the

effectiveness of a very simple trend following strategy for individual stocks.

Review of 2005The stock market “rules”

We look at how the various rules performed last year.

JAN/FEB

>15

>18

> 29

>

> >

WELCOMEThe end of a year allows us the opportunity to assess market forecasts made 12

months ago and to look at how the various market “rules” performed. These includethe January Barometer, Sell in May and the 5-year rule, all of which we have featured

in previous issues of the magazine. In this issue we assess how well these rulesworked for both the US and UK markets. We also look at the US yield curve to see

whether historical data supports the likelihood of a US recession this year.

We hope you enjoy this issue of the magazine.

Matthew Clements, Editor

January/February 2006 THE TECHNICAL ANALYST 1

www.vni.com

The IMSL Numerical Libraries™

The most sophisticated, flexible, scalable and highly accessible mathematical and statistical algorithms available for numerical analysis today. Available in pure C, C# for .NET, Java™, and Fortran.

Professional ServicesThe PV-WAVE Product FamilyVisual Data Analysis (VDA) solutions experts trust for rapidly importing, manipulating, analyzing and visualizing data of any size and complexity. Includes the PV-WAVE development environment for sophisticated visualization, TS-WAVE™ for advanced time series analysis and JWAVE™ for Web-based VDA.

Utilizing the best tools, highly skilled technical experts and over three decades of experience, Visual Numerics collaborates with its customers to develop optimal solutions and achieve the highest possible return on investment.

Advanced Analytics for Powerful Minds

VISUAL NUMERICS INTERNATIONAL, LTD. Soane Point, 6-8 Market Place, Reading, Berkshire, RG1 2EG United Kingdom

®

TEL: +44 (0) 118-925-5910 EMAIL: [email protected]

Editor: Matthew ClementsManaging Editor: Jim BissAdvertising & subscriptions:Louiza Charalambous Marketing: Vanessa GreenEvents: Adam CooleDesign & Production: Paul Simpson

The Technical Analyst is published byClements Biss Economic Publications LtdUnit 201, Panther House,38 Mount Pleasant, London WC1X 0AN

Tel: +44 (0)20 7833 1441Web: www.technicalanalyst.co.ukEmail: [email protected]

SUBSCRIPTIONS

Subscription rates (6 issues) UK: £160 per annumRest of world: £185 per annumElectronic pdf: £49 per annumFor information, please contact: [email protected]

ADVERTISING

For information, please contact:[email protected]

PRODUCTION

Art, design and typesetting by all-Perception Ltd.Printed by The Friary Press

ISSN(1742-8718)

INDUSTRY NEWS

MARKET VIEWS US stocks: Yield curve inversion may confirm bear marketGold: The nature of the bull EURUSD: Possible resumption of a bigger up trend?

TECHNIQUES Does trend following work on stocks?Review of 2005The Ichimoku Indicator Combining Bollinger, Fibonacci and trendlines

INTERVIEWGuido Stiel, Deka Bank

SUBJECT MATTERS Towards the fundamentals of TAJapanese candlesticks & simple trading rules

SOFTWAREBerkeley IQ-Trader

BOOK REVIEWMastering Technical Analysis by John, C. Brooks

COMMITMENTS OF TRADERS REPORTLONG-TERM TECHNICALSEVENTS

04

070910

15182025

29

3339

40

43

444648

CONTENTS 2 REGULARS>

20 09

The Ichimoku Indicator US dollar on the up?


George Soros warned in January thatthe US Fed might overshoot in itsbid to tighten monetary policy,deflating housing prices and produc-ing a recession in 2007. Speaking atthe Singapore Institute ofInternational affairs he said that heexpected the federal funds rate topeak at 4.75% but with a risk of

overshooting leading to a hard land-ing for the US economy. He addedthat if the US housing market con-tinues to cool while rates are risingthen it could turn into a hard land-ing," Soros said. "That's why Iexpect a recession to happen in2007, not 2006.

4 THE TECHNICAL ANALYST January/February 2006

Industry News

INSTANT MESSAGING FROM UPDATA Charting provider Updata hasannounced the introduction of sys-tem writing and testing in a new ver-sion of its leading technical analystsoftware to be launched shortly. Thecompany is also launching a newcharting instant messaging serviceenabling users to share analysis andswap charts in real time with a singlemouse click. David Linton, head ofmarketing at Updata, says, "System

writing is the big move for us in2006 with more and more banks andhedge funds wanting to write theirown indicators and test differentstrategies. The real challenge hasbeen to produce a programming lan-guage that most analysts would beable to use and we believe we havedone that". More information can beobtained from their website at:www.updata.co.uk/bloomberg

Berkeley Futures have launchedBerkeley IQ-Trader, a front-end sys-tem which can be used to develop,backtest, optimize and automatemechanical trading strategies ofvarying complexity. The system,which is connected to their online

brokerage services, requires noknowledge of programming lan-guages and can be operated entirelythrough the PC's mouse. For furtherinformation contact Marc Quinn atBerkeley Futures.www.bfl.co.uk

Berkeley launch IQ-Trader

NYSE Regulation has fined UBSFinancial Services in the US $49.5million for failure to supervisedeceptive market-timing activitiesengaged in by UBS. Almost half thefine will be used to compensateclients who lost money in the trades.Richard Ketchum of the NSYEcommented, "When a brokeragefirm permits a hedge fund to gain anunfair advantage over other

investors, it has violated the trustthat forms the foundation of ourcapital markets," Between January2000 and December 2002, NYSEsay brokers in at least seven UBSbranch offices engaged in deceptivemarket-timing to benefit their hedgefund customers by concealing theiridentities, enabling them to trade inmutual funds that sought to limit orcurtail their market-timing.

UBS fined $50m by NYSE

Leading hedgefund managersees sharp dollardecline in 2006 David Murrin, chief investment offi-cer with Emergent AssetManagement says investors should sell their US assets, especiallythe dollar against the yen, asAmerica's economic and politicalpower is likely to decline this year.Murrin says the dollar could fall to104 within the next 6 months beforeheading towards 90. "The biggestexpression of America's decliningpower and the ascension of Asia willbe seen in dollar/yen." Murrin said.

Charting and order routing softwareprovider, CQG, is now offeringorder execution on the MontrealExchange (Bourse de Montréal).Traders will have the ability to tradefrom a full range of CQG products,including CQGNet, CQGTrader andWebTrader which will use the FIXprotocol to connect to the MontréalExchange's SAM electronic tradingengine.www.cqg.com

CQG CONNECTS TO MONTREALEXCHANGE

Soros sees 2007 recession


Industry News

E X C H A N G E N E W S :

In 2005 the London Stock Exchange attracted a record 129international companies, an increase of 82% on 2004, raisinga total of £5.9 billion in new issues, the highest value since2000. ETF trading at the LSE saw an increase of 108% inDecember on the previous year, with daily value averaging£47.8 million.

Euronext.-liffe traded a total of 606 million futures andoptions contracts in 2005, up 7% on 2004. Interest rate prod-ucts saw the strongest growth, with a total of 347 millioncontracts traded, up 11% on 2004. Equity products traded250 million contracts, up 2% on 2004.

Eurex has announced that it is in talks with several USexchanges over a possible partnership to further develop itsUS business. Eurex currently distributes 25-30% of its globalbenchmark futures contracts into the US and has seen totalvolume growth of around one million contracts after theextension of US afternoon trading hours in November.Eurex say any agreement in the US would call for the partnerexchange to take a stake in Eurex US, their US subsidiary.

The London Metal Exchange has said its annual volume fig-ures for 2005 show a record year for the exchange with over78 million lots traded. This is an increase of 9.3% on 2004with total traded volume being worth over $4,500 billion.

Research conducted by the US’sNational Bureau of EconomicResearch has found evidence forcognitive dissonance – a behaviouralbias commonly used to explain assetprice trending – in US presidentialelections. Sendhil Mullainathan andEbonya Washington compared thepresidential opinion ratings of nine-teen and twenty year olds two yearsafter the President's election andfound that twenty year olds (whowere eligible to vote in the election)showed greater polarization of opin-ions than comparable nineteen year

olds (who were ineligible to vote).They conclude that the very act ofvoting may influence political atti-tudes and that a vote for a candidateoften leads to more favourable inter-pretations of their actions in thefuture. The results may suggest thatsimilar dissonance can affect stockmarket investors by inducing a morefavourable view of a company’s per-formance, than that warranted by thefundamentals, once that stock hasbeen purchased. For full story seeNBER Working Paper No. 11910,“Sticking with your Vote”.

Further evidence forinvestor bias

Ralph Acamora's technical researchteam, recently layed-off byPrudential Equity Group in NewYork, has been hired by KnightCapital of New Jersey. Knight haveannounced the formation of KnightTechnical Research, to be headed byAcampora, to provide market analy-sis to their client base. Acampora isjoined by Peter. J. Martin and PeterW. Austin both previously ofPrudential. Knight say their researchwill focus on price momentum, sen-timent and time cycles across allasset classes.

ON THEMOVE...


Market Views

US STOCKSYIELD CURVE INVERSION MAY CONFIRM SECULAR BEAR MARKET FOR 2006 by George Slezak

Many market technicians dividethe last century of US stockmarket history into 16 to 20

year periods of either secular bull mar-kets or secular bear markets. 1908 to1929, 1949 to 1966, and 1982 to 2000could be considered secular bull mar-kets. 1929 to 1948 and 1966 to 1982could be considered secular bear mar-kets and now, possibly, 2000 forward 15to 20 years may also be another bearmarket.

We can see the rolling four year cyclein the market throughout the last cen-

tury. In secular bull market periods, thefour year cycle lows tend to be relative-ly shallow declines of less than 30%. Insecular bear market periods the fouryear cycle lows tend to have declinesranging from 30 % to 60%. The nextfour year cycle lows are expected by theend of 2006 or early 2007 and inpreparing for that coming four yearcycle low, the issue is whether we are ina secular bull or bear market.

The decline in the Dow from the2000 high to the 2002 low was about33%. That decline, like the 27% decline

in 1966, is not enough on its own todeclare the onset of a secular bear. Theextent of the decline into the comingfour year cycle low will tell us impor-tant information about the condition ofthe current secular market cycle.

Energy prices, real estate prices andprice movements of the dollar and goldare already important cautionary signalsof a secular bear. Interest rates are alsoan important variable. Rather thanlooking at the absolute interest rate,Figures 1 and 2 suggest the yieldspreads may be an additional

Figure 1.

→→


Market Views

important factor in preparing for thepossibility of a secular bear market.They show the daily Dow plotted overthe spread between the yield on the oneand ten year Treasurys. When the oneyear yield is less than the ten year yieldthe spread is plotted in green. Whenthe one year rate is higher than the tenyear yield the spread is plotted in red.

Figure 1 shows the secular bull mar-ket of 1982 to 2000 when there are twoperiods of interest rate inversions. In1989 the Dow rose and the yield curvewas inverted, while yields were general-ly decreasing. In 2000 the curve wasagain inverted and yields were alsodeclining. The stock market, though

topping, was generally flat through theyear. Contrast this to Figure 2 whichcovers the secular bear market of 1966to 1982. The repeated periods of mar-ket declines that accelerated into thefour year cycle lows show the yieldspreads to be inverted and the individ-ual yields to be rising.

The conclusions to be drawn from asimple visual review of the charts isthat if the current inversion of the yieldcurve continues with rising yields it willbe a strong additional indicator (alongwith rising energy costs, falling dollar,etc) that we may be in a secular bearmarket with likely four year cycledeclines expanding to the range of 30%

to 50%. Consider the potential impacton the US stock market of up to threeadditional rate increases by the Fed thisyear. The continued stability of longterm rates during short term monetarytightening will steepen the yield curveinversion and increase the chances thatwe are in a secular bear market. If thisis so, we could see a decline of 30% to50% in the Dow by the time we get tothe four year cycle low in late 2006 toearly 2007.

George Slezak is editor of US-based research report:www.commitmentsoftraders.com

“THE NEXT FOUR YEAR CYCLE LOWS ARE EXPECTED BY THE END OF 2006 OR IN EARLY 2007.”

Figure 2.


Market Views

The decline in the dollar seen sofar this year follows the high of1.1642 reached in November.

The euro stalled just before the 38.2%Fibonacci retracement level of 1.1583,measured from the .8230 all time eurolow to the 1.3666 all time high. Theholding of this Fibonacci level held andthe subsequent piercing of the resist-ance line shown in Figure 1 suggeststhe decline to 1.1642 was merely a tem-porary pullback within a still valid,long-term uptrend for the dollar.However, a closer looks suggests thatdollar bulls still have their work cut outas some strong resistance levels loomlarge. Firstly, there is an upside

Fibonacci level at 1.2410/15 that needsto be surmounted although this couldprove only a minor obstacle.

Upside targets The bigger barrier for the dollar is1.2589. This level represents both thelast major peak and the last shoulder ina larger reversal pattern. As such, anyrally might fail here completely and leadto a larger move lower to test 1.1642again. Alternatively, if the market canmuster the strength needed to make itabove 1.2589 then a concerted effort topush up past the previous low seen ear-lier in 2005 at 1.2732 should follow.

A move above this level will open the

way to the remaining upside Fibonaccilevels at 1.2895, 1.3195 and even as faras the 1.3666 all-time-high, whichwould then look vulnerable. On thedownside, the main support still lies at1.1642 and the target beyond this cur-rently stands at 1.1582 and a projectedbottom of 1.1470 where some signifi-cant upside could kick in again.However, at the time of writing thedownside pressure has eased off andmore upside looks likely, with the nextbig challenge at 1.2589.

Steve Wesiak is technical analyst forABN Amro in Amsterdam.

EURUSDPOSSIBLE RESUMPTION OF BIGGER UP TRENDby Steven Wesiak

Figure 1.


Market Views

The market remains fixated asgold powers above that $500barrier for the first time since

1985. Yet perhaps more importantly wehave also reached an interesting stand-off. A place where bearish camps pre-pare to throw in the towel and the con-sensus view amongst speculative goldbugs gathers pace. After such anoverextended run, it would be healthyto see some temporary profit taking, asinvestors prepare for the next legupwards and the retail physical marketadjusts to higher gold prices.

The secular trend The fact remains that in 2001, goldceased its twenty year secular declineand started to rise (see Figure 1).Following its price slump at the lows of$252, a double bottom formed andfrom this base the resilient pattern ofhigher peaks and troughs would laterform. More recently, a trading rangebreakout and failed head & shoulderssignal launched price momentumthrough that all-important $500 level,extending into overhead resistance at$550/60. Indeed, we have now retracedmore than one-half of what had beenlost in the 1980s bear market. The infla-tion-adjusted gold chart reminds usthat the underlying demand for goldremains strong and that confirmationabove the $550/60 supply zone willunlock upside scope.

Ultimately, it was the market action inSeptember 2005 that confirmed gold'spotential, when prices began to postmulti-year highs across the reserve cur-rencies. The broadening out of goldstrength has fuelled greater liquidityinto the market, increasing its portfolio

appeal within the investment communi-ty. We have entered the second psycho-logical perception phase for gold wherethe general public, financial media andinvestors recognize that the bull marketis real.

Gold's relationship with USDTraditionally, gold prices share a nega-tive relationship with movements in theUSD. Periods of USD strength are usu-ally accompanied by declining goldprices (see Figure 2). However, Figure 3

GOLD THE NATURE OF THE BULL by Ron William

Figure 1. Inflation-adjusted gold price chart illustrates the end of a secular bear market andinterim resistance zone at $550/60.

Figure 2. Rolling annual correlation coefficient showing the negative relationship between goldand USD index following a cyclical pattern. We highlight its dramatic transition in recent monthsfrom -0.9 to historic highs at +0.6.


illustrates that the relationship betweengold and the USD index (trade weight-ed) follows a cyclical pattern, and look-ing at the 1-month annual rolling corre-lation it is clear that the correlation hasrecently reached historic highs.

While the past couple of years haveseen the fall in USD index matchedalmost exactly by the advance in gold

prices (which have mirrored theEUR/USD uptrend), in recent monthsthis trading relationship has eased dra-matically from a low of -0.9 in June toa current historic high of +0.6 (simul-taneous with gold's multi-year breakoutacross the major currencies). The loos-ening relationship between these twopowerful asset-classes continues to pro-

vide welcome traction for speculativegold bugs.

The traditional negative correlationbetween gold prices and USD has beenrationalised by the simple fact thatwhen the USD appreciates, gold pricesin local currencies automaticallybecome more expensive. This has anegative impact on demand and wouldexplain the decline in gold prices fromthe mid-1980s to 1999.

However, there are internal differ-ences this time around. The first differ-ence stems from increased global liq-uidity and growing interest from fundmanagers. The latter has been fuelledby restricted sales on the part of centralbanks, reduced downside hedging activ-ity by producers and the non perform-ance of other asset classes. With pock-ets estimated to be as deep as one tril-lion dollars, the hedge fund communi-ty's appetite for gold will continue toleave footprints.

Furthermore, according to the WorldGold Council, the fastest rate of golddemand has come from identifiableinvestment demand, which saw a rise of90% in Q1 2005. Of this, exchangetraded funds (ETFs) saw a 465%growth in Q1 2005.

Market Views

Figure 3. Gold overlain with USD index (inverted). Between 1990 and early 2000 both marketsdiverged, remaining negatively correlated, (i.e. gold up, USD down). However, within the last fewyears gold and the USD index (inverted) have tracked each other closely and moved into a pos-itive relationship (gold up, USD up). →→


Market Views

Relative performanceGold's heightening attractiveness is alsodue to how undervalued it is and moreimportantly, it's improving perform-ance relative to other asset-classes. Wecan see this on the S&P/gold ratiochart in Figure 4, which shows that theS&P peaked against gold in 1999.

The gold/oil ratio shows that gold isturning away from historic lows andhas recently broken its short-termdowntrend. Once oil reasserts itsupwards trend, it is quite plausible thatsome of these extra petrodollars will beused to purchase gold, especially asgold continues to outperform othercurrencies.

The greatest structural difference isthe emergence of India, China andcountries in the Middle East as majormarkets for gold. Indeed, the fastestgrowing markets for gold have reachedcritical mass and are enjoying a periodof unprecedented growth in theirdomestic economies. Increasing pros-perity and higher disposable incomeswill translate into higher gold demandeven with increasing local gold prices.(Note: Jewelry demand accounts for75% of total gold demand). The onlycaveat is that we may see temporarydips in retail demand in the short-termas people adjust to the $560 gold price.A sustained push above this levelshould resume demand as the seculartrend gains wider acceptance.

ConclusionThe major upward trend in gold perse-veres. However, trend followers arewell advised not to expect linear projec-tions upwards. In practice, marketswings remain complex. A twenty-yearsecular bear market is such a large frac-tion of an adult's working life thatinvestors may be inclined to believeprices only ever move in one direction.Inevitable corrections to this view willprovide short to medium-term profit-taking opportunities.

These corrective setbacks shouldpresent longer-term investors withanother bite at the cherry and greatbuying opportunities for the secondhalf of 2006. By this time the Fed'sinterest rate cycle is expected to havepeaked and oil is likely to reassert itself

above its Katrina top. Both factors arecatalysts for USD weakness, with con-sequent tailwind for gold. This effect issupplemented by petrodollars and theincrease in gold demand stemmingfrom the Middle East, as excesspetrodollars stoke demand throughwealth creation. Moreover, the non per-formance of varied asset classes, cou-pled with the generalized increase inoverall commodity demand fromemerging markets, looks set to enticesecular gold impetus back above the1980 peaks of $850.

Ron William is technical analyst atInvestors Intelligence in London, adivision of Stockcube plc.

Figure 4. S&P/Gold ratio

Figure 5. Gold/Oil ratio

“WITH POCKETS ESTIMATED TO BE AS

DEEP AS ONE TRILLIONDOLLARS, THE HEDGEFUND COMMUNITY'SAPPETITE FOR GOLDWILL CONTINUE TO

LEAVE FOOTPRINTS.”

Desktop Qualityin a mobile environment

Real-Time Quotesand Charts on your

Blackberry orMobile/Cell Phone

Information at your f ingertips

-Don’t just look at where the market is trading now. Get the whole picture.

-Access to high quality real-t ime charts defined by us or the user.

-No software instal lat ion means it is quick and easy to set up.

-Reduce the number of mobile devices you carry.-Accessible world-wide 24/7

8 Archers Court48 Masons Hil lBromleyKent BR2 9JGUnited Kingdom

Phone: +44 (0) 20 8313 0992Fax: +44 (0) 20 8313 0996E-mail: [email protected]

The RIM and BlackBerry famil ies of related marks, images and symbols are the exclusiveproperties of and trademarks or registered trademarks of Research In Motion Limited -used by permission.


Techniques

A SIMPLE TREND FOLLOWING STRATEGY FORINDIVIDUAL STOCKS by Cole Wilcox and Eric Crittenden

We decided to put a long-onlytrend following strategy tothe test by running it against

a comprehensive database of USstocks.

The stocksThe database we used included24,000+ individual securities from theNYSE, AMEX & NASDAQ exchangesfrom January 1983 to December 2004.About half of these were stocks thatwere delisted at some point between1983 and 2004 and all were back adjust-ed for corporate actions such as divi-dends, splits, mergers, etc. We used aminimum stock price filter to avoidpenny stocks and a minimum daily liq-uidity filter to avoid stocks too illiquidto generate realistic results.

The strategyThe trade entry level chosen for thetests was the all-time highest closingprice. This means that if today's close isgreater than or equal to the highestclose during the stock's entire historythen we buy on tomorrow's openingprice. We chose this method to avoidambiguity. A stock that is at an all-timehigh must be in an uptrend by any rea-sonable person's definition. This is atrend following entry in its purest form.

Figures 1 and 2 illustrate what wouldhave been notable trade entries for thesystem for two stocks; EMC Corp andTaser Corp. The green dots denoteinstances where the closing price forthe week was at a new all-time high.The horizontal pink line represents theprevious all-time high that would havetriggered the initial entry:

Exits are essential to any trend fol-lowing strategy. We decided to useaverage true range trailing stopsbecause they are universally applicableand commonly used by trend followingprograms. The average true range is aderivative of the true range indicatorwhich measures the daily movement ofa security by calculating the greater of:

The true range illustrates the maxi-mum distance the security's price trav-eled from the close of one business dayto the close of the next business day,capturing overnight gaps and intradayprice swings. The average of this

Trend following strategies are typically applied to stock indices but can they alsobe successfully used on individual stocks?

Figures 1 and 2

Figure 3. A 10 ATR stop on a volatile internet stock might be 55% away from the stock price

Today's high minus today's lowToday's high minus yesterday's close Yesterday's close minus today's low

•••

→→


Techniques

value can be used to integrate thevolatility of a security into a universallyapplicable trailing stop. Average truerange stops effectively account forvolatility differences between individualsecurities (see Figures 3 and 4). For thepurposes of this project we chose toexit a stock on the open the day afterthe exit level was breached.

Trading results To determine how well these entriesand exits would have worked in thepast, we need to test the strategyagainst the historical database. Figure 5shows the results from using an all-timehigh entry strategy along with a 10-unitATR stop. There were 18,000+ trades

over the 22 year test period and trans-action costs of 0.5% round-turn werededucted from each trade to accountfor estimated commission and slippage.

The horizontal axis represents the netreturn from the trade. The vertical axisindicates how many trades would haveachieved the indicated net return. Thelong volatility component resultingfrom the combination of a trend fol-lowing entry and trailing volatility stopis immediately recognizable by the sig-nificant right skew of the distribution.17% of trades would have gained 50%or more while less than 3% of tradeswould have registered a loss equal to, orworse, than -50%.

At first glance a winning percentage

of 49.3% might seem less than impres-sive, but it is relatively high for a trendfollowing system. Trend following sys-tems can be very effective with muchlower winning percentages if the prof-itable trades are significantly larger thanthe more frequent unprofitable trades.In the case of this system the ratiobetween average winning trade andaverage losing trade is 2.56; a healthynumber in our experience.

A positive mathematical expectancy(the weighted average of the observedprobability distribution) is the bareminimum needed to justify the tradingsystem. The results yield an expectan-cy of approximately 15.2% with anaverage holding period of 305 calendardays. Considering the significance ofthe sample size, depth of the sampleperiod, realistic assumptions used, andthe right skewed return distribution, wefelt this was a very solid foundation tobuild from.

Other settings for the ATR stop weretested ranging from 8 to 12 in steps of0.5. The middle setting of 10 was cho-sen for illustration purposes. Therewere no material differences in resultsamong the various settings. HigherATR levels (looser stops) resulted inslightly higher winning percentages andslightly lower win/loss ratios. Theinverse was true of lower ATR levels(tighter stops).

Normalizing risk Figure 6 illustrates a collection of alltrades, each normalized for its own risk.This concept typically requires someexplanation. Every trade ultimately hasa recorded percent return. Every tradealso has a recorded percentage initialrisk from the day of entry. The resultis that we know what the percent returnof each trade would have been and weknow how much risk each trade wouldhave subjected us to. The ratio betweenthese two numbers is the focus of thissection.

The simplest way to interpret the fol-lowing distribution is to focus on a cou-ple of specific numbers on the hori-zontal axis. First the -100% columncontains trade results where the

Figure 4. A 10 ATR stop on a quiet utility stock might only be 15% away from the stock price

Figure 5.


Techniques

absolute value of the net loss approxi-mately equaled the initial risk (lost thefull amount that was expected).Likewise, the 100% column containstrades where the net gain approximate-ly equaled the initial risk. Results worsethan -100% represent trades where wewould have lost more than that budget-ed for on the trade (negative outliertrades). This is usually the result of alarge, overnight price decline. Resultsgreater than 100% represent tradeswhere we would have gained more thanthat initially risked (positive outliertrades). Consider the following twoscenarios:

We purchase XYZ stock at $15.50.The 10 ATR stop is $11.32. Initialrisk in this case is 27%. Two yearslater we sell XYZ at $30.75 for a gainof 98%. The ratio between gain andinitial risk is 3.63 or 363%. This datapoint would therefore go in the350% column in the following distri-bution. The return would have been363% the size of the initial risk.

We purchase ABC stock at $32.35.The 10 ATR stop is $26.53. Initial

risk in this case is 18%. Threemonths later the company misses itsearnings estimate and gaps downwell below the stop. We sell ABC at$21.15 for a loss of -35%. The ratiobetween gain and initial risk is -1.94or -194%. This data point wouldtherefore go in the -200% column.The loss would have been almostdouble what was budgeted for.

From the above distribution one canget a feel for how realistic a 10 ATRstop is for real world trading. Datapoints to the left of -100% reflecttrades that couldn't be controlled.There were less than 400 trades thatcaused worse than expected losses.This amounts to approximately 2% ofall historical trades. In some ways thissecond distribution is more importantthan the first. Normalizing each tradeby its own risk reduces the possibilitythat highly volatile stocks will unjustifi-ably dominate the results.

Short sellingFor the purposes of this project wedecided against testing short selling

strategies. Our reasons for this have todo with the following issues:

Forced buy-ins: A short seller has toborrow shares before they can shortsell them. Likewise, the short sellermust return (deliver) the shares shouldthe brokerage firm call them back.From the historical data available thereis no way to know when or if a shortseller would have been subject to aforced buy-in.

Borrowing shares: Short selling asecurity requires borrowing sharesfrom an investor who holds them in amargin account. Not all stocks meetthese criteria all the time; some nevermeet these criteria at all. There is noreliable method to determine whatstocks from the investable universewould have been realistically shortablein the past.

Limited expectancy: With respect tolong term trend following, short sellingoffers a severely limited mathematicalexpectancy. The price of a stock canonly decline by a maximum of 100%.However, it can rise by an infiniteamount. This is a significant disabilityto overcome.

ConclusionsThe evidence suggests that trend fol-lowing can work well on stocks. Buyingstocks at new all time highs and exitingthem after they've fallen below a 10ATR trailing stop would have yielded asignificant return on average.Moreover, test results show the poten-tial for diversification exceeding that ofthe typical mutual fund. The traderesults distribution shows significantright skew, indicating that large outliertrades would have been concentratedamong winning trades rather than los-ing trades.

Eric Crittenden is director ofresearch at Blackstar Funds, a USbased hedge fund. Cole Wilcox ismanager of the fund.

Figure 6.

“TREND FOLLOWING SYSTEMS CAN BE VERY EFFECTIVEWITH MUCH LOWER WINNING PERCENTAGES IF THE

PROFITABLE TRADES ARE SIGNIFICANTLY LARGER THANTHE MORE FREQUENT UNPROFITABLE TRADES.”

REVIEWING2005HOW DID THE STOCK MARKET

"RULES" PERFORM?


Techniques

The January Barometer The January Barometer states simplythat if the stock market rallies and endshigher in January then it will also endthe year up. Similarly, Januarys that enddown will mean the year ends lower.Since 1960, the barometer has an 80%success rate for up and down yearscombined. One popular explanationfor this effect says that with the USpresidential Inauguration Day takingplace in January, several key politicalevents take place in this month thatmay impact on the US economy.Perhaps the most important is the pres-ident's State of Union address thatpresents forthcoming economic poli-cies. This speech can set the tone forthe markets for the year ahead.

The Dow had a very poor month inJanuary 2005 and the index did indeedend the year lower, but only just. TheFTSE ended the year some 800 pointshigher although in January it closedonly 38 points higher. So, on the face ofit the January Barometer worked wellfor both markets. However, it failed forthe S&P500 and in the case of theFTSE, a leap of faith was required backin January after such a moderate rise ifyou were considering staying in themarket all year.

The 5-year Rule The 5-rear rule states that years endingin '5' (1975, 1985 etc) have a strong ten-dency to end the year higher. Indeed,between 1880 and 1995, all 12 yearsending in '5' have seen the Dow Jonesclose higher. This would appear to bestrong proof of a '5' year upward biasfor the stock market probably due, ifanything, to the US presidential elec-

tion cycle. However, the small samplesize and the tendency of stock prices toend the year higher rather than lower inthe majority of years means the '5' yearrule does not stand up to rigorous sta-tistical investigation. Although the '5'rule worked (just) for the S&P500, itfailed for the first time for the DowJones.

Sell in MaySell in May and come back on St.Leger's Day (or thereabouts) has beenone of the more reliable stock marketrules. Since 1970, all markets (US,Europe and Japan) have shown a cleartendency to outperform during thewinter months. The relatively poor per-formance of stocks during summermay be due to disruption caused byholidays and the tendency of marketplayers to be more optimistic towardsthe end of the year in anticipation ofnew company earnings data and thestart of a new year ahead.

Nevertheless, the rule failed to per-form for both US and UK stocks lastyear. Indeed, stocks enjoyed somethingof a rally over the summer as a sharprise in oil prices boosted oil stocks (andthe indices) and renewed optimism inthe US economy supported a positivemarket outlook. Surprisingly this hap-

pened despite a backdrop of steadilyincreasing US rates from the Fed.

The January Effect The January Effect refers to the ten-dency of small capitalisation stocks torally in January after a sell-off inDecember. Possible explanations forthis include year-end tax loss sellingcombined with January corporate earn-ings results and new year investor opti-mism. For the US, the Russell 2000Index is used as the small-cap barome-ter and this ended January lower on themonth. However, for the UK, theFTSE SmallCap index opened January2005 at 2761.30 and closed the monthat 2877.50.

There is a tendency to explain awayyears that fail to perform according tothe rules as being exceptional in someway. 2005 certainly falls into that cate-gory as the US and UK economies sawvery high oil prices along with sharplyrising interest rates. Given these factors,it is surprising that the stock marketsperformed as well as they did.Nevertheless, the turmoil caused bythese factors resulted in a poor overallperformance for the rules, especially inthe US, with only the JanuaryBarometer working for both the USand UK stock markets.

2005 DJIA S&P500 FTSE100January 1 10783. 75 1211.92 4814.30January 31 10489.94 1181.27 4852.30May 1 10192.51 1191.50 4801.70October 1 10568.70 1228.81 5477.70December 31 10717.50 1248.29 5618.80

Rule DJIA S&P500 FTSE100January Barometer Yes No Yes 5 Year Rule No Yes Yes Sell in May No No No

Russell 2000 Index

FTSE SmallCapJanuary Effect

No

Yes

US and UK stock market closes in 2005

Did the rules work in 2005?

The various stock market"rules" for the US and UKmarkets continue to

receive widespread press coverageand be of interest to fund andportfolio managers from year toyear. We look at how the rules per-formed in 2005.


Techniques


Techniques

THE ICHIMOKUby Jean Paul van Straalen

Despite being incorporated intomost advanced technicalanalysis software packages,

the Ichimoku Kinko Hyo is not widelydescribed in the technical analysis liter-ature nor is it fully embraced by thetechnical analyst community. Yet it pro-vides the technical analyst with anothervaluable tool for money management.

In this article, I will first decomposethis indicator and interpret the varioussignals derived from it. I will then high-light its major advantages before usingit to analyse the S&P 500 index andgold. Finally, I will conclude with someobservations and additional remarks.

Decomposing the IchimokuKinko Hyo indicator Ichimoku Kinko Hyo charts consist offive sub-indicators:

Tenkan-Sen - the turning or conversion lineKijun-Sen - the standard or baselineChikou Span - the delayed line orlagging spanSenkou Span A - the first precedingspanSenkou Span B - the second preceding span

All five sub-indicators are displayed inFigure 1.

The Ichimoku charts can easily beconstructed in Microsoft Excel byusing the following formulas.

The Tenkan-Sen, the turning line, iscalculated as follows: (Highest high +Lowest low)/2 for the past 9 daysincluding today's high and low.

The Kijun-Sen, the standard line, is cal-

culated as follows: (Highest high +Lowest low)/2 for the past 26 daysincluding today's high and low.

The Chikou Span, the delayed line, istoday's closing price recorded 26 daysprior to today.

The Senkou Span A, the first preced-ing span, is calculated using the stan-dard and turning lines: (Standard line +Turning line)/2. The time span is pro-jected 26 days ahead including today.

The Senkou Span B, the second pre-ceding span, is calculated as follows:(Highest high + Lowest low)/2 for thepast 52 days including today and is pro-jected 26 days ahead including today'.

The area between Senkou Span A andSenkou Span B is known as the Kumo,which means cloud.

InterpretationThe Ichimoku indicator provides arange of signals which vary in strength.When two sub-indicators cross eachother, this triggers a bullish or a bearishsignal. The strength of a bullish orbearish signal depends on where thecrossover between the sub-indicatorsoccurs in relation to the Kumo. Thesignificance of a signal increases whenthe various sub-indicators confirm thesignal and thus increases the probabili-ty that the up or down trend will takeplace or continue.

The Tenkan-sen and the Kijun-Senlines: A signal is triggered when theTenkan-Sen line crosses the Kijun-Senline. An upwards crossover of the

INDICATOR

1.

2.

3.

4.

5.

This Japanese charting technique was invented prior to World War II by GoichiHosoda, a Tokyo newspaper journalist who called himself Ichimoku Sanjin,which translates as ‘a glance of a mountain man’. The technique is known asIchimoku charting, which means ‘at a glance’. The full name of the techniqueis actually the Ichimoku Kinko Hyo. Kinko means ‘balance’ or ‘equilibrium’,and hyo is the Japanese word for ‘table’. The full translation, therefore, is the‘table of equilibrium prices at a glance’. The Ichimoku guidebook was finallypublished in 1968.

Hosoda was a lover of nature and therefore he used expressions of nature todenote whether sentiment towards an equity, bond, commodity or index wasbullish or bearish. When the price of the asset was above the clouds (where thesun was shining), this triggered a buy signal for the technical analyst. When theprice of the asset was below the clouds (where it was raining), this would trig-ger a sell signal. According to Hosoda, the man standing on the mountaincould easily observe past market behaviour and current market sentiment,which enabled him to predict future directions of a particular asset.

Hosoda's choice of time spans is interesting, as he used 9, 26 and 52 days inhis calculations. When Hosoda constructed the Ichimoku indicator, the tradingweek consisted of six days (including Saturday). So nine days reflect one-and-a-half trading weeks, while 26 days and 52 days reflect one and two monthsrespectively.

History of the Ichimoku Kinko Hyo


Techniques

→→


Techniques

Kijun-sen line provides a bullish signal,while a downwards crossover gives abearish signal. The bullish signal isstronger when the crossover occursabove the Kumo, weaker when itoccurs within the Kumo, and weakestwhen it occurs below the Kumo. Theopposite is true for the bearish signal.

In Figure 1, for example, the Tenkan-Sen line crosses the Kijun-Sen linefrom below in September 2004, trigger-ing a bullish signal. The crossoveroccurred on the upper side of theKumo, indicating a relatively strongbullish signal.

The Kumo: The Kumo provides sup-port and resistance levels. If the price

of the asset falls into the Kumo fromabove, then it is at a support level. In adowntrend, the Senkou-Span A will bethe first support level and the SenkouSpan B operates as a second supportlevel.

When the price of the asset enters theKumo from below, the price is at aresistance level.

The thickness and steepness of thecloud reflects the acceleration of theprice. A thick cloud corresponds to anacceleration of the Senkou Span Acompared to the Senkou Span B. Theexplanation lies in the time span of theSenkou Span A (26 days versus 52 days)and the Senkou Span B. When thecloud is thin, the new highs or lows

behave in such a way that the values ofthe Senkou Span A and Senkou Span Bconverge, even if they have differenttime spans.

Where the asset price is positionedrelative to the Kumo indicates whetherit is in a bullish or bearish trend mode.When the price is below the Kumo, thetrend is downwards and in a bearishmode. Similarly, when the price is abovethe Kumo, the trend is upwards and ina bullish mode.

The Chikou Span: The Chikou Spanoperates as an additional convictiontool and indicates the current momen-tum or strength of the asset price.When the price is below the ChikouSpan line, this indicates that there aremore buyers than sellers in the currentmarket, reflecting that the asset pricehas strengthened. When the price isabove the Chikou Span line, this indi-cates there are more sellers active in themarket than buyers reflecting that theasset price has declined the last 26 days.

Advantages of the IchimokuOne of the advantages of theIchimoku indicator is that it uses his-torical highs and lows during the trad-ing day instead of the closing pricesused by standard moving averages tech-niques. As a result, the midpoints ofthe various lines provide a differentinsight into the daily volatility of theprice than do the moving average meth-ods.

The Ichimoku combines backwardand forward-looking indicators, whichenables the trader to make future pro-jections. This indicator can be appliedto any market and any time frame.

The Ichimoku indicator can vary instrength and triggers various tradingsignals. These signals can be easily com-bined with Western technical tech-niques, but the most logical choice is tocombine the Ichimoku indicator withJapanese candlesticks. The uniquereversal or continuation patterns ofJapanese candlesticks in combinationwith the signals of Ichimoku charts can

Figure 1. The Ichimoku chart including the five sub-indicators. Source: Bloomberg, Metastock

Figure 2. Ichimoku chart for the S&P 500. Source: Bloomberg, Metastock


Techniques

provide the technical analyst with aneven stronger conviction to enter orexit a trade.

Analysing the S&P 500 and gold To illustrate the usefulness of theIchimoku indicator, I have analysed theprice of gold and the S&P 500 indexusing Hosoda's different time spans.Gold is a commodity that is currently ina significant trending mode, while theS&P 500 has a less prominent trend. Iwill show that the Ichimoku indicator isbetter at analysing assets that are in atrending mode.

S&P 500: An Ichimoku chart of theS&P 500 is shown in Figure 2. At thebeginning of 2005, the Tenkan-Sen linefirst crossed the Kijun-Sen line fromabove and the S&P 500 followed with adrop into the Kumo. The Tenkan-Senline then crossed the Kijun-Sen linefrom below within the Kumo inFebruary 2005, indicating a weak bull-ish signal. It's clear that when the assetis in a non-trending market, the fre-quency of the trades will undoubtedlyhave a negative impact on results.

In early November 2005, the Tenkan-Sen line crossed the Kijun-Sen linefrom below, triggering a bullish signal.Unfortunately, this bullish signal isweak, as the crossover occurs below theKumo. The more aggressive trader cantake this opportunity to go long. In this

case, the trading strategy would pay off.As the index broke above the Kumo'stwo resistance levels and closed signifi-cantly higher than the Kumo, this sup-ports the bullish outlook for the S&P500.

The bullish trend of the S&P 500 willremain intact as long as the Kumo, theTenkan-Sen and the Kijun-Sen lineshold as support levels. So the traderwith a long position should maintainhis current position, while short posi-tions should be closed out unless theindex falls below the Tenkan-Sen line.

Gold: An Ichimoku chart for Gold isshown in Figure 3. Between May andOctober 2004, the upward trend ofgold remained intact, as the SenkouSpan A held as a strong support level.A bullish signal was generated in June/July 2005, when the Tenkan-Sen linecrossed the Kijun-Sen line from below.This crossover took place above theKumo, triggering a more bullish viewon gold. The bullish signal will weakenwhen the gold price first falls below theTenkan-Sen line and then below theKijun-Sen line. The Kumo formed astrong support level in the uptrend, soif this does not hold, the outlook forgold will deteriorate significantly. Theshort-term technical picture for goldremains bullish, as the current goldprice is still above the Tenkan-Sen line,the Kijun-Sen line and - more impor-

tantly - above the Kumo.

ConclusionThe Ichimoku indicator is a trend-fol-lowing system and defines support andresistance levels within the market inorder to generate buy and sell signals.This Japanese technical indicator cannot only be applied to equity marketsbut also to bonds, indices, commoditiesand every other market. The Ichimokuindicator is less suitable for use in cur-rency markets, as it is difficult to deter-mine high, low, open and close pricesgiven these markets' 24-hour trading.

The default time spans can bechanged for various markets, as mostsecurities are not traded on Saturdays.Therefore, the determining of the timespans will be purely subjective. Tradersand technical analysts will have to runbacktests for their specific markets inorder to obtain the optimal time spans.

The main advantage of the Ichimokutechnique is that it allows you to timeyour trades effectively. In a non-trend-ing market, commission and bid-askspreads will impact negatively on theresults of short-term horizon trades.Therefore, the Ichimoku charts shouldbe combined with other technical indi-cators such as oscillators in non-trend-ing environments.

The mountain man's view is a uniquecombination of a backward-lookingindicator - the Chikou line - and for-ward-looking indicators such as theSenkou Span A and the Senkou Span B.The Ichimoku indicator is therefore auseful additional tool for the technicalanalyst.

Jean Paul van Straalen FRM is aquants strategist and co-portfoliomanager of the Global Equity ValueFund at ABN AMRO AssetManagement.

This article is written from a personal perspec-tive and does not necessarily represent the viewsof ABN AMRO Asset Management.

Figure 3. Ichimoku chart for gold. Source: Bloomberg, Metastock


Techniques

Drawing trendlinesThere is an understandable and accept-ed tendency to expect widely recog-nised horizontal support and resistancelevels to be tested when these levels arere-approached. Many traders set profittargets referenced to such levels.Unfortunately, disappointment fre-quently results. Very often the marketwill reverse just short of these isolatedhighs/lows much to the annoyance ofmany traders who'd been counting thepennies before they were securely inthe till. It is my contention that thishappens because the keysupport/resistance levels aren't neces-sarily located at these highs and lows asis often believed.

Consider the stylised bar chart inFigure 1. Areas [A] and [B] represent atwo period reversal structure that iscommonly seen in the markets. Thus inan up market - as typified by [A] - fol-lowing a strong close, the next periodinitially sees further topside priceaction. However, the power balancebetween bull and bear is changing,prices back-off sharply and we have alower close (the reverse situation ofthis for a down market). Taking thehighlighted area [A], our bulls wouldhave been contented at the close of thefirst period (close "c"). However, whilstwe did initially push higher (suckingmore weakly committed bulls in), bythe next close (close "d") there were alot of unhappy traders - contrastingstrongly with their feelings at the closeof the previous period.

It is my experience that it is the highof the first period "b" that constitutes a

much more potent resistance levelwhen re-tested than the actual isolatedhigh of the structure marked "a". Thisis not to say that "a" won't be attacked,but time and again you'll find it's "b"where the real battle is fought, andwhere the subsequent reversal or con-solidation will be initiated from. Onecould reckon "b" as being more signifi-cant than "a" due to a greater degree ofmarket confidence at "b", and almostcertainly "b" will be associated withhigher volume. I call "b" the maximumbullish consensus point.

So rather than blindly drawing mytrendlines through high and lows, Idraw them through the maximumbull/bear consensus points of a pricestructure even though this means thatpart of the price action may be left out-side of the trendline/channel. I havefound that trendlines drawn using thismethod are a much more powerful tool

for understanding market action.

Bollinger bandsI have done a lot of research into opti-mal parameters for Bollinger Bands andam very much struck how 8 or 9 peri-ods with 1.6 or 1.7 standard deviationskeep cropping up as giving the best sys-tem testing results. Because I'm a userof Fibonacci, I've now settled on 8

COMBINING BOLLINGER, FIBONACCI &TRENDLINES by Glyn Bradney

Glyn Bradney, technical analyst at Reuters, provides an insight into how he com-bines several techniques in his market analysis.

Figure 1.

These are pivotal as to how I look at markets (in association with their CounterConsensus counterparts). I use them in a number of ways but the most impor-tant is in drawing trendlines and trend channels. You'll notice in Figure 1 that I donot draw trendlines through isolated highs or lows and also that I don't necessar-ily draw them using the bull bear consensus point of a candle that is itself an iso-lated high or low. Looks a fiddle at first glance, but in fact it's entirely mathemat-ical and very much underpinned with a psychological justification. See text.

The black dots associated with the candles are what I term Bull Bear Concensuspoints. The formula is:If(CLOSE>=OPEN,(HIGH+CLOSE)/2, (LOW+CLOSE)/2)

The purple dots are Bull Bear Counter Concensus points. The formula is:If(CLOSE>=OPEN,(LOW+OPEN)/2,(HIGH+OPEN)/2)

Bull/Bear Consensus points

→→


Techniques

and 1.618 as my normal defaults.Closes outside the envelope (Close

field, 8, 1.618) with the top and bottombands uni-directional is usually anexcellent signal for at least a temporaryhalt in the trend, if not a full blownreversal.

Adding an outer envelope based uponthe High field (for the top band) andthe Low field (for the bottom band) issomething I've found to be extremelyvaluable - especially when looking to

target exit (profit) points dynamically ina strongly trending market.

The inside "snake" based on Close, 8,0.618 is another Bollinger variant I use.It's observable that in strongly trendingmarkets price action will occur in theouter part of the envelopes (between8/1.618 and 8/0.618).

Closes outside both 8/1.618 and thecommonly used 21/2 envelopes withthe 8/1.618 having moved outside the21/2 band, signals a very heavily

extended market. A significant isolatedhigh/low usually needs to have beentaken out for a subsequent close toremain outside both bands. After thistype of excess a correction is followedby one more move to thetopside/downside which will result in asignificant top/bottom being formed.

Glyn Bradney is head of technicalanalysis at Reuters in London.

Figure 2. See boxes for description. (The red analysis line in the subchart is a Trend indicator which I've developed which I've called "Icarus".The Bollingers in the subchart (dashed blue lines) are taking the Icarus analysis as the input, parameters are 8 periods and 1.618 Std devs).

----------------- : Bollingers based on Close, 8 periods, 1.618 Std deviations---------------- : Top Bollinger based on High, 8, 1.618 parameters ; Bottom Bollinger on Low, 8, 1.618___________ : Bollingers based on Close, 8, 0.618 parameters.___________ : Bollingers based on Close, 21 periods, 2 Std devs.

All Bollingers use a Simple Moving Average as the underlying basis.

Bollinger settings (Figure 2)

Asia and Japan Hedge Fund Directory

Qsjodjqbm!Tqpotps;!

Tfdpoebsz!Tqpotps;!

2005

EurekaProductAD220405.pdf 19/05/2005 12:18:12


Interview

THE TECHNICAL ANALYST TALKS TO...

Guido Stiel is a research manager in behavioural finance at Deka Investment inFrankfurt. He is also a portfolio manager of EuropaTrend, a fund based onDeka's technical strategy.

TA: How many members are there in your team and whatare their different roles?

GS: There are currently four members in our team. Theteam is headed by Manfred Huebner, a former chairman ofthe German society for technical analysis (VTAD).Manfred's focus is on behavioural finance, sentiment analy-sis and the European equity markets. Also in the team isPatrick Hussy who is responsible for bond market researchand the US equity markets. My focus is on the Asian equitymarkets, commodities and the development of systematictrading strategies. The team is completed by Stanimir Denevwho does a lot of the programming and statistical tests.

TA: Who are your main customers?

GS: We are buy side analysts so our clients are the mutualand institutional fund managers, the research analysts andstrategists at Deka Investments. Besides that, we run ourown funds.

TA: Which markets do you look at?

GS: We have a very broad coverage - simply spoken, wecover almost everything Deka can invest in. We evaluateabout 2,500 stocks worldwide and all the important marketand sector indices. In the bond markets we analyze about 20different countries, the spreads between countries and therelative attractiveness of different maturities. In the FXmarkets we have a look at 20 liquid currencies and crossrates. For the purpose of intermarket analysis we investigatethe trends of the commodities that make up the GSCIindex family.

TA: How is Deka's research different?

GS: We do our technical research very systematically other-wise we couldn't handle such a great range of markets. Wehave developed our own models and computer programswhich screen the markets and help us to identify markettrends or contrarian opportunities. We use not only 100%mechanical trading models but try also to implement →→


Interview

mechanical rules wherever possible and effective. DekaInvestment has with its four member team an unusual com-mitment to technical analysis for a fund management com-pany. We are convinced that we have to look at markettrends to prevent costly misallocations and optimise ourentries in, and our exits out, of the markets.

TA: Would you say that Deka's application of technicalanalysis is especially enthusiastic?

GS: It is part of our investment philosophy. Technicalanalysis for us is the most effective way to determine thecurrent trend and momentum of the market and our pre-ferred tool for market timing. We want to go with the trend,at least most of the time. Only when we can see that themarket is positioned one way do we try to establish a con-trarian position. Employing technical analysis in our invest-ment process is one way to control our investment risks.

TA: You mentioned that you combine both fundamentaland technical analysis. Do you use TA mainly for markettiming? Is this applied to both short and long term analysis?

GS: Our approach as an asset manager is fundamental innature. With our fundamental research team we evaluate theprospects of a company, the earnings outlook and so forth.From that we calculate a fair value which reflects our esti-mates. Then we compare this fair value with the currentmarket price and determine how much upside or downsidethe price has. The next step is to formulate the expectation;when and how fast the market price will converge with thefair value we have calculated. This is the time when techni-cal analysis comes into play. We identify the current markettrends, the relative sector trends and overbought or over-sold conditions. From that we score each stock or market.Every score reflects the current state of the trend and theprobability of a continuation of the trend. We maintain a

“TECHNICAL ANALYSIS FOR US IS THE MOST EFFECTIVE WAYTO DETERMINE THE CURRENT TREND AND MOMENTUM OF THE

MARKET AND OUR PREFERRED TOOL FOR MARKET TIMING.”

Left to right: Stanimir Denev, Guido Stiel, Manfred Hübner, Patrick Hussy


Interview

medium to long term investment horizon rather than mak-ing short term calls.

TA: What are the indicators you favour?

GS: As mentioned above we try to identify the mediumterm trends and calculate a probability for these trends tocontinue. For that purpose, moving averages, the ADX andRMI (relative momentum index), simple momentum andBollinger bands have proven to be successful. But we alsohave a look at classic chart patterns. Currently we are tryingto incorporate a systematic approach to Elliott Waves intoour research.

TA: You also mentioned your use of sentiment analysis.Which indicators do you use for this? Do you apply con-trary opinion to market sentiment indicators?

GS: Sentiment indicators play an important role, mainly ascontrarian indicators. Because Manfred Huebner andPatrick Hussy run the "Sentix" behavioural indices, one ofthe biggest sentiment databases worldwide, we have naturalaccess to this data and incorporate it into our framework.With sentiment indicators you get a good insight into mar-ket psychology. To see where the overcrowded trends are isimportant for establishing take-profit rules, for example.The problem is getting real behavioural data. Here Sentixrepresents a very unique approach.

Trend following techniques and momentum strategies aregreat but have two major drawbacks. Firstly, in sidewaysmarkets they lack efficiency and produce unprofitable sig-nals. Secondly, after v-shape reversals you have to give up alot of your profits until you know that the trend has ended.Because v-shape reversals are common in markets wheregreedy or anxious investors have driven markets to unsus-tainable levels, a sentiment analysis in full can be of greatvalue.

TA: To what extent is TA used in German markets?

GS: Among professional investors technical analysis is com-mon, but seldom done systematically. Because everyone hasa Bloomberg, Reuters or other charting software, everyonedoes some sort of technical analysis. But it makes a greatdifference whether you have a look at a chart from time totime and draw some standard indicators from it or screenthe markets systematically with a proven set of indicatorsand rules. For many investors technical analysis is only atool to confirm their preconceived opinions.

TA: To what extent do you use behavioural finance in youranalysis?

GS: Behavioural finance is for us the theoretical foundation

and the reasoning behind technical analysis. Technical analy-sis means observing what markets are doing; behaviouralfinance means understanding why they are doing whatthey're doing. Many well known technical rules have beenconfirmed by academics. As an example I want to mentionthe "reference point effect" and the "anchoring heuristics",which mean that people consider their entry prices or pasthighs and lows when they make investment decisions. Thedynamics behind these effects explain the principles of sup-port and resistance.

TA: What behavioural finance techniques interest youmost?

GS: We only want to use those tools and techniques fromour technical toolkit that are compatible with the latestbehavioural finance research. That is why we rely onmomentum analysis, are aware of the crowd at marketextremes and do our research in as systematic a way as pos-sible.

TA: Is there any difference in the analysis you apply to thecapital markets side of the business and the fund manage-ment side?

GS: From a behavioural perspective we have to differentiatebetween macro and micro. On a macro basis we spotinvestor behaviour and try to take advantage of investormistakes. On a micro basis, we have to build a robustframework that minimizes our own mistakes.

TA: You mentioned using moving average crossovers and abreak out rule. Can you enlarge on this technique?

GS: We use moving average crossovers and a breakout ruleas our medium-term trend following system. The reasonwhy we incorporate a simple breakout rule is that it helps tominimize the unavoidable problems with moving averagesin a sideways trend.

“TECHNICAL ANALYSISMEANS OBSERVING WHAT

MARKETS ARE DOING;BEHAVIOURAL FINANCE

MEANS UNDERSTANDINGWHY THEY ARE DOING IT.”


Subject Matters

A common feature of most of thesestudies is that TA is generally equatedwith trend-following indicators calcu-lated on a closing price basis.Unfortunately, this does not pay justiceto a wide range of technical indicatorsthat assign special importance to differ-ent constellations of Open, High, Lowand Close prices. In two recent paperswe provide evidence that TA involvingHigh, Low and Close prices mightindeed keep such promises (Fiess andMacDonald, 1999 and 2002).

Discrete time seriesCandlestick analysis represents proba-bly the most exhaustive attempt to clas-sify price forecasts according to Open-High-Low-Close constellations. Withincandlestick analysis, as well as in otherforms of TA, the difference betweenopening and closing prices serves as ameasure of the direction of intradaytrends. The difference between highand low prices marks the intraday trad-ing range and represents a basic meas-

ure of volatility. It is the interactionbetween trend and volatility that isassumed to be informative about futureprice developments.

But is it appropriate to think of High,Low and Close prices as different timeseries? After all, they are derived fromthe same asset price. We argue that it isentirely justifiable to treat High, Lowand Close prices as time series with dif-ferent dynamics. The theoretical justifi-cation is derived from Takens' timedelay concept (1981) of Chaos Theory.This concept is easily illustrated fordaily exchange rates: the Open priceseries represents an embedding of theClose price series since Open and Closeare always recorded at the same pointsin time; i.e. the opening and closingtime of the corresponding market. Lowand High prices, however, carry differ-ent information - they do not coincidewith certain times of the day but withthe extreme prices of a trading day.

Since Low and High prices do notoccur at a specific time, neither can beexpressed as an embedding of theClose series, i.e. we cannot simply shiftthe clock forward or backward to getfrom one market High to the next.Thus, even though originating from thesame asset price series, High, Low andClose do not have the same dynamiccharacteristics; it should therefore bepossible to treat them as three differentdiscrete time series. This opens thedoor to modelling the dynamicsbetween High, Low and Close priceswith the help of multivariate time seriestechniques.

A simple way to illustrate the differ-ent dynamics of the High, Low andClose is to look at the autocorrelationfunction (ACF) of the differencesbetween these series. If High, Low andClose had the same time series proper-ties, the difference between these seriesshould be a white noise process:

Orthodox economicthinking is at odds withtechnical analysis (TA).

Under the generally acceptedparadigm of market efficiency,asset prices fully reflect all avail-able information at any giventime, implying that prices donot follow any trends or pat-terns. Any attempt to use pastprices to predict future prices istherefore doomed to fail andshould not deliver any excessprofits. This claim is upheld bymany academic studies of TA.

TOWARDS THE FUNDAMENTALS OFTECHNICAL ANALYSIS by Norbert Fiess

Figure 1: ACF of the difference between High and Low. Daily observations of GBPUSDexchange rate.

→→


Subject Matters

the ACF of High less Close, High lessLow, and Close less Low should there-fore decay immediately to zero. This isnot the case. Figure 1 shows the ACFplot of the difference between theHigh and Low price for the first 500lags of GBPUSD. The slow decay ofthe ACF is significant and hence thedifferences between the two series arestructural and not random in nature. Ittakes approximately 200 lags before theACF reaches zero for the first time.

The slowly decaying ACF reflects ahigh degree of structure between thesetwo series. This is expected since theabsolute difference between High andLow denotes the daily trading range,which is also a simple measurement ofdaily volatility. A slowly decaying ACFindicates the presence of autoregres-sive conditional heteroscedasticity(ARCH), a feature of daily asset priceswidely accepted within the marketmicrostructure theory. Thus, like mod-els from the generalized ARCH family,trading range-based volatility estima-tors seems to have the potential forcapturing serial dependencies in thevolatility of asset prices.

One way of demonstrationing thepredictive power of trading-rangebased estimators is to compare theirACFs with the ACF of absolutereturns, calculated on consecutive clos-ing prices (see Figure 2). Interestingly, ittakes approximately the same numberof lags for both ACFs to become zero.However, the ACF of the differencebetween Highs and Lows decays at amuch slower rate. What explains thishigher degree of persistence? TA andits concept of support and resistanceprovides a ready explanation: High andLow prices do not only determine theintraday trading range but also oftencoincide with interday support andresistance levels; they are therefore lesslikely to change as frequently as closingprices.

Support and resistance levels whichare not breached during the course of atrading day retain their importance andprovide a proxy for the next day'svolatility. This volatility proxy stands a

fair chance of being superior to areturn-based volatility estimate if sup-port and resistance levels are also notbreached during the next trading day.

Forecasting pricesSo, does the trading range have any pre-dictive power for future closing prices?A study of lagged correlations of thetrading range and absolute returnsindeed seems to point to a causal rela-tionship between the trading range andfuture closing prices.

Note that a strong correlation of time

series A (trading range) with time seriesB (absolute return) at a positive lag,indicates information flow from timeseries A to time series B where theinformation manifests itself with a pos-itive lag. Since it is very likely that thereis a third common cause influencingthe behaviour of both series A and B,we cannot always assume direct causal-ity. However, even in the case of a thirdcommon cause, we have to accept thatthe information flow to series B isslower than that to series A.

Figure 3 shows a significant

Figure 2: ACF of absolute returns (solid line) and ACF of difference between High and Low (dashed line).exchange rate.

Figure 3: Lagged correlation function between daily trading range and daily returns. Thelower line shows the difference between positive and negative lags and indicates causality fromthe trading range to return-based estimates of volatility based on daily observations for USD-JPY.

→→


Subject Matters

asymmetry in the difference betweenpositive and corresponding negativelags of these two series; informationseems therefore to flow more slowly toa volatility measure consisting ofabsolute daily returns and indicates thata volatility measure based on the trad-ing range systematically predicts Close-to-Close volatility, at least within anintra-week time frame.Once we treatHigh, Low and Close prices as differenttime series, we can use cointegrationanalysis (Johansen, 1988) to explorestructural relationships suggested bytechnical indicators such as candle-sticks. While the empirical evidence infavour of cointegration between differ-ent asset prices is modest, we find thatHigh, Low and Close prices of thesame asset price are highly cointegrat-ed.

Further, we find evidence of Grangercausality (a technique for determiningwhether one time series is useful inforecasting another) between High,Low and Close prices. This spells goodnews for technical analysts: as cointe-gration establishes a linear relationshipbetween random walks, it should bepossible to exploit this structural rela-tionship to predict future prices frompast prices.

In Fiess and Macdonald (1999 and2002) we formally investigate the infor-mational content of Open, High, Lowand Close prices and their value in fore-

casting volatility, as well as future levelsof daily exchange rates. We find thatwith the help of dynamic modellingtechniques it is possible to identify andexploit structural relationships betweenHigh, Low and Close prices to derivedynamic out-of-sample forecasts overdifferent time horizons. These modelshave good out-of-sample forecastingperformance and perform well as trad-ing models. Perhaps most pleasingly theforecasting performance of these mod-els does not disappear when risk andtransaction costs are allowed for. Toillustrate, the US dollar/Yen exchangerate model shows good directionalforecasting ability over a forecastinghorizon of up to ten days and correct-ly predicts the direction of a one-dayahead exchange rate change with a57.3% success rate.

The existence of a stable structuralrelationship between High, Low andClose prices and the evidence of coin-tegration and Granger causality showsthat by following a modelling strategythat combines High, Low and Closeprices it is indeed possible to generatesuperior forecasts of volatility as well asfuture levels of asset prices. Technicalanalysis of high, low and closing pricesmay indeed present a crude but usefulway of exploiting any latent Grangercausality which exists in high frequencyexchange rates.

Norbert Fiess is an economics lec-turer at the University of Glasgow.([email protected]).

ReferencesJohansen, S. (1988): Statistical analysis ofcointegration vectors, Journal of EconomicDynamic and Control, 12, 231-254.

Fiess, N.; MacDonald, R.(1999): TechnicalAnalysis: A cointegration-based approach,Multinational Finance Journal, 3, 3, 147-172.

Fiess, N.; MacDonald, R.(2002): Towards theFundamentals of Technical Analysis,Economic Modelling, 19, 3, 353-374.

Takens, F. (1981): Detecting strange attractorsin turbulence, in: D. Rand, L. Young (Eds.),Dynamical Systems and Turbulence, Berlin.

“BY FOLLOWING A MODELLING STRATEGY THAT COMBINESHIGH, LOW AND CLOSE PRICES IT IS INDEED POSSIBLE TOGENERATE SUPERIOR FORECASTS OF VOLATILITY AS WELL

AS FUTURE LEVELS OF ASSET PRICES.”


Subject Matters

Japanese candlesticksMy colleagues and I conduct the firstknown robust test of candlestick chart-ing in a paper entitled "CandlestickTechnical Trading Strategies: Can TheyCreate Value for Investors", forthcom-ing in the Journal of Banking andFinance. We test all the commonly-usedsingle and multiple candlestick patternson the Dow Jones Industrial Averagecomponent stocks for the 1992-2002period.

We are particularly careful to apply arobust methodology. The traditionalapproach of comparing the returnsfrom a technical analysis signal to themean return using t-tests is often inad-equate because it does not account forwell known characteristics of stockreturn data such as skewness, leptokur-tosis (fat tails), autocorrelation, andconditional heteroskedasticity. Due tothis, we extend the superior bootstrap-ping methodology for use in an open,high, low, and close environment. Thisinvolves fitting a null model of stockreturns (e.g. GARCH-M) to each indi-vidual stock series (based on thereturns from close to close) and deter-mining each of the parameters. Theresiduals are then resampled and used,together with the model parameters, toform a random series of closing pricesthat has similar characteristics to theoriginal series. Random open, high,and low series are then formed. Thisprocess is repeated 500 times to gener-ate 500 random series.

If a candlestick rule produces moreprofit on the original stock series than itdoes on 25 or less of the random series(i.e. 500 series × 5%), then the candle-stick rule has statistically significantforecasting power at the 5% level.

We find that candlestick analysis does

not out-perform a buy-and-hold strate-gy. This result stands up to rigoroussensitivity analysis, such as two, five,and ten day holding periods and buyingat the opening price and the closingprice on the day after a signal. We con-clude that candlestick technical analysismay have value on DJIA stocks when itis combined with other technical analy-sis techniques but it is not profitablewhen used in isolation.

Simple trading rulesIn a recent article in the Journal ofFinancial Econometrics, Po-HsuanHsu and Chung-Ming Kuan ofColumbia University and AcademiaSinica respectively, examine the prof-itability of 39,832 simple trading rules,contrarian trading rules, and investorstrategies (these are strategies based onthe assumption that investors continu-ally change their trading rules accordingto which rules have been the mostprofitable in recent times) on Nasdaqcomposite, Russell 2000, DJIA, andS&P 500 index data for the 1989-2002period.

The authors pay particular attentionto the issue of data snooping. Lo andMacKinlay (1990) and many otherspoint out that by repeatedly examiningdifferent trading rules using the samedata sets, some rules may appear to beprofitable simply due to luck. Toaddress this concern, Hsu and Kuanapply a Reality Check (White 2000 andHansen 2004) which determineswhether there exists a superior rule in auniverse of rules.

Hsu and Kuan find that statisticallysignificantly profitable simple tradingrules and investor strategies are avail-able for relatively "young" markets(Nasdaq Composite and Russell 2000)

but not for more "mature" markets(DJIA and S&P 500). They then takethe common approach in the academicliterature and deduct 0.05% from thetransaction price for each one-waytrade - assumed to be a reasonable esti-mate of the costs faced by market mak-ers. After taking transactions costs intoaccount in this way, the authors findthat the 11 year average return for thebest trading rule on the Nasdaq is27.9% compared to a buy-and-holdreturn of 21.5%. The best trading ruleon the Russell 2000 index performseven better. The average return aftertransaction costs is 37.2% compare to abuy-and-hold return of 11.4%. Theyconclude that technical analysis cangenerate superior returns for theinvestor.

Ben Marshall is a Lecturer in theDepartment of Finance, Bankingand Property, Massey University,New Zealand. His research inter-ests include investigating the prof-itability of technical analysis tech-niques, with a focus on the applica-tion of rigorous statistical method-ologies. [email protected]

ReferencesHansen, P.R. (2004). A test for superior predictive abil-ity. Working Paper, Department of Economics, StanfordUniversity.Hsu, P-H. & Kuan, C-M. (2005). Re-examining theprofitability of technical analysis with White's RealityCheck and Hansen's SPA Test. Journal of FinancialEconometrics, 3(4), 606-628.Lo, A.W. & MacKinlay, A.C. (1990). Data-Snooping Biases in Tests of Financial Asset PricingModels. Review of Financial Studies, 3, 431-467.Marshall, B.R., Young, M.R., and Rose, L.C. (2005).Candlestick technical trading strategies: Can they createvalue for investors. Journal of Banking and Financeforthcoming. White, H. (1999). A reality check for data snooping.Econometrica, 68, 1097-1126.

TEST RESULTS FOR JAPANESE CANDLESTICKS AND SIMPLE TRADING RULES by Ben Marshall

BBRROOKKEERRSS CCOONNTTIINNUUEE TTOO OOFFFFEERR EEVVEERR GGRREEAATTEERR CCHHAARRTTIINNGG PPOOWWEERR TTOO TTHHEEIIRR CCLLIIEENNTTSS..

Berkeley Futures, one of the UK lead-ing independent brokerages, has nowupped the ante with its new BerkeleyIQ-Trader, a front-end system designedfor institutional and private tradersalike.

Berkeley IQ-Trader gives BerkeleyFutures' clients the ability to develop,backtest, optimise and automaticallyimplement technical trading strategiesof great complexity, on the world'sfutures and options markets withoutany need for programming knowledge.

As such, Berkeley IQ-Trader can beused to support a wide range of tradingstyles, including algorithmic trading,system trading, automated trading andscalping.

To explain the advantages of usingIQ-Trader, three key aspects of the sys-tem stand out: automation; chartinganalytics; backtesting & optimisation.

Simplicity of automationIQ-Trader is designed to perform com-plex operations that would normallyrequire a high level of programmingknowledge. This is all thanks to an easy-to-use programming language calledExpress. Whatever your trading strate-gy, everything can be done with a clickof a mouse. There is a choice of man-ual or automatic entry and you canplace orders directly from charts, createexit strategies with stops and bracket

orders, and use fully automated entriesand exits. Profit targets and stops canbe set up simultaneously with the order,or they can be automated for a setnumber of points, percentage gain, oreven ATR multiples. Another valuedfeature is 'TradeGuard', where you canset up multiple types of stops so, slip-page aside, you should never exceedyour predetermined loss levels. There iseven an 'emergency exit all positionsand cancel open orders' function.

Powerful charting analyticsA wide variety of charts and studies canbe used to build your trades. Powerfulcharting tools such as Bollinger Bands,candlesticks, and Fibonacci retrace-ments are all included. The functionali-ty is intuitive, allowing you to createindividual studies and layouts with fil-ters, drawing tools and trading aidscalled sentimentors. Information canbe dragged and dropped into charts,building up a trading profile and thenused to place orders directly from thecharts. Colour-coded indicators assistby identifying breakout, support andresistance levels (see Figure 1).

Sentimentors are an interesting and

very useful feature of the front-end.They can be described as a key indica-tor or a trading insight, and they pro-vide a 'sentiment' value in a numericalrange. The value is linked to a point intime for a particular underlying feature;these features could be interest rates, aspecific market, or perhaps businessnews. You can program the sentimen-

BERKELEY IQ-TRADER

It's common practice for bro-kers and spread-betting com-panies to offer non-execution

related services in an effort to pullin new customers. One of theways they do this is by offeringcharting analytics and trading soft-ware as a part of their onlineoffering.

Figure 1.

Figure 2.


Software

“I recommend this system to those who require a platform that can accu-rately trade moves in multiple markets with speed and efficiency in both

small and large volumes.”- Lisa Jarvis is an investment adviser at ABN Amro Morgans Ltd, and provides advice on

equities and options.

tors or create them manually (Figure 2),and you can combine a range of differ-ent sentimentors to form a so-calledmeta-sentimentor (Figure 3). The metasentimentor shapes the overall strategyor study, based on an analysis of all theelements, and stop techniques and fil-ters can be used to enhance it.

Backtesting and optimisationBacktesting enables the user to examinepotential trades using historical data,based on a given strategy, to see howthe trades might have performed in thepast. Once you have backtested a strat-egy and are satisfied it has performedwell, you can begin optimising it so thatyou have the most effective setting fora particular strategy. Another usefulfeature is walk-forward testing, whichallows you to continually re-optimise asnew data is collected.

SummaryTechnology that automates trades andminimises manual intervention is muchin demand today, particularly with thetide of new regulations promised myMifid (the EC's markets in financialinstruments directive). These regula-tions will change the way derivatives (aswell as equities and bonds) are traded in

Europe and place a premium on solu-tions that reduce operational risk andeliminate the errors that creep in withmanual processing. Berkeley IQ-Traderis therefore not just a sophisticatedfront-end system, it's also an applica-tion that is very much in tune with thetimes.

Berkeley Futures Ltd is based in theWest End of London and has beenoffering dealing services in deriva-tives since 1986.

Berkeley IQ-trader is part of a suite ofproducts provided by Berkeley On-linewhich is a trading name of BerkeleyFutures Ltd. Berkeley Futures Ltd isauthorised and regulated by theFinancial Services Authority. For moreinformation on Berkeley IQ trader orto register for a free demo please con-tact Marc Quinn on + 44 (0) 207 77584777 or by email at [email protected] please see their websitewww.bfl.co.uk.

Berkeley Futures will be exhibitingBerkeley IQ-Trader in conjunction withthe CBOT® at the Technical AnalystEuropean Conference 2006, London8th/ 9th February 2006.

Key Facts

IQ-Trader was originally released inJune 2005 by Patsystems, the inde-pendent software vendor. BerkeleyIQ Trader is the same front-end sys-tem, but connected to BerkeleyFutures online brokerage services.Three versions of Berkeley IQ-Trader are available to BerkeleyFutures clients: Standard, Advancedand Premier:

Standard Advanced chartingChart trade indicatorsChart order entryFills and orders visualised in chartsBracket ordersTrendline stopsLadder order entryOrder managementPaper tradingLatest market news24 hour online help facility & support

AdvancedAs standard, plusAutomated tradingTrading signals from indicatorsStrategy backtestingStrategy optimisationPosition management on signals

Premium As Advanced, plusInter-product spreadingStrategy scripting tool (Express)Strategy simulation tool

Data requirementsBerkeley Futures provides the real-time data and brokerage service,which includes access to UK,European, US and most worldwideFutures and Options exchanges.Users however will need to sub-scribe to either eSignal or Fides forthe historical data. Out of the twodata providers, Berkeley Futuresencourages connectivity to FIDES.

Figure 3.


Software

Berkeley Futures Ltd announce the launch ofBerkeley Futures Ltd Announce the Launch of

For more information on Berkeley IQ-Trader or the services that Berkeley Futures Ltd offer please contact Marc Quinn on +44 (0)207 758 4777 or by email at [email protected] or see our website, www.bfl.co.uk

• Strategy simulation tool• Position management on signals• Advanced charting• Chart trade indicators• Chart order entry• Fills and orders visible in charts

Berkeley Futures Ltd is authorised and regulated by the Financial Services Authority. Please note that dealing in equities, futures, options, foreign exchange and CFD’s are all areas of investment in which it is possible to lose money. The risks attached to dealing in off-exchange products such as foreign exchange and CFD’s differ from those attached to trading in on-exchange products. If you trade in any geared/contingent liability product it is possible to lose in excess of the funds you may have put in as your initial deposit. Investing in any of the products mentioned may not be suitable for you and if you are in any doubt you should consult your financial adviser.

Berkeley Futures Limited has been offering dealing services in Derivatives to institutions and individuals since 1986.

We deal in Futures, Options, CFDs, Bullion, Forex and Equities for Individuals, Corporates, Hedge Funds, Introducing Brokers and SIPPs.

Berkeley, in conjunction with CBOT®, will be exhibiting Berkeley IQ-Trader at the Technical Analyst European Conference 2006, London 8th/9th February.

• Bracket orders• Trendline stops• Ladder order entry• Order management• Paper trading

• Automated trading• Trading signals from indicators• Strategy backtesting• Strategy optimisation• Inter-product spreading• Strategy scripting tool

Jackson House, 18 Savile Row, London W1S 3PW


Mastering Technical Analysis By John C. Brooks McGraw Hill279 pages, £31.99ISBN 0-070144882-9

‘Mastering Technical Analysis’can be purchased from theTechnical Analyst bookshop. Toorder please call 01730 233870and quote "the Technical Analystmagazine".

Although many new books on technical analysis contain little that addsto the various publications by John Murphy and the like, some authorsattempt to combine an overview of traditional charting techniques with

new thinking. This kind of publication may go over old ground simply toremind readers of useful existing techniques. But it will also present some newideas that add real value and John. C. Brooks'new book falls firmly into this cat-egory.

Brooks has an impeccable pedigree as a technical analyst. A founder of theMarket Technicians Association (MTA) and the International Federation ofTechnical Analysts (IFTA), he is now senior analyst at Lowry's Reports, arespected US research house. With this book he has made an honest effort tocondense his 40 years of market experience into one volume that will be of realpractical value to the trader or investment manager. More than anything else,his book is a practical book on technical analysis and avoids the more theoret-ical approach taken by other publications. Brooks concentrates less on commontechniques such as chart patterns and trendlines but instead focuses more onmore using volume, flow of funds, market sentiment and stock selection.

As Brook's points out, volume should be a significant element in technicalresearch, especially as a trend confirmation signal, a warning that a trend maybe coming to an end and in the trading of various patterns. However, it doestend to be under discussed, perhaps because real-time volume data is not avail-able. Indeed, if "volume leads price" as Brooks emphasises, then even delayedexchange volume statistics will have value for the short-term trader. Moreover,Brooks argues that the introduction of ETFs and various derivative productsmeans volume data has more to offer the trade today than ever before.

Unusually for book on technical analysis, Brooks devotes a chapter to mone-tary policy or 'money flow'. Although this is a purely fundamental subject, mosttraders and investment managers do use a combination of technical and funda-mental factors in their analysis. Brooks discusses policy changes by the FederalReserve and draws attention to how a directional shift to higher rates by the Fedcan signal the "three steps and stumble rule". This states that after the third risein the discount rate (rather than the fed funds rate), the stock market stumblesinto a bear market. Furthermore, if the yield curve implies a stock market topwhen the yield curve inverts then an effective way of confirming a shift in thecurve is to divide long-term corporate paper yields by 3-month paper yields andapplying a moving average to the result. However, one might argue against theneed to apply a rate-of change indicator or a moving average to the fed fundsrate in order to distinguish a short term move from a long-term shift in policydirection.

John C. Brooks has written a highly readable and useful book that containsmany interesting ideas on using volume, sentiment indicators and relativestrength but it is not a book for the technical purist. Nevertheless, as the sub-ject of technical analysis continues to develop, more books of this sort can beexpected to appear on the bookshelves.

MASTERING TECHNICAL ANALYSIS

Book Review


COMMITMENTS OF TRADERS REPORT11 January 2005 - 10 January 2006Futures only (open interest) commercial and non-commercial net positions

10-year US Treasury Source: CBOT

Dow Jones Industrial Average Source: CBOT

5-year US Treasury Source: CBOT

Swiss franc Source: CME

Pound sterling Source: CME Yen Source: CME

-250000

-200000

-150000

-100000

-50000

0

50000

100000

150000

200000

11/01/2005 05/04/2005 28/06/2005 20/09/2005 13/12/20050

200000

400000

600000

800000

1000000

1200000

1400000Non commercial (LHS)Commercial

-300000

-250000

-200000

-150000

-100000

-50000

0

50000

100000

150000

11/01/2005 05/04/2005 28/06/2005 20/09/2005 13/12/20050

200000

400000

600000

800000

1000000

1200000

Non commercial (LHS)Commercial

-60000

-50000

-40000

-30000

-20000

-10000

0

10000

11/01/2005 05/04/2005 28/06/2005 20/09/2005 13/12/20050

10000

20000

30000

40000

50000

60000

70000

80000


-40000

-30000

-20000

-10000

0

10000

20000

30000

40000

50000

11/01/2005 05/04/2005 28/06/2005 20/09/2005 13/12/20050

10000

20000

30000

40000

50000

60000

70000

80000


-80000

-60000

-40000

-20000

0

20000

40000

11/01/2005 05/04/2005 28/06/2005 20/09/2005 13/12/20050

20000

40000

60000

80000

100000

120000

140000

160000

180000


-10000

-5000

0

5000

10000

15000

20000

11/01/2005 05/04/2005 28/06/2005 20/09/2005 13/12/20050

5000

10000

15000

20000

25000

30000

35000


Commitments of Traders Report


Commitments of Traders Report

Review of the Commitments of TradersReport released on January 10th 2006

In November the net commercial positions forcertain markets in the COT report were atextreme levels, suggesting we could see a ris-ing US stock market, falling longer term interestrates, a weakening dollar, and another up-leg incommodity prices led by the grains. This couldhave been viewed as a special opportunitybecause the interrelationship between the mar-kets supported the forecast. Indeed, lookingback over the last two months, this has indeedbeen the behaviour of the markets. Since theNovember issue, we have seen four year highsin the stock market, a turn towards lower longerterm interest rates, a weakening of the dollar,and another up leg in commodity prices includ-ing important lows in the grains.

In the current report those extreme net com-mercial positions in the currencies, interestrates, and grains have moved back to neutral.Markets tend to move back and forth betweenundervalued and overvalued, and so the bias inthose markets should be to continue the newtrend until they get to a new extreme. Technicaltrading tools should be used to confirm the con-tinuation of those new trends and we shouldconsider the possibility of retesting the recenttrend reversals. In October, the "adjusted com-bined" Stock Index Futures net commercial posi-tion was at the smallest net hedge in the lastfive years. In the three month up move in themarket this position moved to near the largestnet hedged position in the last five years. Thissuggests the US stock market has moved froman undervalued position to a perceived overval-ued position relative to the fundamentals.

George Slezak www.commitmentsoftraders.com

Euro Source: CME

Nasdaq Source: CME

Gold Source: CEI

-30000

-20000

-10000

0

10000

20000

30000

40000

11/01/2005 05/04/2005 28/06/2005 20/09/2005 13/12/2005

0

20000

40000

60000

80000

100000

120000


-20000

-15000

-10000

-5000

0

5000

10000

11/01/2005 05/04/2005 28/06/2005 20/09/2005 13/12/20050

10000

20000

30000

40000

50000

60000

70000

80000

90000


0

20000

40000

60000

80000

100000

120000

140000

160000

180000

200000

11/01/2005 05/04/2005 28/06/2005 20/09/2005 13/12/20050

20000

40000

60000

80000

100000

120000

140000



LONG-TERM TECHNICALSProvided by Thomas Anthonj, ABN Amro, Amsterdam

Whether the consolidation from 1.9550 has finally bottomed outcan't be confirmed at the moment. The least it would take for a big-ger rebound or a resumption of the former bull-trend would be abreak above strong resistance between 1.7820 and 1.8005.

GBP-USD: Chances for a minimum rally to 1.8595/1

Strong resistance at 121.20/122.35 triggered the expected setbackbut overshooting 114.40 raises the question whether this decline isjust corrective in nature or the resumption of the 2002-2004 beartrend. The key level is the 113.60 trend line support.

USD-JPY: Straight resumption of the up trend pos-sible as long as 113.60 holds

The market may have just finished a triangle consolidation patternthat could signal the bigger up-trend has already resumed targetingat least .7255. A break above .6925 would support this but only a tri-angle breakout at .7010 would confirm it.

EUR-GBP: Triangle consolidation still favoring anupside break

Unless Gold breaks the row of higher lows at 489.90, the focusremains clearly up. A break below 489.90 could trigger a setback to440.62/437.89, which looks to be the worst case scenario at themoment.

Gold: Roaring bull-trend with no sign of weaknessyet

Long-Term Technicals


The market leaves no doubt that it wants to reach the head-and-shoulders projection at 16800 before taking a temporary break. Afterthis, we see no reason why the market should not extend to theupside straight to 17950/18190.

Nikkei: Healthy uptrend but resistance at 16800 or17950 could trigger setback.

DJ EURO STOXX 50: Further gains expected but a resistance at 3685/3886

It looks at the moment the bulls are clearly ruling and once 65.30has been broken, the next targets are the next Fibonacci-projec-tions at 77.50 and 84.60 from where a consolidation could followagain.

Brent Crude Oil future: Straight resumption of theup-trend very likely

Failing to clear 1312/67 would increase the risk of having completedthe accumulation phase, which would see strong support on a breakbelow the trend channel support at 1191 and the last low at 1168. Asetback to 1094/61would then be the best case scenario.

S&P: Bull-trend intact but massive resistancebetween 1312 & 1367

Long-Term Technicals

Having reached the 50 % retracement of the 2000-2003 bear trendat 3685 we'd very much doubt that any setback could go lowerthan 3263/56 at the moment. If it does then we'd have to watch outfor a minimum setback to 2983.


EVENTS 2006

Technical Analysis InsightsOne day seminar

March09

Marriott Hotel, Zurich,

Switzerland

[email protected]+44 (0) 207 833 1441

www.ta-conferences.com

Short Term TradingTechniques

One day training course

March01

Cass Business School,London

[email protected]+44 (0) 207 833 1441

www.technicalanalyst.co.uk/training

Elliott Wave TheoryOne day training course

March02


[email protected]+44 (0) 207 833 1441


Market CyclesOne day training course

April04


[email protected]+44 (0) 207 833 1441


Behavioural Finance &Market Psychology

One day training course

April05


[email protected]+44 (0) 207 833 1441


DeMark IndicatorsOne day training course

April06


[email protected]+44 (0) 207 833 1441


The Technical AnalystIndia 2006Conference

May03

Taj Hotel,Mumbai,

India

[email protected]+44 (0) 207 833 1441


Date Event Venue Contact

The Technical AnalystEuropean Conference 2006

Glaziers Hall - London, 8th and 9th February 2006

Europe’s premier event for traders and investment managers


Contact: [email protected] Tel: +44 (0) 207 833 1441

Events

deka investment - the technical analyst · 2016-11-23 · at deka bank, discusses his methodical...

Documents