ALGORITHMICTRADINGCOURSE MODULE2 BACKTESTING&QUANTITATIVETRADING

2628March2011 0930hrs1715hrs ThomsonReuters,OneRafflesQuay Learnhowtocarryoutrigorousquantitativeanalysisofatradingstrategy Classsizeiscappedtotutorialgroupsize 50%fundingfromFinancialTrainingSchemegrants (basedonMASqualifyingcriteria)

TechnologyPartners

Introduction
Algorithmictradingofteninvolvestheuseofmathematicalmodelstodescribeandpredictmarketmovements.Thesemodelsare thenimplementedoncomputersystemsforautomaticexecution.Thejobofanalgorithmictraderistofirstdevelopamarket intuitionorideaofhowpricesshouldevolve.Usingmathematics,thetraderthenturnstheideaintoaquantitativemodelfor analysis,backtestingandrefinement.Whenthisquantitativemodelproveslikelytobeprofitableafterrigorousstatisticaltesting, thetraderimplementsthestrategyoncomputersystemsforexecution. Thisisa3dayintensivecoursedesignedtoprovideparticipantswithagoodunderstandingofthecoreconceptsandquantitative techniquesusedinthebacktestingandoptimizationofatradingstrategywithparticularemphasisonpairtradingandrelated strategies.ParticipantswilluseMATLABsoftwaretosolvebacktestingproblemsusingrealmarketdata.

Outcome
Attheendofthecourse,participantsareexpectedtodevelop:

anunderstandingofthecoreconceptsinquantitativetrading adeepappreciationoftheprocessofusingmathematicsandstatisticstoanalyzetheprofitabilityofatradingmodel handsonexperienceofhowbacktestingisdone anunderstandingofpairtradinginstocks,ETFs,futuresandcurrencies

HighlyRecommendedfor

Traderswishingtoapplytheirmathematicalandstatisticalstrengthsinthetradingarena Algorithmictradersseekingadeeperappreciationoftheroleofquantitativetraders Regulators,riskmanagersandauditorswhoneedagoodunderstandingofthenatureofquantitativeanalysis Anyonewhoaspirestobecomeaquantitativetrader

PreferredBackground

Someexperienceintradingispreferredbutnotessential Somebasicstatisticsbackground Someprogrammingexperienceispreferred

www.ntusgxcfe.ntu.edu.sg

Contents
Topic1: Introductiontobacktesting Whatisbacktesting? Theimportanceofbacktesting Thelimitationsofbacktesting:asurveyofcommonpitfalls Howtodecidewhethertobacktestastrategy:aseriesofexamples Criteriaforchoosingabacktestingplatform:prosandconsofvariousplatforms WhyisMATLABsuperiortoExcel/VBA/Java/C++/C#forportfoliotradingresearch?Overviewofcapabilitiesas researchandbacktestingplatform Topic2: MATLABTutorial

TheprosandconsofusingMATLABasautomatedtradingplatform Quicksurveyofsyntax Exercise:buildsomeutilitiesusefulfortradingresearch

Topic3: Nutsandboltsofbacktesting Backtestingasingleinstrument Exercise:buildatradingstrategyonES Exercise:computevariousperformancemetricsonESstrategy Exercise:computevariousperformancemetricsonESstrategywithrealistictransactioncosts Performancemeasurement:commonmetrics Transactioncosts:discussionofvarioussourcesoftransactionscosts Choosingahistoricaldatabase:importantpitfallstoavoid Reuterspresentationontheirdatabases Backtestingaportfolio Exercise:backtestalongshortportfoliotradingstrategy Exercise:findwaystoimproveonperformanceonlongshortportfoliotradingstrategy Exercise:testatechniqueforlookaheadbiasdetectiononamodifiedlongshortportfoliostrategy Strategyrefinement Waystoavoidlookaheadanddatasnoopingbiases Whyislivetradingperformanceusuallyworsethanbacktestperformance? Topic4: Kellyformula RiskmanagementusingKelly Exercise:howtoadjustyourportfoliosizebasedonrealizedP&L Exercise:allocatecapitalinaportfolioof3ETFsbasedonhistoricaldata CapitalallocationusingKelly

Topic5: Theoreticalfoundationofpairtrading Conceptofstationarity,andwhyitisuseful Conceptofcointegration,andwhyisituseful Howiscointegrationdifferentfromcorrelation? Howarestationarityandcointegrationdifferentfrommeanreversion? Testformeanreversion:computinghalflifebasedonOrnsteinUhlenbeckformula Whyiscomputinghalflifebetterthancomputingaverageholdingperiod? Exercise:computethehalflifeofmeanreversionforAUDCAD

www.ntusgxcfe.ntu.edu.sg

Topic6: Tradingapplicationsofstationarity Statisticaltestforstationarity:adf Exercise:useMATLABandspatialeconometricstoolboxtofindoutifAUDCADandEURCHFisstationary Exercise:backtestaBollingerbandstrategyforAUDCADandEURCHF

Topic7: Cointegrationandpairtrading Statisticaltestsforcointegration:cadfandJohansen Exercise:findoutifGLDGDXiscointegrating Exercise:backtestaBollingerbandstrategyonGLDGDXandotherpairs Exercise:computehalflifeofGLDGDXstrategy Findingthebesthedgeratio

Backtestvs.cointegration PastFuture Parameterlesspairtrading Exercise:constructaparameterlesspairtradingstrategy Stoploss? Tradingcointegratedtriplets Optionalexercise:testforcointegrationofatriplet,andconstructmeanrevertingstrategyonit Whatarethebestmarketstopairtrade?Prosandconsofeachmarket Automatedpairtrading Topic8: Relatedstrategies Indexarbitrage:Tradinganindexagainstabasketofitscomponentstocks Optionalexercise:backtestatradingmodelofXLEagainstitscomponents Statisticalarbitrage Momentumvs.meanreversal Momentumpairtrading:examples Otherstockmeanreversiontrades

Topic9: ReuterspresentationonQuantitativeResearch&Tradingworkflow

Trainers& Speakers
Dr.ErnestP.CHAN Dr. Ernest P. Chans career since 1994 has been focusing on the development of statistical models and advanced computer algorithmstofindpatternsandtrendsinlargequantitiesofdata.Hehasappliedhisexpertiseinstatisticalpatternrecognitionto projects ranging from textual retrieval at IBM Research, mining customer relationship data at Morgan Stanley, and statistical arbitrage trading strategy research at Credit Suisse First Boston, Mapleridge Capital Management, Millennium Partners, and MANEFundManagement. While in the Human Language Technologies group at IBM T. J. Watson Research Center (Yorktown Heights, NY), Ernest spearheadedIBMsresearchefforttodevelopasystemforsearchinglargetextdatabasessuchastheWorldWideWeb,catapulting IBMs reputation as a top player in the field. His system was placed seventh among some forty competitors in a competition sponsoredbytheNationalInstituteofScienceandTechnologyandtheDepartmentofDefensein1996.AttheDataMininggroup in Morgan Stanleys headquarter in New York, Ernest pioneered the application of some of these sophisticated statistical algorithmstothecomplextaskofextractingcustomerrelationshipsintheMorganStanleycustomeraccountsdatabase. Ernest was invited to join a proprietary trading group at Credit Suisse First Boston in New York in 1998 to develop statistical models for futures trading, stock pairtrading as well as trading based on earnings revisions, surprises and analyst recommendationchanges.HejoinedMapleridgeCapitalManagementCorp.in2002asaSeniorQuantitativeAnalystworkingon futurestradingstrategies,andthenMapleSecurities/MANEFundManagementInc.in2003asaseniorresearcherandtrader.
www.ntusgxcfe.ntu.edu.sg 3

Ernest consults for money management companies and also manages various accounts including EXP Quantitative Fund, L.P. which he cofounded. He has served as an expert witness in a matter related to algorithmic trading. He writes the Quantitative Trading blog which is syndicated to www.tradingmarkets.com and Yahoo Finance,and has published inthe Automated Trader magazine. He was quoted by the New York Times and CIO Magazine, and interviewed on CNBCs Closing Bell program and TechnicalAnalysisofStocksandCommoditiesmagazineontopicsrelatedtoquantitativetrading.HeistheauthorofQuantitative Trading:HowtoBuildYourOwnAlgorithmicTradingBusinesspublishedbyJohnWiley&Sonsin2008. ErnestholdsaBachelorofSciencedegreefromUniversityofTorontoin1988,graduatingwithHighDistinctionandreceivingthe Lieutenant Governors Gold Medal. He also holds a Master of Science (1991) and a Doctor of Philosophy (1994) degree in theoreticalphysicsfromCornellUniversity.Inrecognitionofhisexpertiseinstatisticaldatamining,hewasinvitedtoserveonthe Program Committees of the International Conference of Knowledge Discovery and Data Mining in 1998 and also of the SPIE ConferenceonDataMiningandKnowledgeDiscoveryin1999.HewasaninvitedpanelistonEffectiveArbitrageStrategiesatthe ETFEvolution2007Summit.HewasaninvitedspeakerattheAutomatedTradingconferenceinLondon,UK,inOctober2009.He conductsworkshopsontopicsfromPairTradingtoBacktestinginNewYork,LondonandHongKong. MrNeerajMASKARA NeerajMaskaraistheAsiaBusinessManagerforQuantitativeandEventDrivenTradingSolutionsatThomsonReuters.Thomson Reutersisthelargestglobaltechnologyandmarketdataproviderwithanendtoendsolutionforquantsandautomatedtraders, fromhostingfacilities,lowlatencymarketdatafeedstopricing,newsandfundamentalfeedsandanalyticsacrossglobalmarkets. ThomsonReutersisaleadingsolutionsprovidertobuysideandsellsidemarketplayerswhoareinvolvedinquantitativetrading. NeerajisresponsibleforthecontentandtechnologysolutionsofferedbyThomsonReutersacrossAsiaforhighfrequencytrading. Hehasextensiveexperienceworkingwitharangeofbuysideandsellsidequanttradingfirms/desksacrossEMEAandAsia.Prior tohisrolewithThomsonReuters,NeerajworkedforaquanthedgefundbasedinSwitzerland.

Venue
ThomsonReuters,OneRafflesQuay,#2801NorthTower,Singapore048583

FeesandRegistration
ProgramFees: NonSGXmember SGD4,230(excluding7%GST)beforefundingsupportfromFTSgrants* SGXMember SGD3,830(excluding7%GST)beforefundingsupportfromFTSgrants* EarlyBird1,Group2Discount&SGD3,380(excluding7%GST)beforefundingsupportfromFTSgrants* ParticipantsofATC13
1 2

earlybirddiscountissubjecttopaymentmadebefore11Feb2011 groupdiscountappliestogroupregistrationof5ormoreparticipants 3discountappliestoparticipantswhohaveattendedATC1withNTUSGXCFE Feesincludelunches,teabreaks,coursematerialsandtheuseofanindividualhandsontradingterminal. *TheMonetaryAuthorityofSingapore(MAS)administersFinancialTrainingScheme(FTS)grantstofinancialsectororganisations thatsponsoreligibleSingaporebasedparticipantstotrainingprogrammesthatmeetthequalifyingcriteria.Formoredetails,please visitwww.mas.gov.sg,orcontacttheMASat62299396orfsdf@mas.gov.sg. ToRegister: Logontowww.ntusgxcfe.ntu.edu.sg/courses.asp Closingdateforallregistrationsis4Mar2011 Allpaymentsmustbereceivedby11Mar2011 ForEnquiries: PleasecontactMsMichelleChahorMsAhneesahOmarattel:67905736/6078oremailnscfe@ntu.edu.sg

www.ntusgxcfe.ntu.edu.sg