OnegoodexampleisGroundhogDay。

IfPunxsutawneyPhilseeshisshadowonFebruary2,therewillbesixmoreweeksofwinterweather。

WeallknowhowaccuratePhilis

As2007getsunderway,thestockindicatorgettingthemostpressthisweekistheFirstFiveDaysgauge。

Butbewarneditsalotofhype,anditjustdoesntwork。

Firstfivedaysindicator:popular,butuseless

TheFirstFiveDaysindicatorwaspopularizedbyYaleHirschsStockTradersAlmanac。

ItholdsthattheFirstFiveDaysofJanuarypredictsthedirectionofthemarketfortheremainderoftheyear。

Asproofoftheindicatorseffectiveness,itsproponentstracetheindicatorsperformancebackover55years,andstatethatof35FirstFiveDaysperiodsthatfinishedup,thestockmarketsubsequentlyfinishedupin30ofthoseyearsanimpressive85%winrateforthepredictor。

Consequently,venerablesourcessuchasTheNewYorkTimes,U。

S。

NewsWorldReport,CNNandMoneyMagazinehavequotedtheindicator。

Theproblemis:Itsauselessindicator,basedonshakylogicandevenshakierstatistics。

Andevenworse,itspotentiallydangeroustoyourwealth。

Firstfivedays:ignorethisinvestmentmyth

Letmebeblunt:TheFirstFiveDaysindicatoristhelowestformofanalysis。

Itstheoppositeofcause-and-effect。

Asyoureabouttosee,itsthekindofanalysisthatlooksforanycausetotietoanendeffect,regardlessoflogic,andstatisticalsupport。

ItsreallynomorevalidorusefulthanpredictingthestockmarketbasedonSuperBowlwinnersorgroundhogshadows。

Herearethreereasonswhy:

1。

TheLogicIsArbitrary:TherawnumbersforthisindicatorshowthatthemarkethasgonedownduringthefirstfivedaysofJanuary20timesoverthepast55years。

Inthose20occurrences,themarketfinishedtheyearup10timesanddown10times。

Lookingatthedata,theauthorsconcludethattheindicatorhasbeenright30outof35timesifthemarketbeginstheyearontheup。

Butwaitthismeansthattheindicatorhasnopredictivevalueifitstartsouttothedownside。

Thisworkinginonedirection,butnottheotheristooarbitraryformeandisaclassicexampleofwhereifthenumbersdontfitthehypothesis,thenyouchangethehypothesistofitthedata。

Thisiscurve-fittingmentality。

Doyouwanttoriskyourmoneybasedonthatlogic?

2。

TheTriggeringEventIsNotStatisticallySignificant:Thisindicatorstatesthatallyouneedtotriggerayearlongmarketpredictionisanymoveforfivedays。

Thismeansthattrivialmovesinthemarketcouldshapeyouroutlookforthecomingyear。

Butsupposethatafterthefirstfivetradingdays,themarketwasuponlyone-quarterofapoint。

Thiswouldstilltriggertheindicatorspredictionforanupyear。

Andtheproblemwithhavingamoveofanymagnitudetriggeranindicatoristhatatinymovelikethisdoesnttellusanythingaboutwhatthemarketisdoing。

Asmallmoveeitherupordownisjustrandompartofthebackgroundnoiseofthemarket。

Sohowdowedecidewhatismeaningfulandwhatisjustbackgroundnoise?

Onemeasurethatmanyanalystsuseistheaveragevolatilityofapricemovement。

Long-timereadersknowthatIusetheAverageTrueRange(ATR)ofpriceasameasureofvolatility。

Insimpleterms,theATRmeasurestheaveragesizeofthedailyrange(thehighminusthelow),whileaccountingforgapsbetweenbars。

IfwelookattheATRforafive-daymove,wedwantourtriggertomoveupordownbyatleasthalfoftheaverage。

Anythinglesswouldalmosthavetobeconsideredrandom。

Withthatinmind,yourindustriouseditordugdeepintothedetailsoftheFirstFiveDaysindicatorsrawdata。

IcalculatedtheSP500indexsATRforthelast20yearsandcheckedtoseehowmanyoftheFirstFiveDaystriggersignalscouldbeconsideredmorethanrandom。

Theanswer:Only4!

3。

TheSamplePopulationIsTooSmall:Whenweeliminatethetriggersignalsthataremerenoise,wenowonlyhave12to15triggersoftheindicatoroverthelast55years。

Thisisnotastatisticallysignificantsampletobaseanypredictionson,andthisindicatorisuncoveredasjustsomesimplisticcurve-fittingthatdoesntmeanathingfortradersandinvestors。

Thereisplentyofothergoodanalysisforyoutousetohelpguideyourtradingandinvestingdecisions。

Soitmakesalotofsensetothrowoutoverlysimplistic,statisticallymeaninglessonesliketheFirstFiveDaysindicator。

Andeventhoughtheindicatorworkedlastyear,dontfallintothetrapofassigninganexcessiveamountofmeaningtothemostrecentdatapoints。

Byallmeansuseitforcocktailpartydiscussions,butdontwasteanymoneytryingtouseittohelpyoumakesenseofthemarkets。

ByD。

R。

BartonJr。

QuantitativeAnalyst,Mt。

VernonResearchfortheSmartOptionsReport,www。

smartoptionsreport。

com

NakedTextNakedText