Tải bản đầy đủ - 0 (trang)
Hack 33. Predict the Length of a Lifetime

Hack 33. Predict the Length of a Lifetime

Tải bản đầy đủ - 0trang

privilegedplaceintheuniverse.

It'simportanttochoosesubjectsatordinary,unprivileged

moments.Biasingyourtestbychoosingsubjectsthatyou

alreadybelievetobenearthebeginningorendoftheirlife

spansuchasthehumanoccupantsofaneonatalwardora

nursinghomewillyieldbadresults.Further,Gott'sPrincipleis

lessusefulinsituationswhereactuarialdataalreadyexists.

Plentyofactuarialdataisavailableonthehumanlifespan

already,soGott'sPrincipleislessusefulhere.

Havingchosenamoment,let'sexamineit.Allelsebeingequal,

there'sa50percentchancethemomentissomewhereinthe

middle50percentofthephenomenon'slifetime,a60percent

chanceit'sinthemiddle60percent,a95percentchanceit'sin

themiddle95percent,andsoon.Therefore,there'sonlya25

percentchancethatyou'vechosenamomentinthefirstfourth

ofitslifetime,a20percentchanceit'sinthefirstfifth,a2.5

percentchanceit'sinthelast2.5percentofthesubject's

lifetime,andsoon.

Table3-11providesequationsforthe50percent,60percent,

and95percentconfidencelevels.Thevariabletpastrepresents

howlongtheobjecthasexisted,andtfuturerepresentshow

longitisexpectedtocontinue.

TableConfidencelevelsunderGott'sPrinciple



Confidencelevel



Minimumtfuture



Maximumtfuture



60percent



tpast/3

tpast/4



3tpast

4tpast



95percent



tpast/39



39tpast



50percent



Let'slookatasimpleexample.Quick:whoseworkdoyouthink

ismorelikelytobelistenedto50yearsfromnow,Johann



SebastianBach'sorBritneySpears'?Bach'sfirstworkwas

performedaround1705.Atthetimeofthiswriting,that's300

yearsago.BritneySpears'firstalbumwasreleasedinJanuary

1999,about6.5yearsor79monthsago.

ConsultingTable3-11,forthe60percentconfidencelevel,we

seethattheminimumtfutureistpast/4,andthemaximumis

4tpast.SincetpastforBritney'smusicis79months,thereisa

60percentchancethatBritney'smusicwillbeheardfor

between79/4monthsand79x4monthslonger.Inotherwords,

wecanbe60percentsurethatBritneywillbeaculturalforce

forsomewherebetween19.75months(1.6years)and316

months(26.3years)fromnow.



Sixtypercentisagoodconfidencelevelforquickestimation;notonlyis

itabetter-than-evenchance,butthefactors1/4and4areeasytouse.



Bythesametoken,wecanexpectpeopletolistentoBach's

musicforsomewherebetweenanother300/4and300x4years

atthe60percentconfidencelevel,orsomewherebetween75

yearsand1,200yearsfromnow.Thus,wecanpredictthat

there'sagoodchancethatBritney'smusicwilldiewithher

fans,andthere'sagoodchancethatBachwillbelistenedtoin

thefourthmillennium.



HowItWorks

Supposewearestudyingthelifetimeofsomeobjectthatwe'll

callthetarget.Aswe'vealreadyseen,there'sa60percent

chancewearesomewhereinthemiddle60percentofthe

object'slifetime(Figure3-4).3



Figure3-4.Themiddle60percentofthelifetime



Ifweareattheveryendofthismiddle60percent,weareat

thesecondpointmarked"now?"inFigure3-4.Atthispoint,

only20percentofthetarget'slifetimeisremaining(Figure35),whichmeansthattfutureisequaltoone-fourthoftpast(80

percent).Thisistheminimumremaininglifetimeweexpectat

the60percentconfidencelevel.



Figure3-5.Theminimumremaininglifetime(60

percentconfidencelevel)



Similarly,ifweareatthebeginningofthemiddle60percent

(thefirstpointmarked"now?"inFigure3-4),80percentofthe

target'sexistenceliesinthefuture,asdepictedinFigure3-6.

Therefore,tfuture(80percent)isequalto4xtpast(20percent).

Thisisthemaximumremaininglifetimeweexpectatthe

currentconfidencelevel.



Figure3-6.Themaximumremaininglifetime(60

percentconfidencelevel)



Sincethere'sa60percentchancewe'rebetweenthesetwo

points,wecancalculatewith60percentconfidencethatthe

futuredurationofthetarget(tfuture)isbetweentpast/4and

4xtpast.



InRealLife

Supposeyouwanttoinvestinacompanyandyouwantto

estimatehowlongthecompanywillbearoundtodetermine

whetherit'sagoodinvestment.YoucanuseGott'sPrincipleto

doso.Althoughit'snotpubliclytraded,let'stakeO'ReillyMedia,

thepublisherofthisbook,asanexample.



Icertainlydidn'tpickO'ReillyMediaatrandom,andplentyofhistorical

informationisavailableabouthowlongcompaniestendtolast,butlet's

tryGott'sPrincipleasarough-and-readyestimateofO'Reilly's

longevityanyway.Afterall,there'sprobablygooddataonthelongevity

ofBroadwayshows,butGottdidn'tshrinkfromanalyzingthemandI

hesitatetosaythatnowthatO'ReillyhaspublishedMindPerformance

Hacks,itsimmortalityisassured.



AccordingtotheWikipedia,O'Reillystartedin1978asa

consultingfirmdoingtechnicalwriting.It'sJuly2005asIwrite

this,soO'Reillyhasexistedasacompanyforapproximately27

years.HowlongcanweexpectO'Reillytocontinuetoexist?

Here'sO'Reilly'slikelylifetime,calculatedatthe50percent

confidencelevel:



Minimum

27/3=9years(untilJuly2014)



Maximum

27x3=81years(untilJuly2086)

Hereareourexpectationsatthe60percentconfidencelevel:



Minimum

27/4=6yearsand9months(untilApril2012)



Maximum

27x4=108years(untilJuly2113)

Finally,here'sourpredictionwith95percentconfidence:



Minimum



27/39=0.69years=about8monthsand1week(until

mid-March2006)



Maximum

27x39=1,053years(untilJuly3058)

Inthepost-dot-comeconomy,thesefigureslookprettygood.

Forexample,AppleComputer'saren'tmuchbetter,and

Microsoftwasfoundedin1975,sothesamecanbesaidforit.A

realinvestorwouldwanttoconsidermanyotherfactors,such

asannualrevenueandstockprice,butasafirstcut,itlooksas

thoughO'ReillyMediaisatleastaslikelytooutlivea

hypotheticalinvestorastotankinthenextdecade.



Endnotes

1. Ferris,Timothy."HowtoPredictEverything."TheNew

Yorker,July12,1999.

2. Gott,J.RichardIII."ImplicationsoftheCopernican

PrincipleforOurFutureProspects."Nature,363,May27,

1993.

3. Gott,J.RichardIII."AGrimReckoning."

http://pthbb.org/manual/services/grim.

RonHale-Evans



Hack34.MakeWiseMedicalDecisions



Medicaltestsprovidediagnosticscreeninginformation

thatisoftenmisunderstoodbypatientsand,sometimes,

evenbydoctors.Understandingtheprobability

characteristicscalled"sensitivity"and"specificity"can

provideamoreaccurateand(sometimes)reassuring

picture.

Asaconsumerofmedicalinformation,youhavetomake

decisionsaboutbehavior,treatment,seekingasecondopinion,

andsoon.Youlikelyrelyonmedicalinformationnewspaper

stories,yourdoctor'sadvice,testresultstomakethose

decisions.However,muchofthemedicalinformationyouget

fromyourdoctorhasaknownamountoferror.Thisisespecially

trueaboutdiagnostictestresultsthatindicatetheprobability

thatyouhaveacertaincondition.

Thishackisallaboutusinginformationaboutthecharacteristics

ofthosemedicalteststogetamoreaccuratepictureofreality

and,hopefully,makebetterdecisionsabouttreatment.



StatisticsandMedicalScreening

Tousemedicaltestinformationwisely,wehavetolearnjusta

bitaboutwhattheconceptofaccuracymeansforthesetests.

Thefourpossibleoutcomesofmedicaltests,intermsof

accuracy,areshowninTable3-12.

TablePossiblemedicaltestoutcomes







Patientactuallyhas Patientactuallydoes

nothavecondition(B)



thecondition(A)

Testresultindicatespatient

hascondition

Testresultindicatespatient

doesnothavecondition



Truepositive(scoreis

correct)

Falsenegative(scoreis

wrong)



Falsepositive(scoreis

wrong)

Truenegative(scoreis

correct)



Thereliability[Hack#6]ofmedicalscreeningtestsis

summarizedbytwoproportionscalledsensitivityandspecificity.

Essentially,thosewhorelyonthesetestsareconcernedwith

threequestionsofaccuracy:

Ifapersonhasthedisease,howlikelyisthepersonto

scoreapositivetestresult?Thislikelihoodissensitivity.Of

thosepeopleincolumnA,whatpercentwillreceivea

positivetestresult?

Ifthepersondoesnothavethedisease,howlikelyisthe

persontoscoreanegativetestresult?Thislikelihoodis

specificity.OfthosepeopleincolumnB,whatpercentwill

receiveanegativetestresult?

Ifapersonscoresapositivetestresult,howlikelyisthe

persontohavethedisease?Fromthepatient'sperspective,

thisistheultimatequestion,anditcanbethoughtofasthe

basicvalidityconcernwiththesetests.Doctor,canItrust

thesetestresults,orcouldtherebesomemistake?



NoticeinTable3-12thattherearedifferentpeopleincolumnsAandB.

PeoplewiththediseaseareincolumnAandpeoplewithoutthedisease

areincolumnB.IfyouareincolumnA,youcannotscoreafalse

positiveonthetest,becauseapositiveresultiscorrect.Ifyouarein

columnB,youcannotscoreafalsenegative,becauseanegativeresult

iscorrect.



Tài liệu bạn tìm kiếm đã sẵn sàng tải về

Hack 33. Predict the Length of a Lifetime

Tải bản đầy đủ ngay(0 tr)

×