Chapter9Curiositiesanddilemmas
Atthebeginningofthisbook,Isomeaspectsofprobabilityappear,atfirstsight,todefyonsense。Exampleshaveariseoryhasunfolded。Herearesomeotherceswhereintuitionbemisleading,but,withsuffitcare,theseapparenttradibeexplaiofprobabilityiswhollyfreefromrealparadoxes。Butalthoughideasofprobabilityhelpusmakesensibledeayalsofihinkingabouttheprobabilitiesofeventsmightleadtounfortabledilemmas。
Parrondo’sparadox
GrahamGreeakesAllisafibasedonafalsepremise:thatthereissomeclevermathematibisoewheeltogivetheplayeranadvaherthanthehouse。Orary:mathematicshasprovedthat,whenallindividualbetsfavourthehouse,nobinatiournmattersroundandfavourtheplayer。Sorry,folks。
JuanParrondohasshownthatyouhavetobeveryprehowyouformulateageneralclaimthat,whesfavouroisimpossibletobihattheothersidehasanadvantage。Idescribehereamodifiofhisidea,duetoDeanAstumian,whodescribedasimplegameplayedontheboardwithfiveslots,shownihisisname。Itwasstructedmerelytomakethispoint。)
11。TheboardforAstumian’sgame
Youneedsomewayarawilloe:perhapsabagwith99WhiteballsandoneBlackball,oraspiisequallylikelytoetorestosonehuobeginthegame,plaomarked‘Start’。Everymovewilltakethetoke,orht,andyouwiokenreabeforeithitsLose。
&wobasicsetsofrules,callthemAhAndy,fromStartyoualwaysmovetoLeft,andfrht,youalwaysmovetoWi,youusethespiogivea1%ovingtoLose,anda99%ovingbacktoStart。WithBert,thespinarttogivea99%ht,anda1%gtht,
youalwaysreturntoStart,whilefromLeft,itisthesameasinAndy–thespinnergivesa1%ovingtoLose,a99%cetoStart。
Analysisofthesegamesissimple。InAndy,thereisnoprovisioht;youshufflearouailrandomcetakesyoufromLefttoLose。I,youusuallyshufflebetweenStartandRight,withoalvisitstoLeft。Eventually,ohesesojour,randomcetakesyoutoLose。IheceWiniszero。
Forthenewgame,Chris,youalsoneedafair。Atea,tossthis:ifitshowsHeads,usetherulesofAndy,ifitshowsTails,usetherulesofBert。
ItturnsoutthatyChrisexceeds98%!Itiseasytoseewhyyfavourite:ifeveryougettoLeft,youarelylikelytoreturyofStart。FromStart,youplayBerthalfthetime,withits99%ceofgettingtht;andinRight,youplayAndyhalfthetime,iablywinning。
FollowiherA,you mustlose:flipbetweerandom,andyouwiime!Framiheoremthatexcludesexampleslikethis,butsthatGreerestsonshakyground,requiresverypreguage!
2+2=4,or2+2=6?
SupposewecarryoutBernoullitrialswithafair,i。e。eachtoss,ily,isequallylikelytobeHeadsorTails。AtypiewillbeHHTHTTTHT。。。。ThemeaossesuntilHeadsappearsistwo;butwhatisthemeaossesu,orHH?
&iveanswerisfour,asweexpecttosforthefirstsymbol,thehrowsforthesedthemeahrowsuisindeedfour,butthisisnotthecaseforHH。Toseethatpattern,themeahrowsissix!
Thereasonforthediffere,togetHT,itiscorrecttuethatweexpecttotaketwettheH,theogettheTthatpletesthepattern。AndTwoplusTwoequalsFour。ButforHH,afterwehavethefirstH,thehrowwillbeThalfthetime,abeginagain–allthrowsuptothatpointwillhavebeehealgebraleadiaheAppendix。
&weenHandT,eachisequallylikelytoappearfirst;whataboutbetweenHHandHT?Again,eachisequallylikelytoarisebeforetheother,siwaitforthefirstHead,ahehrowdetermiheanswer。However,betweenHHaeristhreetimesaslikelytoappearfirst!Thereasonissimple:thesequencewillbeginwithHHohetime,butuhishappens,itisiHappearsfirst。(Thinkaboutit。)
Thegame Peeisbasedontheaboveideas。YouinviteyouroppoapossibletripleslikeHHT,orTHT,etc。,thatmightothreesecutivethrowsofafair。Youselectadiffereralpersoherepeatedly,andthewihepersoripleisseenfirst。Despitetheappareyofallowingyouroppoohavefirstpick,thisgamefavoursyou–ifyouknowwhaty。Whatevershechooses,youselectatriplethatearbeforehersatleast23ofthetime!ThewinniheAppendix。
Givemeaclue…