Askyourself:amIassessietythingshappening,orjustthealceofthetweh,giveeenhavehappened?
(e)Neersareoftenprodudergreattimepressure,soitisnthatsomeartitaiherearethreereportsthatshouldhavebeenspiked。
Whenallsixeggsinaboxboughtiweredouble-yolked,itwasclaimedthatatrulyastohadhappened。Ohousandhasthisproperty,sotheceofgettingaboxfulisthistinyfraultipliedtogethersixtimes。Theresultingnumberissosmallthatifyouopenedoneboxeverysed,youwouldexpecttotakeoverthirtybilliooeacrossaboxgonlydouble-yolkedeggs!
Butthatakessenseonlyifalltheeggsinaboxaredepelyfromavastofeggs,ohousandofwhichisdouble-yolked。Thisdoesn’thappen。Eggsaresortedbysizebefboxed。Someboxesareeveainingonlydouble-yolkedeggs。。。。
&ionsabouttheprivatelifeofanEnglaainsurfaced。Areporteruresforthe‘likelihood’ofeachoffourpossibleasbytheteammanager:
(a)expelhimfromtheplayingsquad–110
(b)retaininthesquad,butin–310
(inthesquad,butremovethecy–610
(d)takenoa–810
Ahesefourestimatedprobabilitiesisplausible。Butsietomutuallyexes,theirsummusty:however,these‘probabilities’addupto1。8。
Thirdly,itortedthat,amongpeoplewhohadwo£50,000oioheurningupmostoftenincludedJohn,David,Michael,Margaret,Susan,andPatricia。Sood:butitwasabsurdtoclaimthat thereforeyoushouldtrytoihthosenamesinyourLotterysyndicate!
Desgignorance
Drawingoneawell-shuffledpack,IexpectBladRedtobeequallylikely,andItlyattachthefigureof‘12’totheceofRed。Inaslightlyvolutedway,Icouldsaythatthe distributionIattachtotheceofgetties100%tothequantity‘12’–myceisshownbythechoiceof100%。IfIamsurethataprobabilityissomedefinitefigure,Iattachprobability100%tothatfigure。
Butfrequently,Iotseleglefigure:mybestestimateofthecemytrainwillmissitsaybe34,but23and45earlyasplausible,andeveotheextremesofzeroandupletelyruledout。Iuseauousdistributiefromzerotounitytodescribemyfeelingsaboutthisunknownprobability。
Totalignoranceofaprobability–veryrareces–wouldbedescribedbyusiinuousuniformdistributionofFigure5。Moreoften,thereissomeiefigurewhichisyourbestsiheprobability,andyourinstinctaboutbothhigherandlowervaluesisthattheircesfallawayture 8showsaofgraphsfromwhatiscalledthe betafamilyofdistributionsthathavethisproperty。
Figure 8airettyignorantofthevalueoftheprobability:weattachhighestbelieftovaluesnear12,butvaluesassmallas15has45arestillquitepossible;withFigure 8b,wearemuchmoretthatthevalueiscloseto12,whilestillnouttheextremes。With8c,hestbeliefisforvaluescloseto13,whilewith8dthevaluesareverystronglygregatednear23,withverylittleexpethatthevalueisbelow12。
Howfarwillmycartravelorol?,aeadyspeed,Iexpecttogetabout90miles,butifImakeshortjourneysoveracoupleofweeks,itwillbemorelike60miles。Ihedistancewillhavesomeuy,expressedviasomeuousdistributioahakindofdistributioiosmallcupsofsize10cc。Therewillbe1,000suchdthetotaldistancecoveredwillbethesumofthedistancesachievedusingthese1,000pohetralLimitTheorem,whichsaysthatthe sumenumberofrandomquaendtofollowaGaussiandistribution。
&afamilyofdistributions
8。ued
Forsteadymotorwaytravel,IwouldselectaGaussiandistributioredon90,sharplypeakedtoshowlowvariance。ForsporadictravelouldalsouseaGaussiandistributioredon60andreadtreateruy。
Utility
Yodmothermakesyouaonlyoffer。Eithershewillgiveyou£1,orshewilltossadifyoucallHeadsorTailscorrectly,shewillgiveyou£10,otherwiseyougetnothing。Whatake?
&ivesare£1for,orafifty-fiftygambleayoutof£5。Nearlyeverybodyprefersthelatter。Butscalethemoneyupbyafaillily,preferenge。Tohaveonemillionpouainisfarmoreattrabeingequallylikelytohavezero,ortenmillionpouof utilityliesbehindthisdifference。
Forsmallsumsofmoney,havingtwiceasmuchusually isworthtwiceasmuch,butifonemillionpoundswouldgeainlevelofpleasureforyouandyourfamily,doublethatamountwouldodoublethepleasure。Isyoueasurethe‘worth’ofamoaryamount,theusualshapeoftherelationshipwillfollowthatshownihegraphalwaysrises,initiallylikeastraightlihensteadilymoreslowly。
‘Utility’explainswhyahouseholderandaninsuranpanyagreethatasumof£250isareasonableaoinsureahouseworth£250,000againstperilssuchasfire,subsidence,。Thesubstaheinsuraneanthat,onanyindividualhouse,itactasthoughtheutilityisidenticaltotherelativelysmallsumsinvolved。Solohatitwillpayoutinanyyearislessthahousaedvalueforthistransaispositive,itrofit。Oheuninsuredhouseholderwouldfaousyoffinding£250,000ifshelostherhousetoohoseperils,sovoluntarilysurre£250toremovethatpossibilityisagooddealtohertoo。
9。Thegeneralshapeofautilitycurve
Takingainstbreakdownoftelevisions,microwaveovens,andsoonisalmostalwaysabadidea。Thesumsaremuchsmaller,utilityandmoiallythesame,andthepanyremiumlargeenoughforthemtoexpectaprofit。Insteadofbuyingthisexpensiveinsurance,builduparepairfundbyplagthehypotheticalpremiumintoyourbankat。Veryfewpeoplewillregretthatmove。
Ifyousuccessfullystructyourownutilityfun,youuseittohelpmakeachoitheavailableadersofuy。Forea,calculatethe expectedutilityoftheoutes,i。e。weighttheutilitiesbytheirrespeselectthataforwhichthisexpectedutilityisaslargeaspossible。
Thatistheprobabilist’suniversalreakiheavailablechoices,whatevertheces。