Game TheoryOutlineAdminSlide 4Utility TheoryExpected UtilityHow to evaluateSlide 8ExampleRationalityZero-sumNon-zero sumDegenerate StrategyNash EquilibriumClassic ExamplesPrisoner's DilemmaRock-Paper-ScissorsNo dominant strategyMixed StrategyMixed Strategy 2Mixed Strategy 3DegeneraciesSlide 23Lie or Not LieExpectationNoteSlide 27BUTRepeated InteractionsDecision TreeFuture CostReducing degeneracyWednesdayGame TheoryRobin BurkeGAM 224Spring 2004OutlineAdminGame TheoryUtility theoryZero-sum and non-zero sum gamesDecision TreesDegenerate strategiesAdminDue WedHomework #3Due Next WeekRule AnalysisReaction papersGrades availableGame TheoryA branch of economicsStudies rational choice in a adversarial environmentAssumptionsrational actorscomplete knowledge•in its classic formulationknown probabilities of outcomesknown utility functionsUtility TheoryUtility theorya single scalevalue with each outcomeDifferent actorsmay have different utility valuationsbut all have the same scaleExpected UtilityExpected utilitywhat is the likely outcomeof a set of outcomeseach with different utility valuesExampleBet•$5 if a player rolls 7 or 11, lose $2 otherwiseAny takers?How to evaluateExpected Utilityfor each outcome•reward * probability(1/6) * 5 + (1/18) * 5 + (7/9) (-2) = -2/9MeaningIf you made this bet 1000 times, you would probably end up $222 poorer.Doesn't say anything about how a given trial will end upProbability says nothing about the single caseGame TheoryExamine strategies based on expected utilityThe ideaa rational player will choose the strategy with the best expected utilityExampleNon-probabilisticCake slicingTwo playerscutterchooser Cutter'sUtilityChoose bigger pieceChoose smaller pieceCut cake evenly½ - a bit ½ + a bitCut unevenlySmall piece Big pieceRationalityRationalityeach player will take highest utility optiontaking into account the other player's likely behaviorIn exampleif cutter cuts unevenly•he might like to end up in the lower right•but the other player would never do that•-10if the current cuts evenly,•he will end up in the upper left•-1•this is a stable outcome•neither player has an incentive to deviate Cutter'sUtilityChoose bigger pieceChoose smaller pieceCut cake evenly(-1, +1) (+1, -1)Cut unevenly(-10, +10) (+10, -10)Zero-sumNotefor every outcome•the total utility for all players is zeroZero-sum gamesomething gained by one player is lost by anotherzero-sum games are guaranteed to have a winning strategy• a correct way to play the gameMakes the game not very interesting to playto study, maybeNon-zero sumA game in which there are non-symmetric outcomesbetter or worse for both playersClassic examplePrisoner's DilemmaHold Out ConfessHold Out [-1, -1] [-3, 0]Confess [0, -3] [-5, -5]Degenerate StrategyA winning strategy is also calleda degenerate strategyBecauseit means the player doesn't have to thinkthere is a "right" way to playProblemgame stops presenting a challengeplayers will find degenerate strategies if they existNash EquilibriumSometimes there is a "best" solutionEven when there is no dominant oneA Nash equilibrium is a strategyin which no player has an incentive to deviatebecause to do so gives the other an advantageCreatorJohn Nash Jr"A Beautiful Mind"Nobel Prize 1994Classic ExamplesCar DealersWhy are they always next to each other?Why aren't they spaced equally around town?•Optimal in the sense of not drawing customers to the competitionEquilibriumbecause to move away from the competitoris to cede some customers to itPrisoner's DilemmaNash EquilibriumConfessBecausein each situation, the prisoner can improve his outcome by confessingSolutioniterationcommunicationcommitmentRock-Paper-ScissorsPlayer 2Rock Paper ScissorsPlayer 1 Rock [0,0] [-1, +1] [+1, -1]Paper [+1, -1] [0,0] [-1, +1]Scissors [-1, +1] [+1, -1] [0,0]No dominant strategyMeaningthere is no single preferred option•for either playerBest strategy(single iteration)choose randomly"mixed strategy"Mixed StrategyImportant goal in game designPlayer should feelall of the options are worth usingnone are dominated by any othersRock-Paper-Scissors dynamicis often used to achieve thisExampleWarcraft II•Archers > Knights•Knights > Footmen•Footmen > Archers•must have a mixed armyMixed Strategy 2Other ways to achieve mixed strategyIgnoranceIf the player can't determine the dominance of a strategy•a mixed approach will be used•(but players will figure it out!)CostDominance is reduced•if the cost to exercise the option is increased•or cost to acquire itRarityMixture is required•if the dominant strategy can only be used periodically or occasionallyPayoff/Probability EnvironmentMixture is required•if the probabilities or payoffs change throughout the gameMixed Strategy 3In a competitive settingmixed strategy may be called foreven when there is a dominant strategyExampleFootballthird down / short yardagehighest utility option•running play•best chance of success•lowest cost of failureButif your opponent assumes this•defenses adjustincreasing the payoff of a long passDegeneraciesAre not always obviousMay be contingent on game stateExampleLiar's Diceroll the dice in a cupstate the "poker hand" you have rolledstated hand must be higher than the opponent's previous rollopponent can either•accept the roll, and take his turn, or•say "Liar", and look at the diceif the description is correct•opponent pays $1if the description is a lie•player pays $1Lie or Not LieMake outcome chartfor next playerassume the roll is not good enoughRollerlie or not lieNext playeraccept or doubtExpectationKnowledgethe opponent knows more than just thisthe opponent knows the previous roll that the player must beat•probability of lyingNoteThe opponent will never lie about a better rollOutcome cannot be improved by doing soThe opponent cannot tell the truth about a worse rollIllegal under the rulesExpected UtilityWhat is the expected utility of the doubting strategy?P(worse) - P(better) When P(worse) is greater than
View Full Document