Endogenizing the order of moves in matrix games

Players often have flexibility in when they move and thus whether a game is played simultaneously or sequentially may be endogenously determined. For 2 × 2 games, we analyze this using an extended game. In a stage prior to actual play, players choose in which of two periods to move. A player moving at the first opportunity knows when his opponent will move. A player moving at the second turn learns the first mover's action. If both select the same turn, they play a simultaneous move subgame.If both players have dominant strategies in the basic game, equilibrium payoffs in the basic and extended games are identical. If only one player has a dominant strategy or if the unique equilibrium in the basic game is in mixed strategies, then the extended game equilibrium payoffs differ if and only if some pair of pure strategies Pareto dominates the basic game simultaneous play payoffs. If so, sequential play attains the Pareto dominating payoffs. The mixed strategy equilibrium occurs only when it is not Pareto dominated by some pair of pure strategies.In an alternative extended game, players cannot observe delay by opponents at the first turn. Results for 2×2 games are essentially the same as with observable delay, differing only when only one player has a dominant strategy.     

On the Effect of Risk Aversion in Bimatrix Games

Nash equilibria with identical supports are compared for bimatrix games that are different with respect to the risk aversion of player 2. For equilibria in 2× 2-bimatrix games and for equilibria with efficient supports in coordination games it is established for which cases increased risk aversion of player 2 benefits or hurts player 2.     

Long-Term Behavior in the Theory of Moves

This paper proposes a revised Theory of Moves (TOM) to analyze matrix games between two players when payoffs are given as ordinals. The games are analyzed when a given player i must make the first move, when there is a finite limit n on the total number of moves, and when the game starts at a given initial state S. Games end when either both players pass in succession or else a total of n moves have been made. Studies are made of the influence of i, n, and S on the outcomes. It is proved that these outcomes ultimately become periodic in n and therefore exhibit long-term predictable behavior. Efficient algorithms are given to calculate these ultimate outcomes by computer. Frequently the ultimate outcomes are independent of i, S, and n when n is sufficiently large; in this situation this common ultimate outcome is proved to be Pareto-optimal. The use of ultimate outcomes gives rise to a concept of stable points, which are analogous to Nash equilibria but consider long-term effects. If the initial state is a stable point, then no player has an incentive to move from that state, under the assumption that any initial move could be followed by a long series of moves and countermoves. The concept may be broadened to that of a stable set. It is proved that every game has a minimal stable set, and any two distinct minimal stable sets are disjoint. Comparisons are made with the results of standard TOM.     

A theory of forward induction in finitely repeated games

A forward induction solution for finitely repeated games with complete information is developed. This notion is motivated in terms of its implications on the way deviations affect the opponents' expectations about the future behavior of the deviating player. We argue that the inability of the notion of perfect equilibrium to take account of forward induction is a key factor responsible for a number of difficulties encountered in the use of perfect equilibria in repeated games. It is then shown that the solution proposed in this paper remedies some of these problems in the study of three important classes of repeated games: (i) finitely repeated coordination games; (ii) repeated games where one long-term player plays a sequence of short-term players; (iii) repeated battle of the sexes games.     

Cycles versus equilibrium in evolutionary games

Mixed-strategy equilibria are typically rather unstable in evolutionary game theory. “Monocyclic” games, such as Rock–Paper–Scissors, have only mixed equilibria, some of which are “stable” in the sense that sequential best replies lead to them; yet, even these games are prone to stable cycles under discrete-time simultaneous best replies, giving an unusual equilibrium-selection problem. This article analyzes such games in a random-utility setting where changing strategies is costly, and the speed of the dynamic is, thus, endogenous. The stochastically stable outcome is determined by the cost of switching strategies; when switching costs are high, mixed equilibria are selected, whereas when switching costs are low, cycles are selected.     

K–Player Additive Extension of Two-Player games with an Application to the Borda Electoral Competition Game

In this note we introduce the notion of K–player additive extension of a symmetric two-player game and prove a result relating the equilibria in mixed strategies in the two games. Then we apply the result to the Borda electoral competition game.     

Staying power in sequential games

Staying power is the ability of a player to hold off choosing a strategy in a two-person game until the other player has selected his, after which the players are assumed to be able to move and countermove sequentially to ensure their best possible outcomes before the process cycles back to the initial outcome and then repeats itself (rational termination). These rules of sequential play induce a determinate, Paretosuperior outcome in all two-person, finite, sequential games in which the preferences of the players are strict.In 57 of the 78 distinct 2 × 2 ordinal games (73 percent), it makes no difference who the (second-moving) player with staying power is, but in the other 21 games the outcome is power-dependent. In all but one of these games, staying power benefits the player who possesses it.If no player has staying power, the outcomes that result from sequential play and rational termination are called terminal; they coincide with staying power outcomes if they are Pareto-superior. Normative implications of the analysis for rationally justifying cooperation in such games as Prisoners' Dilemma and Chicken, and implementing Pareto-superior outcomes generally, are also discussed.We are grateful to D. Marc Kilgour for very valuable comments on an earlier version of this paper, causing us to rethink and redefine staying power. The earlier version was presented at the Seventeenth North American Conference, Peace Science Society (International), University of Pennsylvania, November 9–11, 1981.     

Two-Speed Evolution of Strategies and Preferences In Symmetric Games

Agents in a large population are randomly matched to play a certain game, payoffs in which represent fitness. Agents may have preferences that are different from fitness. They learn strategies according to their preferences, and evolution changes the preference distribution in the population according to fitness. When agents know the preferences of the opponent in a match, only efficient symmetric strategy profiles of the fitness game can be stable. When agents do not know the preferences of the opponent, only Nash equilibria of the fitness game can be stable. For 2 × 2 symmetric games I characterize preferences that are stable.Jel Codes: C72, A13     

Outsourcing with identical suppliers and shortest-first policy: a laboratory experiment

We study experimentally in the laboratory two 2-player games that mimic a decentralized decision-making situation in which firms repeatedly outsource production orders to multiple identical suppliers. The first game has a unique (inefficient) equilibrium in mixed strategies, while the second game has two (efficient) equilibria in pure strategies and an infinite number of (inefficient) equilibria in mixed strategies. In both games, the optimal social costs can also be obtained via dominated strategies. We find that only in the second game subjects manage to reach an efficient outcome more often when matched in fixed pairs than when randomly rematched each round. Surprisingly, this is because subjects coordinate on dominated strategies (and not an efficient pure strategy equilibrium). We show theoretically that preferences for efficiency cannot explain our experimental results. Inequality aversion, on the other hand, cannot be rejected.     

A differential approach to the repeated prisoner's dilemma

The repeated prisoner's dilemma game is converted into a differential game by assuming that the players, instead of making decisions individually for each repetition of the prisoner's dilemma game, make decisions on the ratio of cooperative and noncooperative games that they wish to play over the next few moves, and that the actual plays are then determined using this ratio and a randomizing procedure. Although it sounds like a significant departure, this assumption is probably not too different from reality.Since each player can always obtain by his own action at least the payoff which he would receive from the noncooperative-noncooperative pair of strategies, that part of the differential game which is dominated by such a dual noncooperative strategy for either player is an unlikely outcome. This dominated area can be readily computed for any game, including those with more than two players.Formal testing with empirical data was impossible because of uncertainty about the proper null hypothesis. Nevertheless, experimental results reported by Rapoport are consistent with the theory.The authors are, respectively, Graduate Research Asistant in Economics and Professor of Economics and Public Choice at Virginia Polytechnic Institute and State University.     

Ellsberg games

In the standard formulation of game theory, agents use mixed strategies in the form of objective and probabilistically precise devices to conceal their actions. We introduce the larger set of probabilistically imprecise devices and study the consequences for the basic results on normal form games. While Nash equilibria remain equilibria in the extended game, there arise new Ellsberg equilibria with distinct outcomes, as we illustrate by negotiation games with three players. We characterize Ellsberg equilibria in two-person conflict and coordination games. These equilibria turn out to be related to experimental deviations from Nash equilibrium play.     

Punishing greediness in divide-the-dollar games

Brams and Taylor 1994 presented a version of the divide-the-dollar game (DD), which they call DD1. DD1 suffers from the following drawback: when each player demands approximately the entire dollar, then if the least greedy player is unique, then this player obtains approximately the entire dollar even if he is only slightly less greedy than the other players. I introduce a parametrized family of 2-person DD games, whose “endpoints” (the games that correspond to the extreme points of the parameter space) are (1) a variant of DD1, and (2) a game that completely overcomes the greediness-related problem. I also study an n-person generalization of this family. Finally, I show that the modeling choice between discrete and continuous bids may have far-reaching implications in DD games.     

LEXICOGRAPHIC ADDITIVITY FOR MULTI-ATTRIBUTE PREFERENCES ON FINITE SETS

Payoff dominance, a criterion for choosing between equilibrium points in games, is intuitively compelling, especially in matching games and other games of common interests, but it has not been justified from standard game-theoretic rationality assumptions. A psychological explanation of it is offered in terms of a form of reasoning that we call the Stackelberg heuristic in which players assume that their strategic thinking will be anticipated by their co-player(s). Two-person games are called Stackelberg-soluble if the players' strategies that maximize against their co-players' best replies intersect in a Nash equilibrium. Proofs are given that every game of common interests is Stackelberg-soluble, that a Stackelberg solution is always a payoff-dominant outcome, and that in every game with multiple Nash equilibria a Stackelberg solution is a payoff-dominant equilibrium point. It is argued that the Stackelberg heuristic may be justified by evidentialist reasoning.     

Non-strict ordinal 2 × 2 games: A comprehensive computer-assisted analysis of the 726 possibilities

Game theory has provided many tools for the study of social conflict. The 2 × 2 game has been found to be a particularly useful model. This paper describes the enumeration and analysis of all 726 distinct 2 × 2 games. A computer is used to generate the complete set, and a wide variety of maximin, equilibrium and stability calculations is performed for each player for every outcome in every game. The resulting data set is of great value for both the modeling and analysis of social conflict.     

Stubbornness,Power, and Equilibrium Selection in Repeated Games with Multiple Equilibria

Axelord’s [(1970), Conflict of Interest, Markham Publishers, Chicago] index of conflict in 2 × 2 games with two pure strategy equilibria has the property that a reduction in the cost of holding out corresponds to an increase in conflict. This article takes the opposite view, arguing that if losing becomes less costly, a player is less likely to gamble to win, which means that conflict will be less frequent. This approach leads to a new power index and a new measure of stubbornness, both anchored in strategic reasoning. The win probability defined as power constitutes an equilibrium refinement which differs from Harsanyi and Selten’s [(1988), A General Theory of Equilibrium Selection in Games, MIT Press, Cambridge] refinement. In contrast, Axelrod’s approach focuses on preferences regarding divergences from imaginary outmost rewards that cannot be obtained jointly. The player who is less powerful in an asymmetric one-shot game becomes more powerful in the repeated game, provided he or she values the future sufficiently more than the opponent. This contrasts with the view that repetition induces cooperation, but conforms with the expectation that a more patient player receives a larger share of the pie.      

Binary 2 × 2 games

The 2 × 2 game is the simplest interactive decision model that portrays concerned decision makers with genuine choices. There are two players, each of whom must choose one of two strategies, so that there are four possible outcomes. Binary 2 × 2 games are 2 × 2 games with no restrictions on the players' preference relations over the outcomes. They therefore generalize the strict ordinal 2 × 2 games and the ordinal 2 × 2 games, classes which have already been studied extensively. This paper enumerates the strategically distinct binary 2 × 2 games. It also identifies important subsets defined by the number of pure Nash equilibria and the occurrence of dominant strategies.     

Double deception: Two against one in three-person games

This article examines deception possibilities for two players in simple three-person voting games. An example of one game vulnerable to (tacit) deception by two players is given and its implications discussed. The most unexpected findings of this study is that in those games vulnerable to deception by two players, the optimal strategy of one of them is always to announce his (true) preference order. Moreover, since the player whose optimal announcement is his true one is unable to induce a better outcome for himself by misrepresenting his preference, while his partner can, this player will find that possessing a monopoly of information will not give him any special advantage. In fact, this analysis demonstrates that he may have incentives to share his information selectively with one or another of his opponents should he alone possess complete information at the outset.     

On Loss Aversion in Bimatrix Games

In this article three different types of loss aversion equilibria in bimatrix games are studied. Loss aversion equilibria are Nash equilibria of games where players are loss averse and where the reference points—points below which they consider payoffs to be losses—are endogenous to the equilibrium calculation. The first type is the fixed point loss aversion equilibrium, introduced in Shalev (2000; Int. J. Game Theory 29(2):269) under the name of ‘myopic loss aversion equilibrium.’ There, the players’ reference points depend on the beliefs about their opponents’ strategies. The second type, the maximin loss aversion equilibrium, differs from the fixed point loss aversion equilibrium in that the reference points are only based on the carriers of the strategies, not on the exact probabilities. In the third type, the safety level loss aversion equilibrium, the reference points depend on the values of the own payoff matrices. Finally, a comparative statics analysis is carried out of all three equilibrium concepts in 2 × 2 bimatrix games. It is established when a player benefits from his opponent falsely believing that he is loss averse.     

A Banzhaf share function for cooperative games in coalition structure

A cooperative game with transferable utility–or simply a TU-game– describes a situation in which players can obtain certain payoffs by cooperation. A value function for these games assigns to every TU-game a distribution of payoffs over the players. Well-known solutions for TU-games are the Shapley and the Banzhaf value. An alternative type of solution is the concept of share function, which assigns to every player in a TU-game its share in the worth of the grand coalition. In this paper we consider TU-games in which the players are organized into a coalition structure being a finite partition of the set of players. The Shapley value has been generalized by Owen to TU-games in coalition structure. We redefine this value function as a share function and show that this solution satisfies the multiplication property that the share of a player in some coalition is equal to the product of the Shapley share of the coalition in a game between the coalitions and the Shapley share of the player in a game between the players within the coalition. Analogously we introduce a Banzhaf coalition structure share function. Application of these share functions to simple majority games show some appealing properties.     

Game Trees For Decision Analysis

Game trees (or extensive-form games) were first defined by von Neumann and Morgenstern in 1944. In this paper we examine the use of game trees for representing Bayesian decision problems. We propose a method for solving game trees using local computation. This method is a special case of a method due to Wilson for computing equilibria in 2-person games. Game trees differ from decision trees in the representations of information constraints and uncertainty. We compare the game tree representation and solution technique with other techniques for decision analysis such as decision trees, influence diagrams, and valuation networks.