Causal Assessment in Finite-length Extensive-Form Games

We consider extensive-form games in which the information structure is not known and ask how much of that structure can be inferred from the distribution on action profiles generated by player strategies. One game is said to observationally imitate another when the distribution on action profiles generated by every behavior strategy in the latter can also be generated by an appropriately chosen behavior strategy in the former. The first part of the paper develops analytical methods for testing observational imitation. The central idea is to relate a game’s information structure to the conditional independencies in the distributions it generates on action profiles. We present a new analytical device, the influence opportunity diagram of a game, describe how such a diagram is constructed for a given finite-length extensive-form game, and demonstrate that it provides, for a large class of economically interesting games, a simple test for observational imitation. The second part of the paper shifts the focus to the influence assessments of players within a game. A new equilibrium concept, causal Nash equilibrium, is presented and compared to several other wellknown alternatives. Cases in which causal Nash equilibrium seems especially wellsuited are explored.

[1]  J. Pearl.  Causality: Models, Reasoning and Inference , 2000 .

[2]  Robert J . Aumann,et al.  28. Mixed and Behavior Strategies in Infinite Extensive Games , 1964 .

[3]  Eddie Dekel,et al.  Hierarchies of Beliefs and Common Knowledge , 1993 .

[4]  E. Kalai,et al.  Subjective Games and Equilibria , 1993 .

[5]  Judea Pearl,et al.  Probabilistic reasoning in intelligent systems - networks of plausible inference , 1991, Morgan Kaufmann series in representation and reasoning.

[6]  E. Stacchetti,et al.  Towards a Theory of Discounted Repeated Games with Imperfect Monitoring , 1990 .

[7]  David J. Spiegelhalter,et al.  Probabilistic Networks and Expert Systems , 1999, Information Science and Statistics.

[8]  Judea Pearl,et al.  Equivalence and Synthesis of Causal Models , 1990, UAI.

[9]  Paul R. Milgrom,et al.  Rationalizability, Learning, and Equilibrium in Games with Strategic Complementarities , 1990 .

[10]  Pierfrancesco La Mura.  Game Networks , 2000, UAI.

[11]  Daphne Koller,et al.  Multi-Agent Influence Diagrams for Representing and Solving Games , 2001, IJCAI.

[12]  John Nachbar.  Prediction, optimization, and learning in repeated games , 1997 .

[13]  Pierpaolo Battigalli,et al.  Rationalizability in infinite, dynamic games with incomplete information , 2003 .

[14]  D. Madigan,et al.  A characterization of Markov equivalence classes for acyclic digraphs , 1997 .

[15]  Wolfgang Spohn,et al.  Stochastic independence, causal independence, and shieldability , 1980, J. Philos. Log..

[16]  J. Harsanyi.  Games with Incomplete Information Played by 'Bayesian' Players, Part III. The Basic Probability Distribution of the Game , 1968 .

[17]  G. Mackey.  Borel structure in groups and their duals , 1957 .

[18]  David Easley,et al.  Rational Expectations and Rational Learning , 1993 .

[19]  Judea Pearl,et al.  GRAPHOIDS: A Graph-based logic for reasoning about relevance relations , 1985 .

[20]  John C. Harsanyi,et al.  Games with Incomplete Information , 1994 .

[21]  D. Fudenberg,et al.  Self-confirming equilibrium , 1993 .

[22]  A. W. Tucker,et al.  Advances in game theory , 1964 .

[23]  E. Kalai,et al.  Subjective Equilibrium in Repeated Games , 1993 .

[24]  P. Reny,et al.  On the Strategic Equivalence of Extensive Form Games , 1994 .

[25]  Bert Fristedt,et al.  A modern approach to probability theory , 1996 .

[26]  Dan Geiger,et al.  Identifying independence in bayesian networks , 1990, Networks.

[27]  Judea Pearl,et al.  Fusion, Propagation, and Structuring in Belief Networks , 1986, Artif. Intell..