Game Theory, Repeated Games, Computational Methods, Non-Nash Equilibria, Dominated Strategies

C62, C65, D74, D83, D84

The Conflict Analysis approach by Hipel and Fraser (1984) is well equipped to model repeated games. Players are assumed to posses a sequential reasoning that allows them to ( not necessarily correctly) anticipate the reaction of other players to their strategies. An individual?s best response strategy is thus defined based on this projection, adding additional stability conditions to strategic choice and increasing the set of potential equilibria beyond pure Nash equilibria. Yet, the original Conflict Analysis approach lacks the ability to genuinely model dynamic repeated games, in which past play defines the condition for future interactions. This article will illustrate how the original model can be adapted to include endogenous individual preferences that are defined by the strategic choice of players during past play, allowing to model the reciprocal connection between preferential change and best response play in repeated games. A dummy game serves as an exemplar and helps to visualise the results obtained from this extension.