In this paper, we present a model of rationality in extensive form games in which players, during the course of the game, may revise their conjectures about the opponents� preferences, while imposing common belief of rationality at each information set. Since players are assumed to be expected utility maximizers, their preferences, at each of their information sets, are given by a utility function over the reachable terminal nodes and some subjective probability distribution over the opponents� strategy choices compatible with this information set. A player may therefore revise his conjecture about an opponent�s utility function, or his subjective probability distribution, or both, whenever this opponent has chosen a move which would be suboptimal given the previously conjectured preferences.
© 2001-2024 Fundación Dialnet · Todos los derechos reservados