Finnegan Southey

Uncertainty in poker stems from two key sources, the shufed deck and an adversary whose strategy is un-known. One approach is to nd a pessimistic game theo-retic solution (i.e. a Nash equilibrium), but human play-ers have idiosyncratic weaknesses that can be exploited if a model of their strategy can be learned by observing their play. However, games against humans last for at most a few hundred hands so learning must be fast to be effective. We explore two approaches to opponent modelling in the context of Kuhn poker, a small game for which game theoretic solutions are known. Param-eter estimation and expert algorithms are both studied. Experiments demonstrate that, even in this small game, convergence to maximally exploitive solutions in a small number of hands is impractical, but that good (i.e. better than Nash or breakeven) performance can be achieved in a short period of time. Finally, we show that amongst a set of strategies with equal game theoretic value, in par-ticular the set of Nash equilibrium strategies, some are preferable because they speed learning of the opponent's strategy by exploring it more effectively