A method by that players assume that the methods of their opponents are randomly chosen from some unknown stationary distribution. In every amount, a player selects her best response to the historical frequency of actions of her opponents. the method was initial noted by Julia Robinson who conjointly noted that the method converges to the equilibrium for two-player zero add games. whereas the method doesn't invariably converge in different settings, it's known that if it converges, then the purpose of convergence may be a Nash equilibrium of the sport.