A game-theoretic learning process where players iteratively update strategies by best-responding to the empirical distribution of opponents' past actions.