We develop an adaptive learning game to rethink efficient markets. We use the stochastically stable state of this game, which is a mixed Nash equilibrium, to form an adaptive expectation model that provides an estimate of the confidence interval for prices on the next day. The estimate is most...