Why study strategic thinking when with enough experience in a stationary environment, even amoebas—or human reinforcement learners, who need not even know that they are playing a game—usually converge to equilibrium? Many applications of game theory involve situations with no clear precedents. (Should you sell U.S. airline stocks when the market re-opens after 9/11, or buy them on the anticipation that others will overreact?) Comparative statics and design questions inherently involve new games with new equilibria, which players cannot reach by copying behavior from analogous games. In such situations subjects’ initial responses are often plainly "strategic" but nonetheless deviate from equilibrium. Even in settings in which players can be expected to converge to equilibrium, the structure of strategic thinking can influence the rate of convergence and equilibrium selection