An experimental analysis of subgame perfect play: The entry deterrence game
Our experiments model a two-stage, two person non-cooperative game where subjects face a sequence of potential entry situations. Payoffs and entry costs are common knowledge. The subgame perfect equilibrium entails player I choosing an entry barring output and player E not entering. While many subjects played this way, a significant proportion of E players entered when it yielded negative net payoffs, and a non-trivial proportion of I players didn't seek deterrence. While these proportions fall over the course of the experiment, such behavior persists through the final period. Past experience influences I subjects' tendencies to seek deterrence.