Field Centipedes

摘要

Aumann (1998) showed that if the backward induction outcome is not played at some state of the world, then at that state there must be a node in the path of play at which the player whose turn it is to move deliberately chooses an action that he knows yields him a lower payoff than the one he would get by choosing an alternative action. Specifically, at that state there is a node that is reached along the path of play at which a player chooses to continue, even though he knows at the time of his choice that stopping is more profitable. Although this irrational behavior is by no means impossible among humans, our working hypothesis is that it is less likely to occur among chess players, who are familiar with backward induction reasoning. Further, their familiarity with this form of reasoning is common knowledge among many, if not all, humans.