Cooperation, Retaliation and Forgiveness in Revision Games

Revision game is a very new model formulating the situation where players can prepare and revise their actions in advance before a deadline when payoffs are realized. We identify the Limited Retaliation (LR) strategy for revision games which sustains a high level of mutual cooperation and is robust to players' occasional mistakes. The LR strategy stipulates that, (1) players first follow a recommended cooperative plan; (2) if anyone deviates from the plan, the LR player retaliates by using the defection action for a limited duration; (3) after the retaliation, the LR player returns to the cooperative plan. The LR strategy has two good features. First, it is vengeful, in the sense that it deters the opponent from non-cooperative action by threatening a retaliation. Second, it is forgiving, because it returns to cooperation after a proper retaliation. The vengeful feature makes it constitute a subgame perfect equilibrium, while the forgiving feature makes it tolerate occasional mistakes. These are in clear contrast to the existing strategies for revision games which all assume players are extremely grim and never forgive. Besides its contribution as a new robust and welfare-optimizing equilibrium strategy, our results about LR strategy can also be used to explain how easy cooperation can happen, and why forgiveness emerges in real-world multi-agent interactions.