有限繰り返しゲーム

Finitely repeated games occur when the same game is played a specific number of times by the same participants, unlike one-time games played only once. This predetermined endpoint fundamentally changes player behavior. Knowing the game will end encourages players to adopt competitive strategies earlier than they might in indefinite scenarios. The finite nature creates strategic pressure that evolves as players approach the final round.

Backward induction is a method of solving games by working backward from the final round. Players anticipate that in the last round, there is no opportunity for retaliation or future cooperation, so they adopt competitive strategies. Knowing this, players expect competitors to act competitively in earlier rounds too, causing them to reduce cooperation progressively backward through all rounds.

Players lower prices or cut costs from the start because backward induction reveals that competitive behavior is inevitable by the final round. Anticipating this outcome, each player reasons that maintaining high prices or investments earlier is futile. This logic cascades backward through all rounds, causing both competitors to adopt cost-saving strategies immediately, even though sustained cooperation would yield higher collective profits.

In the final round, players are motivated to maximize immediate profits because no future rounds exist for retaliation or cooperation. Both competitors typically cut costs significantly or lower prices to capture more customers. Since there are no consequences for competitive behavior in the last round, players abandon any cooperative agreements they may have maintained in earlier rounds.

The finite nature of a game undermines cooperation because players know the relationship will end. As the final round approaches, the incentive to maintain cooperative agreements diminishes. Players recognize that competitors will eventually defect, so they defect first. This creates a race to competitive behavior that begins well before the final round, preventing sustained cooperation throughout the game.

Maintaining a high-price agreement throughout a finitely repeated game is extremely difficult because backward induction reveals that both players will eventually defect. Each player knows the other will cut prices in the final round to maximize profit, so neither trusts the agreement. This anticipation of future defection causes players to abandon the agreement and adopt competitive strategies from the start.

Game length directly influences when players shift to competitive strategies. In finitely repeated games, knowing the exact endpoint causes players to adopt cost-saving or competitive strategies progressively earlier. The shorter the game, the sooner competitive behavior emerges. This demonstrates that the finite nature and known duration of a game fundamentally shape the dynamic evolution of strategic decision-making between competitors.