Iterated Prisoners Dilemma Quiz

When the Prisoners Dilemma is repeated, players interact in future rounds where they can respond to past behavior. A player who defects today risks triggering retaliation in subsequent rounds, making defection potentially costly over the long run. This possibility of future punishment and reward introduces cooperation as a potentially rational strategy, transforming a game where defection is the only rational single-shot choice into one where sustained cooperation can emerge from rational self-interest.

A trigger strategy, specifically the grim trigger, specifies cooperation as long as the rival cooperates, with permanent retaliation upon any defection. The credible threat of permanent mutual defection deters the rival from defecting if the long-run cost of triggering the punishment exceeds the short-run temptation payoff. This mechanism explains how cooperation can be sustained in oligopoly markets and other repeated strategic interactions without formal binding agreements.

Backward induction applied to a finitely repeated Prisoners Dilemma unravels cooperation. In the final round, there is no future punishment to deter defection, so both players defect. Knowing this, both defect in the second-to-last round as well, and so on back to the first round. This complete unraveling means rational players defect in all rounds of a finitely repeated game, a result known as the backward induction paradox, which is why indefinitely or infinitely repeated games are needed to sustain cooperation through self-interest.

Tit-for-Tat cooperates on the first move and then mirrors the rival's previous action in every subsequent round. It rewards cooperation with cooperation and punishes defection with defection. Axelrod's famous computer tournaments showed that Tit-for-Tat, despite its simplicity, outperformed all other strategies submitted. Its success stems from being nice (starting with cooperation), retaliatory (punishing defection immediately), forgiving (returning to cooperation after the rival does), and clear (easy for the rival to understand and respond to).

The Folk Theorem establishes that in infinitely repeated games, the threat of future punishment can sustain cooperative outcomes as Nash equilibria when the discount factor is sufficiently high, meaning players value future payoffs enough. Specifically, any payoff that exceeds what a player could guarantee themselves through unilateral action can be sustained as a subgame perfect Nash equilibrium. This powerful result explains why long-run business relationships, international agreements, and repeated commercial interactions can sustain cooperation through purely self-interested behavior.

In the iterated game, a firm weighing defection compares the immediate temptation payoff against the long-run cost of triggering retaliation. If the present discounted value of future cooperative profits lost through a retaliatory price war exceeds the one-time gain from defecting, defection is individually irrational and cooperation is sustained. This is the economic logic behind tacit collusion in oligopoly markets: long-run self-interest supports cooperation when firms interact repeatedly and value future profits sufficiently.

Repetition enables cooperation through future punishment threats, the Folk Theorem formalizes this for infinitely repeated games, and Tit-for-Tat is the most studied iterated strategy based on reciprocal cooperation and retaliation. The claim that backward induction predicts cooperation in finitely repeated games is incorrect: backward induction unravels cooperation entirely in finitely repeated Prisoners Dilemmas, predicting defection in every round, which is the opposite of what the iterated setting hopes to achieve.

The discount factor reflects how much players value future payoffs relative to current ones. When the discount factor is high, future cooperation is nearly as valuable as current payoffs, making the long-run cost of triggering a punishment phase large. This high cost of future defection deters present defection. When the discount factor is low, players discount the future heavily, the temptation to defect today outweighs future consequences, and cooperation breaks down. Sustaining cooperation in repeated games requires a sufficiently high discount factor.

Axelrod's tournaments identified four properties that made Tit-for-Tat highly successful: it was nice (never defecting first), retaliatory (immediately punishing defection), forgiving (returning to cooperation once the rival cooperates again), and clear (simple enough that rivals could identify the strategy and respond rationally). These properties collectively make Tit-for-Tat effective at establishing and maintaining mutual cooperation while resisting exploitation, explaining its consistent success against more complex strategies in competitive iterated game simulations.

The iterated Prisoners Dilemma framework explains tacit collusion through the repeated game logic of credible punishment. When firms interact repeatedly, defecting from high pricing triggers retaliatory price cuts that eliminate future cooperative profits. If firms value future profits sufficiently, this threat makes defection unprofitable. Long-term relationships increase the likelihood that firms have high discount factors and multi-period horizons, creating conditions where cooperation is individually rational without any explicit agreement being necessary.

The Folk Theorem establishes that an infinitely repeated game (supergame) supports a vastly larger set of Nash equilibria than the corresponding single-shot game. Any individually rational payoff profile can be sustained as a Nash equilibrium in the supergame when the discount factor is sufficiently high. This means cooperative outcomes that are impossible to sustain as Nash equilibria in the single-shot Prisoners Dilemma can be self-enforcing equilibria in the infinitely repeated version through credible punishment threats.

End-game effects can unravel cooperation even in long-repeated games. As the anticipated final period approaches, backward induction logic applies: if both players expect the relationship to end in the next period, neither has future punishment to deter defection, so both defect. This anticipated defection unravels backward through earlier periods. In business contexts, this explains why firms sometimes behave cooperatively for years but become competitive as contracts near expiration or as industries face consolidation, consistent with the iterated game theory predictions.

In the iterated game framework, defecting from environmental cooperation triggers retaliation that destroys the cooperative equilibrium both countries value. The future cost of a regulatory race to the bottom, where both countries weaken standards and lose the environmental benefits of cooperation, can exceed the short-term industrial gain from unilateral defection. This repeated game punishment threat is self-enforcing without a formal treaty, explaining how informal international environmental cooperation can persist through rational self-interest alone.

Cooperation in the iterated Prisoners Dilemma is sustained by trigger strategies that make defection costly through future punishment, high discount factors that make future cooperative profits valuable enough to protect, and reciprocal strategies like Tit-for-Tat that maintain conditional cooperation. The finitely repeated structure with backward induction does the opposite: it unravels cooperation entirely from the last round backward, making it an obstacle to cooperation rather than a mechanism for sustaining it.

Axelrod's tournaments showed that Tit-for-Tat, submitted by Anatol Rapoport, won both the first and second computer tournaments despite being one of the simplest strategies entered. Its consistent success demonstrated that cooperation based on reciprocity, immediate retaliation, and rapid forgiveness outperforms both unconditional defection and elaborate complex strategies. This result has had wide influence on understanding how cooperation evolves in biology, economics, and political science through repeated interactions among self-interested agents.