Tit-for-Tat
Tit-for-tat is a decision-making strategy that wins in iterated competitive environments by being simple, cooperative, and proportionally retaliatory. It was the consistent winner of Robert Axelrod's 1980 computer tournament of prisoner's dilemma strategies — defeating more complex, cunning programs submitted by leading game theorists.
The strategy:
- Start by cooperating
- Copy the opponent's last move — cooperate if they cooperated, defect if they defected
- Forgive immediately — as soon as the opponent cooperates, return to cooperation regardless of past defections
Why It Wins Without Winning
Tit-for-tat never wins any individual game. One-on-one, the best it can do is draw. Against a pure defector, it loses every round. Against another cooperator, they tie.
It wins the tournament because it cooperates with enough other players — accumulating points across many matchups — while not being exploited by consistent defectors. Strategies that "won" individual games often had defecting wars with other aggressive programs, producing mutual low scores. Tit-for-tat avoided those spirals.
Axelrod's summary of what makes it work:
"Its niceness prevents it from getting into unnecessary trouble. Its retaliation discourages the other side from persisting whenever defection is tried. Its forgiveness helps restore mutual cooperation. And its clarity makes it intelligible to the other player, thereby eliciting long-term cooperation."
The Four Properties
Nice — never defects first. This avoids escalations that would never have started.
Retaliatory — responds to defection immediately. This prevents exploitation; a strategy that can be walked on indefinitely will be walked on indefinitely.
Forgiving — holds no grudge past a single response. This keeps cooperation recoverable; a strategy that punishes indefinitely traps both parties in mutual defection.
Clear — the opponent can always predict it. Predictability enables the opponent to build a model and choose cooperation as the stable long-run equilibrium.
Limits
Game theory tournaments model interactions as discrete, symmetric, two-player games with defined payoffs. Real interactions involve many parties, asymmetric leverage, shifting goals, incomplete information, and human emotion. Tit-for-tat is a principle, not a formula.
In noisy environments (where misunderstanding causes accidental defection), forgiving two-thirds of defections — rather than all or none — performs better than strict tit-for-tat. The underlying principle survives: be nice, be firm, recover quickly, stay predictable.
Connections
- incentive-superpower — game theory is the study of incentive structures; tit-for-tat is what emerges when you optimize across repeated incentive interactions
- second-order-thinking — the strategy is not optimal in any single game; it optimizes for the full sequence, which requires reasoning ahead