★ Free forever — no signup

The Axelrod Tournament Simulator 🏆

In 1980, Robert Axelrod invited game theorists to submit computer strategies for a repeated Prisoner's Dilemma. Everyone expected something clever and ruthless to win. Instead, the shortest program submitted — Tit for Tat — beat them all. Rebuild that tournament here: pick your entrants, add mistakes, and watch cooperation evolve.

1. Choose your entrants

Every strategy plays every other one (and a clone of itself) in a round-robin.

2. Set the rules

Longer matches reward strategies that build cooperation.

Chance a move is flipped by accident. Noise punishes grudge-holders.

Successful strategies reproduce; losers die out.

Why this tournament changed science

The Prisoner's Dilemma has an uncomfortable answer when you play it once: whatever the other player does, you personally do better by betraying them. Two rational players therefore both betray, and both walk away with less than if they had simply cooperated. For decades that looked like a proof that selfishness wins.

Robert Axelrod's insight was to change one thing: play it again and again. When the same two players meet repeatedly, today's betrayal costs you tomorrow's cooperation. He ran an open computer tournament in 1980, and then a bigger rematch with entrants who had studied the first results. Both times the winner was Anatol Rapoport's Tit for Tat — four lines of code that cooperates first and then copies its opponent.

Axelrod identified four traits shared by the strategies that did well. They were nice (never the first to defect), retaliatory (they punished betrayal instead of being exploited), forgiving (they returned to cooperation quickly), and clear (simple enough that opponents could learn to trust them). Complicated strategies that tried to outwit everyone mostly confused their partners into mutual defection.

The mistake slider matters too. In the real world people misread each other, emails get lost, and moves come out wrong. Under noise, a pure Tit for Tat can get locked into an endless revenge spiral with a copy of itself, and Grudger destroys itself over a single accident. Forgiving strategies such as Tit for Two Tats and Generous Tit for Tat take the crown instead — a decent argument for giving people the benefit of the doubt.

These results are still used to explain cooperation among animals, trade between rival nations, why repeat-business firms behave better than one-off sellers, and the evolution of reciprocity in humans. Everything in this simulator runs locally in your browser, and the results are reproducible — same settings, same outcome.

Experiments worth running

  • Run Tit for Tat vs Always Defect alone, then add Always Cooperate. Watch how the presence of a pushover changes who wins.
  • Set mistakes to 5% and compare Grudger with Tit for Two Tats.
  • Drop rounds to 10. Short relationships favour betrayal — this is the "shadow of the future" shrinking.
  • Select only mean strategies. Notice how low everyone's scores become when nobody cooperates.
  • Run the evolution with Detective and Prober included: exploiters thrive early, then starve once their food supply goes extinct.

FAQ

What is an Axelrod tournament?

It is a round-robin competition where computer strategies play the repeated Prisoner's Dilemma against each other, first organised by political scientist Robert Axelrod in 1980.

Why does Tit for Tat win?

It is nice, retaliatory, forgiving and clear, so it earns steady mutual-cooperation payoffs without being exploited.

Does Tit for Tat always win here?

No. Add a mistake rate or change the field of entrants and more forgiving strategies such as Tit for Two Tats or Generous Tit for Tat usually overtake it.

Is the simulator free?

Yes — it is completely free, requires no account, and runs entirely in your browser.

Keep going