Original title: Simple evolution drives direct reciprocity to maximum payoff in social dilemmas
Martin A Nowak · August 2026
Why do individuals help each other when cheating seems easier? Direct reciprocity means individuals interact repeatedly, letting them reward helpers and punish cheaters. However, whether cooperation wins depends on how evolution unfolds. Researchers tested simple evolutionary processes to see how cooperation spreads across different social dilemmas, which are scenarios where personal interests clash with mutual benefit. In their model, players adopt new strategies by comparing payoffs, while random mutations introduce strategies near the boundaries of possible behaviors. They found that this process drives populations to achieve maximum payoff in the Prisoner's Dilemma, Snowdrift, Stag Hunt, and Harmony games. When players remember two past rounds, they consistently reach maximum payoff, escaping traps that defeat simpler strategies.
Imagine you are choosing a route to work every day, and your choice affects both your commute time and your car's fuel efficiency. Normally, you pick what seems best for you alone, but if you and a neighbor commute together, you can share the carpool lane. The catch: you cannot plan everything upfront because your neighbor's choice changes every day based on your previous trip. In this paper, the repeated interactions between two players mirror that carpool partnership, and the simple evolutionary rule finds the best possible payoff for both, even when cheating looks tempting.
Cooperation is the bedrock of human society, but it is fragile when individuals can cheat. This paper shows that under surprisingly simple evolutionary rules, repeated interactions naturally push strategies to the highest possible payoff for everyone, across the classic cooperation dilemmas. That matters for understanding how altruism can arise and persist in nature and economics without complex enforcement.
The abstract does not specify population size thresholds, exact mutation rates, or selection intensity ranges for failure cases beyond noting 'large population sizes, intermediate to high mutation rates and intermediate to strong intensities of selection'. It does not test or claim applicability to repeated games beyond the four social dilemmas studied (Prisoner's Dilemma, Snowdrift, Stag Hunt, Harmony) or to asymmetric or multiplayer settings. Memory lengths beyond 2 are not explored. The 'Hold-trap' in Snowdrift for memory-1 strategies is identified but not analyzed in depth.
Ask anything about this paper. Answers come from its abstract, and say so when the abstract doesn't cover your question.
Simple evolution drives direct reciprocity to maximum payoff in social dilemmas — Martin A Nowak
Abstract licence: CC0
Explanation generated by Synthesis from the abstract — it is not the original paper. This is a preprint, so it hasn't necessarily been peer reviewed.