The Evolution of Cooperation
The book in brief. Robert Axelrod ran a tournament that showed empirically how repeated games change the equilibrium. When the same players meet again and again, with memory and a future that matters, cooperation emerges spontaneously.
Key ideas
- The tournament. Researchers submitted strategies for the Iterated Prisoner’s Dilemma, which were then run against each other in round-robin.
- Tit-for-tat wins. The winning strategy cooperates by default, retaliates against defection and forgives once behaviour is corrected, and it beat entries written specifically to defeat it.
- The horizon drives cooperation. It is the optimisation horizon, not a moral position, that produces the cooperative equilibrium.
Why it matters to my work. In Pareto Frontiers and Lock-In: When Mathematical Sub-Optimisation Is the Right Strategy I use Axelrod to read Nvidia’s refusal of surge pricing. A hyperscaler deals with Nvidia every cycle and remembers every prior interaction, so the function being optimised is cumulative payoff over many cycles and not the margin on this one. The same logic applies to the Nvidia and TSMC relationship.
