The Iterated Prisoner’s Dilemma

The Prisoner’s Dilemma was formalised by Merrill Flood and Melvin Dresher at the RAND Corporation in 1950 and named by Albert Tucker. Two players, each forced to choose between cooperation and defection without consulting the other, each receive a better individual outcome by defecting — and a collectively worse outcome when both do. The puzzle is sharper still when the game is iterated: played many rounds against the same opponent, each player must decide whether and when to trust. In 1979–1980, the political scientist Robert Axelrod ran two computer tournaments inviting strategies from researchers worldwide. The winner of both, submitted by Anatol Rapoport, was Tit-for-Tat — cooperate on the first move, then do what the opponent did last. The result was so striking that Axelrod’s The Evolution of Cooperation (1984) became one of the most-cited works in twentieth-century social science: four lines of code had outperformed elaborate strategies because they embodied being nice, retaliatory, forgiving, and clear. Later work has complicated the picture without erasing the wonder. The iterated Prisoner’s Dilemma is a place where game theory, evolutionary biology, and moral philosophy meet on a single grid.

Merrill Flood and Melvin Dresher, RAND Corporation (1950); the name coined by Albert Tucker in a Stanford lecture the same year. John von Neumann and Oskar Morgenstern, Theory of Games and Economic Behavior (1944), provided the surrounding framework. Robert Axelrod, The Evolution of Cooperation (Basic Books, 1984). Martin Nowak and Karl Sigmund, “Tit-for-Tat in Heterogeneous Populations,” Nature 355 (1992). William Press and Freeman Dyson, “Iterated Prisoner’s Dilemma contains strategies that dominate any evolutionary opponent,” PNAS (2012).