In 1980 Robert Axelrod invited game theorists to submit strategies for a repeated Prisoner’s Dilemma tournament. The winner was the simplest entry of all: tit-for-tat, four lines of code that cooperate first, then copy the opponent. Nice, retaliatory, forgiving, and clear.
The roots of cooperation
Where does cooperation come from among selfish players? Axelrod’s tournaments gave a striking answer: not from saintliness but from repetition and reciprocity. The same math explains why cartels can sustain high prices, and why they collapse when the future stops mattering.
In the repeated PD, grim trigger sustains cooperation when delta is at least...
The folk theorem says many outcomes are equilibria as long as they beat the...
Predict, then play
Run an Axelrod-style setup: pick your strategy against a grim-trigger bot, then read the round-robin table, where every strategy meets the whole field and a copy of itself. It opens at a patient ; drag patience down and watch the ranking turn over.
Why did that happen?
Take a guess, then reveal the reason
Against a one-shot opponent, defecting is dominant. Repeated play changes that: a strategy that provokes lasting retaliation pays for its greed. Tit-for-tat thrives because it earns the cooperative payoff with fellow cooperators and refuses to be exploited by defectors, while never starting fights. Cartels rely on the same logic: firms sustain high prices only because cheating triggers a price war that wipes out the short-run gain. When patience is high, the threat holds; when the future is discounted heavily, or the end is near, cooperation unravels.
The real episodes
Axelrod ran round-robin and evolutionary tournaments of repeated-PD strategies. Tit-for-tat won both, and successful strategies shared four traits: nice (never defect first), retaliatory (punish defection), forgiving (return to cooperation), and clear (easy to read). Real cartels are the same game: collusion on high prices is sustained by the threat of a price war, and it holds only when firms value future profits enough. The Joint Executive Committee railroad cartel of the 1880s alternated collusion and price wars exactly as the repeated-game model predicts.
The condition is the familiar one: grim trigger sustains cooperation when for the standard payoffs. Tit-for-tat, whose punishment lasts one round rather than forever, needs more patience: . Patient firms cooperate; impatient ones defect. Cartels break down when the short-run gain from undercutting beats the discounted stream of future collusive profits, which is why entry, demand shocks, or an approaching end date destabilize them. The folk theorem adds that many collusive prices, not just one, are self-enforcing.
Work it out
Why nice beats nasty over the long run
A purely exploitative strategy defects whenever it can grab the temptation payoff. Against a single opponent it may win, but in a population it provokes retaliation and ends up mired in mutual defection, earning the low punishment payoff round after round. Tit-for-tat, by contrast, reaches mutual cooperation with anyone willing to cooperate, and cuts its losses against defectors after one round.
Cartels show the same arithmetic. Cooperating (holding price high) forever is worth ; undercutting once grabs now but triggers a price war paying thereafter. Cooperation holds when . A firm tempted to cheat weighs a one-time gain against a permanent loss of the collusive premium, and only patience keeps it in line. When the future dims, the cartel cracks, just as tit-for-tat cannot sustain cooperation against a very impatient partner.
Key idea
Cooperation among selfish players comes from repetition and reciprocity, not virtue. Tit-for-tat, nice, retaliatory, forgiving, and clear, wins Axelrod’s tournaments, and the same logic sustains cartels only while the future matters enough.
Check yourself
Sources
- Axelrod, R. (1984). The Evolution of Cooperation. Basic Books.