The Prisoner's Dilemma
Two people can cooperate for mutual benefit or betray each other for individual gain. If both betray, both lose. What do you do? Does it matter if you'll meet again?
Formalized by Merrill Flood and Melvin Dresher at RAND Corporation in 1950, and named by Albert Tucker, the Prisoner's Dilemma is the foundational model for why cooperation breaks down. It has been used to explain everything from nuclear arms races to climate inaction to why people don't hold doors.
Tucker, A. W. (1950). A Two-Person Dilemma. Stanford University memorandum.
The setup
Two suspects are arrested and held separately. The prosecutor offers each the same deal:
- If you betray the other and they stay silent: you go free, they serve 5 years.
- If both betray each other: both serve 3 years.
- If both stay silent: both serve 1 year.
You can't communicate with your partner. What do you do?
Why rational actors defect
If the other person betrays you, you're better off betraying them (3 years vs. 5). If they stay silent, you're still better off betraying them (go free vs. 1 year). Betrayal dominates silence regardless of what the other person does.
So two rational, self-interested actors both betray, and both serve 3 years, when they could have served 1 year each by cooperating. Individual rationality leads to collective failure. That's the dilemma.
What changes with repetition
In a one-shot game, defection is the "rational" choice. But in repeated interactions, cooperation can emerge and stabilize. Robert Axelrod's famous tournaments in the 1980s showed that the winning strategy over many rounds was tit-for-tat: cooperate on the first move, then copy whatever the other player did last time.
The lesson: the shadow of the future changes everything. Long-term relationships, reputation, and the expectation of future interaction are what make cooperation viable.
The real-world echoes
The dilemma maps directly onto arms races (both countries would be safer with fewer weapons, but each is incentivized to build more), climate agreements (every nation benefits from others reducing emissions, but none wants to bear the cost alone), and salary negotiations (workers who share salary information help each other, but individuals fear the personal cost). The same structure, repeated across every domain where people need to cooperate but can defect.