Prisoner’s Dilemma
The Prisoner’s Dilemma (PD) is the canonical 2×2 game in which two players each choose cooperate or defect, and the payoffs are arranged so that defection strictly dominates cooperation for each player individually — yet mutual defection leaves both worse off than mutual cooperation would. It is the minimal formal model of the cooperation-and-defection tension.
Origin
The game was framed in 1950 by Merrill Flood and Melvin Dresher at the RAND Corporation, as part of early Cold War game-theory work on strategic interaction. Albert W. Tucker gave it the now-standard “two prisoners, separately offered a deal to testify” story and the name Prisoner’s Dilemma when presenting it to a Stanford psychology audience — the narrative stuck even though Flood and Dresher’s original experiments used a payoff matrix, not a prison.
Payoff structure
With the conventional ordering Temptation > Reward > Punishment > Sucker (T > R > P > S), and typically the additional constraint R > (T+S)/2 so mutual cooperation beats alternating exploitation:
| Cooperate | Defect | |
|---|---|---|
| Cooperate | R, R (reward) | S, T (sucker / temptation) |
| Defect | T, S | P, P (punishment) |
Defect is the dominant strategy: whatever the other player does, you score higher by defecting. Both players reasoning this way converge on the mutual-defection Nash equilibrium — P, P — which is Pareto-inferior to mutual cooperation R, R. That gap between individual rationality and collective outcome is the entire point.
One-shot vs. iterated
In a one-shot PD, defection is the unambiguous rational play. Everything interesting happens when the game repeats with an uncertain horizon — the iterated PD (IPD), where today’s defection can be punished tomorrow. This is the domain of evolutionary-game-theory and of Robert Axelrod’s famous tournaments (early 1980s): game theorists submitted strategies, played round-robins, and the simple Tit for Tat (TFT) — cooperate on move one, then copy the opponent’s previous move — won both tournaments. Axelrod attributed TFT’s success to four properties: it is nice (never defects first), retaliatory (punishes defection), forgiving (returns to cooperation after a single punishment), and clear (easy for opponents to read). See cooperation-and-defection for the broader evolutionary picture.
Real-world analogues
- Arms races — both sides arm (defect) though both would prefer disarmament (cooperate); the Cold War framing was exactly why RAND cared.
- Price wars / cartels — firms undercut each other to the mutual-defection of commodity pricing; OPEC is an attempt to hold the cooperate-cooperate cell.
- Tragedy of the commons — the N-player generalization: grazing, fishing, carbon emissions.
- Security — the attacker/defender dynamic and defections from cooperative security norms; see cooperation-and-defection.
- Everyday trust — any repeated exchange where short-term exploitation is tempting but long-term reciprocity pays.
Related
- cooperation-and-defection — the underlying tension, generalized
- evolutionary-game-theory — the iterated / population framing
- robert-axelrod — the tournaments and The Evolution of Cooperation
- norms / metanorms — how cooperation is stabilized socially
Sources
- Stanford Encyclopedia of Philosophy — Prisoner’s Dilemma
- Axelrod, R. — The Evolution of Cooperation (1984)
- Poundstone, W. — Prisoner’s Dilemma (1992)