Concepts · 3 min read
The Prisoner's Dilemma: Game Theory Basics With a Payoff Table

The payoff table (years in prison)
| Other stays silent | Other betrays | |
|---|---|---|
| You stay silent | 1, 1 | 3, 0 |
| You betray | 0, 3 | 2, 2 |
Lower is better. Each cell shows (your years, their years).
Best responses
If the other stays silent (C), your best reply is betray (0 year beats 1). If the other betrays (D), your best reply is betray (2 years beats 3). Betraying is best either way: a dominant strategy.
Nash equilibrium
When both players betray, neither can improve by changing alone. That stable outcome (2, 2) is the Nash equilibrium, even though (1, 1) would be better for both.
Why it matters
The same structure appears whenever individual incentives conflict with a shared benefit. Repeated games, reputation and enforcement can make cooperation rational. Compare with probability puzzles like Monty Hall.
What changes when the game repeats
In a single round betrayal is the dominant strategy. If the same two players meet many times, they can reward or punish each other, and cooperation can become sensible. A well-known strategy is tit-for-tat: stay silent in the first round, then copy whatever the other player did last time.
The table below simulates 10 rounds using the same payoffs as above (years in prison, so lower is better). Each row shows the total years for each player.
| Player 1 strategy | Player 2 strategy | Player 1 total years | Player 2 total years |
|---|---|---|---|
| Tit-for-tat | Tit-for-tat | 10 | 10 |
| Always betray | Always betray | 20 | 20 |
| Tit-for-tat | Always betray | 21 | 18 |
| Always silent | Always betray | 30 | 0 |
| Always silent | Always silent | 10 | 10 |
Two tit-for-tat players cooperate every round and serve 10 years each, while two permanent betrayers serve 20 years each. Against a permanent betrayer, tit-for-tat loses only the first round and then matches.
How to find an equilibrium in any payoff table
- For each opponent choice, find your best reply (underline it).
- Do the same for the other player.
- Any cell where both players have an underlined reply is a Nash equilibrium.
- Check whether a dominant strategy exists: a reply that is best against every opponent choice.
A game can have one equilibrium, several, or (in pure strategies) none. The prisoner’s dilemma has exactly one, and it is not the best outcome for the group, which is the point of the example.
Where the structure shows up
The same pattern appears whenever each side gains by defecting but all lose if everyone does. Common textbook examples are firms undercutting each other on price, neighbours over-using a shared resource, and countries deciding whether to limit emissions. Real situations are messier, with communication, repetition and enforcement, so the model is a thinking tool rather than a prediction.
Common misunderstandings
- "Betraying is always right." Only in a one-shot game with these payoffs.
- "Players are irrational." Each is individually rational; the problem is the incentive structure.
- "Cooperation needs trust." It can also come from repetition, reputation or rules that change the payoffs.
Probability also guides decisions under uncertainty. See probability rules and the Monty Hall problem for related reasoning.
Frequently asked questions
Is the dilemma realistic?
It is a simplified model. Real situations involve repetition, communication and enforcement that can change the outcome.
What is a dominant strategy?
One that is best for you no matter what the other player does.
What is tit-for-tat?
A strategy for repeated games: cooperate first, then copy the other player’s previous move.
Is there always a Nash equilibrium?
Every finite game has at least one, but it may require mixed strategies where players randomise.
Keep reading
The Monty Hall Problem Explained (Why Switching Wins 2/3)
The Monty Hall puzzle solved with a full enumeration table: why switching doors wins two times out of three and staying only one in three.
ConceptsStatistics vs Probability: What Is the Difference?
Probability predicts data from a known model; statistics infers the model from data. A coin-flip example computed both ways.
ConceptsThe Birthday Paradox: Why 23 People Is Enough
In a group of 23 people the chance two share a birthday is just over 50%. The full calculation and a table up to 70 people.
ConceptsStandard Deviation Explained With a Worked Example
What standard deviation measures, how to calculate it step by step, the sample vs population difference and the 68-95-99.7 rule.