The prisoner's dilemma is one of the most studied games in game theory: each player is better off defecting, but if both defect, both end up worse off than if they had cooperated. When it is repeated many times with the same person, as happens in almost any real negotiation, the question changes: which strategy works best in the long run?
Political scientist Robert Axelrod organized a computer tournament to find out, and published the results in 1980. One of the simplest strategies won: Tit for Tat.
Here you play 10 to 14 rounds against a hidden opponent using a classic strategy, several of which were entered in Axelrod's tournaments. At the end you'll see:
- Which strategy your opponent was using.
- Which profile describes the way you play.
- Where your way of playing would place in a simulated tournament against the classic strategies.
All possible results
Reciprocal (tit for tat)
You start by cooperating, respond to defections and go back to cooperating as soon as the other player does. It's the pattern that won Axelrod's tournaments.
See the detail
What this profile describes
You play like Tit for Tat, Anatol Rapoport's strategy that won both of Axelrod's tournaments: you cooperate from the start, return a defection and don't hold a grudge.
- Nice: you don't defect first.
- Retaliatory: a defection gets a response.
- Forgiving: the punishment lasts only as long as the defection.
"Tit for Tat never wins a single match: it wins the tournament."
Why it works
Tit for Tat never scores more points than its opponent in a match: at best it ties. It wins the tournament because it generates a lot of cooperation around it. Its weak spot is mistakes: if a cooperation is misread as a defection, two reciprocal players can get trapped in a chain of retaliation. That is why more forgiving versions, such as Tit for Two Tats, do better when there is noise.
Unconditional cooperator
You keep cooperating even when the opponent defects. You trust and give second chances, even if it costs you points.
See the detail
What this profile describes
You cooperate almost always, even after a defection. It is the most generous strategy, and in a world of cooperators it would do as well as any other.
"Trust without a response doesn't teach the other player to cooperate: it teaches them that defecting is free."
The problem
In Axelrod's tournaments, strategies that didn't respond to defection were exploited by the more aggressive ones. Axelrod's finding was that lasting cooperation needs reciprocity: being nice isn't enough; you also have to be provocable.
Grudger
You cooperate until you're betrayed. From then on, even if the opponent goes back to cooperating, you find it very hard to trust again.
See the detail
What this profile describes
You resemble the Grim Trigger strategy: you cooperate as long as the other player cooperates, but a single defection breaks your trust for a long time.
"Never forgiving protects you from a second betrayal. It also rules out a second chance."
Why it pays less than it seems
A grudge deters, but it punishes too much: a single defection, even a mistake, turns the game into a string of mutual defections in which both players lose points. Axelrod pointed to forgiveness as one of the traits of the best-performing strategies: punishing and then cooperating again pays more than punishing forever.
Opportunist
You cooperate some of the time, but defect by surprise when you see a chance to gain an edge, even if the opponent has given you no reason.
See the detail
What this profile describes
You defected at least once against an opponent who had just cooperated, without punishing a defection of theirs. It's the pattern of strategies such as Joss or Tester: probe the other player and take advantage if they don't respond.
"A surprise defection pays off once. After that, it changes how the other player plays you."
Why it pays less than it seems
In Axelrod's first tournament, Joss, which was Tit for Tat with occasional defections, finished below pure Tit for Tat. Against reciprocal opponents, each surprise defection sets off retaliation that costs more than what was gained. It works better against naive cooperators, who are a minority in a real tournament.
Hawk
You defect almost every time. You protect yourself from being exploited, but you miss out on the points that only mutual cooperation brings.
See the detail
What this profile describes
You play like Always Defect: the strategy game theory identifies as dominant in a single round, because defecting leaves you better off whatever the other player does.
"No one can exploit you if you never cooperate. No one can help you either."
Why it fails in the long run
In a repeated game, two hawks get 1 point per round, while two cooperators get 3. In Axelrod's tournaments, aggressive strategies won many of their matches but scored few points overall: in the end, the winner was whoever managed to cooperate with the most opponents, not whoever beat the most.
Indicative result. It describes tendencies based on what you answered today, not a fixed trait or a closed label. It is not a diagnosis.
How this test was built
This game describes your cooperation strategy in a repeated prisoner's dilemma: how you start, how you respond to a defection and whether you forgive or hold a grudge.
This result is not a diagnosis or a validated personality test, and it does not say whether you are a good or bad person.
It is a game played for points, with no real consequences, and people cooperate differently when real money or real relationships are at stake.
See the full methodology
What it measures
This game describes your cooperation strategy in a repeated prisoner's dilemma: how you start, how you respond to a defection and whether you forgive or hold a grudge.
What model it uses
It is based on the iterated prisoner's dilemma and on the tournaments run by Robert Axelrod (1980). Axelrod summed up in four traits what the best-performing strategies had in common:
- Nice: never defects first.
- Retaliatory: responds to a defection.
- Forgiving: goes back to cooperating when the other player does.
- Clear: its way of playing is easy for the opponent to understand.
Where that model comes from
The dilemma was formalized by Merrill Flood and Melvin Dresher at the RAND Corporation in 1950, and Albert Tucker gave it the story of the two prisoners. Axelrod ran two tournaments: the first with strategies submitted by specialists from several disciplines, and the second open to anyone; Tit for Tat, submitted by Anatol Rapoport, won both. He told the story in The Evolution of Cooperation (1984).
How the result is calculated
Axelrod's payoffs are used: 3 points each if you both cooperate, 1 if you both defect, and 5 against 0 if one defects and the other cooperates. The opponent is chosen at random from four strategies that defect at some point on their own initiative (Joss, Tester, Always Defect and Random), so that the game always measures your reaction to a defection. Your profile comes from these rules, in this order:
- Hawk: you cooperate less than 35% of the time and, in addition, you start by defecting or defect two or more times unprovoked.
- Opportunist: you defect at least once unprovoked, that is, when the opponent had just cooperated and you weren't punishing one of their defections (starting with a defection also counts).
- Unconditional cooperator: you respond with a defection to less than 40% of the opponent's defections.
- Grudger: you keep defecting at least 60% of the time after the opponent goes back to cooperating.
- Reciprocal: in all other cases.
For the tournament, your way of playing is approximated as a reactive strategy (your first move and how often you cooperated after a cooperation and after a defection) and is played for 200 rounds against nine classic strategies and against itself.
What your result means
The result describes how you played in this game, not a fixed personality trait. It shows which pattern you spontaneously follow when cooperating is risky.
What it does NOT mean
This result is not a diagnosis or a validated personality test, and it does not say whether you are a good or bad person. A game of a few rounds against one specific opponent is a small sample: another game may give a different profile.
Limitations
It is a game played for points, with no real consequences, and people cooperate differently when real money or real relationships are at stake. The opponent affects the result: against Always Defect, for example, there is hardly any room to show forgiveness. The simulated tournament depends on which strategies take part, as Axelrod himself warned: there is no best strategy in the abstract.
Academic references
- Axelrod, R. (1980). Effective choice in the Prisoner's Dilemma. Journal of Conflict Resolution, 24(1), 3-25.
- Axelrod, R. (1980). More effective choice in the Prisoner's Dilemma. Journal of Conflict Resolution, 24(3), 379-403.
- Axelrod, R., & Hamilton, W. D. (1981). The evolution of cooperation. Science, 211(4489), 1390-1396.
- Axelrod, R. (1984). The Evolution of Cooperation. Basic Books.
- Nowak, M., & Sigmund, K. (1993). A strategy of win-stay, lose-shift that outperforms tit-for-tat in the Prisoner's Dilemma game. Nature, 364(6432), 56-58.
Frequently asked questions
What is the prisoner's dilemma?
A game from game theory in which each player is better off defecting, but if both do, both end up worse off than if they had cooperated. It was formalized by Merrill Flood and Melvin Dresher in 1950, and Albert Tucker gave it the story of the two prisoners.
What is the best strategy in the prisoner's dilemma?
In a single round, defecting. In repeated games there is no best strategy in the abstract, but in Robert Axelrod's tournaments the winner was Tit for Tat: cooperate first, then repeat whatever the other player did.
Which strategies can my opponent have?
One of four, chosen at random: Joss (Tit for Tat with surprise defections), Tester (defects at the start to size you up), Always Defect and Random. Which one it was is revealed at the end.
Why don't I know how many rounds there are?
Because if the last round is known, logic pushes players to defect in it, and by the same reasoning in the one before. In his second tournament, Axelrod made the length of each game uncertain for precisely that reason.
How is the final tournament calculated?
Your way of playing is summed up by your first move and by how often you cooperated after a cooperation and after a defection. That strategy plays 200 rounds against nine classic strategies and against itself, and is ranked by average points per round.
See all 7 questions
What does it have to do with negotiation?
Almost any negotiation that repeats (with a supplier, a coworker or a business partner) has the structure of a prisoner's dilemma: taking advantage may pay today, but reputation and reciprocity carry weight in the next round.
Is my game saved?
No. Everything happens in your browser: your moves are not stored or sent anywhere.
