The prisoner’s dilemma left us somewhere bleak: rational players defect, and mutual cooperation — though better for everyone — falls apart. Yet look around. Firms hold prices instead of warring to the death. Nations sign and keep treaties. Strangers return wallets. Flatmates wash up. Cooperation is everywhere, even among people who are nobody’s saints. So either everyone’s irrational, or something about the real world is missing from the one-shot matrix. It’s the second one. The missing ingredient is the future — the fact that we usually meet the same people again. And once the game repeats, the entire logic flips.
One shot vs. the shadow of the future
The one-shot dilemma has a fatal feature: there is no tomorrow in which your betrayal can be punished. Defect, grab your reward, walk away — the victim never gets to retaliate. But almost no real relationship is one-shot. Your supplier, your rival down the street, your colleague, your country’s neighbour — you’ll deal with them again next week and next year. A repeated game (or iterated game) is simply the same game played over and over by the same players, who remember what happened before.
That memory is everything. Now a defection today can be answered by a defection tomorrow. The prospect that your present choice shapes the other player’s future choices is called the shadow of the future, and when that shadow is long enough, cooperation stops being naive and becomes the smart, self-interested move. You cooperate not because you’re kind but because betrayal isn’t worth the retaliation it buys.
The key shift
In a one-shot dilemma you weigh: temptation to defect now vs. nothing. In a repeated game you weigh: temptation to defect now vs. all the future cooperation you’d destroy. When the future is heavy enough, the second term wins, and a rational player cooperates. Repetition doesn’t repeal the dilemma — it changes the payoffs by adding consequences that stretch into the future.
Axelrod’s tournament
In the late 1970s the political scientist Robert Axelrod ran a now-legendary experiment to find out which strategy actually does best in a repeated prisoner’s dilemma. He invited game theorists, economists, and computer scientists to submit strategies — full programs that would play the dilemma hundreds of times against each other, accumulating points. Some were fiendishly complex, probing and exploiting their opponents. Then he ran them in a round-robin tournament and added up the scores.
The winner stunned everyone. It was the simplest program submitted — just four lines of code — called tit-for-tat, sent in by the psychologist Anatol Rapoport. Its entire strategy: cooperate on the first move, then on every move after, copy whatever your opponent did last time. Be nice to start, then mirror them. That’s it. And it beat every clever, exploitative, scheming program in the room. Axelrod ran the tournament again, telling everyone tit-for-tat had won and inviting them to beat it. They couldn’t — tit-for-tat won the rematch too.
Play it yourself
Below is the repeated dilemma as a live machine. Pick a strategy for each side and watch the match unfold round by round, with running scores. (Payoffs follow the standard convention: mutual cooperation pays 3 each, mutual defection 1 each, and a lone defector grabs 5 while their victim gets 0.) Try the experiments in the caption.
Repeated game
The iterated prisoner's dilemma
Pick a strategy for each side, choose how many rounds they play, then step or play the match. Watch who pulls ahead — and notice that the defector’s edge fades the longer the game runs.
Player A
Player B
Notice what the machine shows. Always Defect beats Tit-for-Tat in their head-to-head by a hair — it steals on the first round and is never a sucker after. But that’s the wrong scoreboard. In a world of many strategies (Axelrod’s tournament), Always Defect poisons every relationship and gets stuck in mutual defection with everyone, while Tit-for-Tat builds long cooperative streaks with all the other nice strategies and only loses a single round to the nasty ones. Summed across the whole population, the nice retaliator wins the war while losing the odd battle.
Two Tit-for-Tat players meet for 12 rounds. What happens, and why?
Why tit-for-tat works: four traits
Axelrod analysed why the simple winner beat the schemers, and distilled four properties that made it strong. They’re a practical recipe for thriving in any repeated game.
- Nice — it never defects first. It’s never the one to start trouble, so it builds cooperation wherever cooperation is on offer.
- Retaliatory — it punishes a defection immediately, the very next round. This makes it impossible to exploit: cheat it once and you pay at once, so cheating doesn’t pay.
- Forgiving — it retaliates once and then, the moment you cooperate again, it goes straight back to cooperating. It holds no grudge, so a single mistake doesn’t spiral into endless mutual revenge.
- Clear — it’s utterly simple and predictable, so opponents quickly learn that cooperation is rewarded and defection is punished, and adjust toward cooperating.
The life lesson hiding in four lines of code
Axelrod’s traits are a startlingly good guide to dealing with people you’ll see again: start by trusting, hit back promptly when wronged, forgive the instant they straighten up, and be predictable about all three. Pure niceness gets exploited; pure nastiness poisons every well; the winning blend is a nice retaliator who forgives. Clarity matters too — a punishment nobody understands teaches nobody.
The catch: a known final round unravels
Repetition rescues cooperation, but there’s a famous crack. What if the game repeats a known, finite number of times — say exactly 100 rounds, and both players know it?
Reason about the last round, round 100. There’s no future after it, so it’s effectively a one-shot dilemma — and in a one-shot dilemma, defection is dominant. So both rational players defect on round 100. But now round 99: since you both know you’ll each defect on 100 no matter what, round 99 has no future worth protecting either — so you defect on 99. And by the same logic, 98, 97… all the way back to round
- The cooperation unravels from the end. This chain is called backward induction, and it predicts defection from the very first move of a known-length game.
The unraveling argument is logically airtight and yet real people (and Axelrod’s winning strategies) cooperate for ages. Why? A few reasons that matter in practice:
- Unknown end. If you don’t know which round is the last — most real relationships have a fuzzy, open-ended horizon — the backward-induction chain has no fixed end to start from, and cooperation is stable. The shadow of the future stays long.
- Indefinite repetition. If there’s even a constant probability the game continues next round, the future never fully vanishes, and cooperation can hold.
- Bounded rationality & reputation. Real players aren’t perfect backward-induction machines, value their reputation, and gain from being known as cooperators. A little of this is enough to sustain cooperation far longer than the cold logic predicts.
The takeaway isn’t that backward induction is wrong — it’s that most valuable real relationships are open-ended, which is precisely what keeps cooperation rational. Want someone to cooperate with you? Make the relationship feel ongoing, not final.
Sort each Tit-for-Tat behaviour under the trait it demonstrates.
Place each item in the right group.
- Punishes a betrayal immediately, not three rounds later
- Never throws the first punch
- Holds no lasting grudge after a single defection
- Opens the match by cooperating
- Defects the round after the opponent defects
- Returns to cooperating as soon as the opponent does
Match each concept to its meaning.
Pick a term, then click its definition.
Summarise the escape.
Pick the right option for each blank, then check.
In a prisoner's dilemma, cooperation can become rational because defecting today destroys valuable future cooperation — the . The simple strategy that exploits this, by cooperating first and then copying the opponent, is called .
Recap
Big picture
How repetition rescues cooperation
- Repeated Games
- The fix
- A future lets defection be punished → cooperation becomes rational (shadow of the future)
- Tit-for-tat
- Nice, retaliatory, forgiving, clear — won Axelrod’s tournaments with 4 lines
- The caveat
- A *known* last round unravels cooperation by backward induction
- Why we cooperate anyway
- Open-ended / indefinite horizons keep the future’s shadow long
- The fix
When to use it
Whenever you’re stuck in a dilemma you can’t escape by force of will, ask whether you can make the game repeat — turn a one-off transaction into an ongoing relationship, lengthen the horizon, make defections visible and punishable, and build a clear reputation as a nice-but-retaliatory player. To get cooperation from others, do the same to their incentives: signal that you’ll reciprocate, that you remember, and that the relationship has a long future. Then play tit-for-tat: open with trust, answer betrayal promptly, and forgive the moment they come back. Next, the final lesson widens the lens beyond the dilemma — to games of trust, nerve, and the costly mistake of seeing every contest as a fight to the death.