The Prisoner's Dilemma: What It Is, How It Began and Why Cooperating Is Still Rational

The prisoner’s dilemma does not prove that people are selfish. It proves something more uncomfortable: that two reasonable people, each doing what suits them, can end up together in an outcome neither of them wanted. And that the way out is almost never outsmarting the other person. It is changing the game.

It is probably the most famous problem in game theory. It is taught in economics, political science, evolutionary biology and negotiation courses.

But it is almost always told halfway: the story of the two prisoners, a table of prison sentences and a quick conclusion about human nature.

Here is the full version. What it is exactly, where it was born (in a nuclear strategy lab, not a police station), what people actually do when they play it, which strategies work when the game repeats and what all of this says about everyday life, from a shared apartment to a salary negotiation.

What the prisoner’s dilemma is

The story of the two prisoners

The police arrest two suspects in a crime and lock them up in separate cells, with no way to communicate. They have enough evidence to convict both on a lesser charge, but not on the main one.

Each suspect gets the same offer, separately:

• If you confess and your partner stays silent, you walk free and he gets the maximum sentence.
• If you both confess, you each get a substantial sentence.
• If you both stay silent, we can only convict you on the lesser charge: a short sentence.

Each prisoner knows the other has been offered exactly the same deal. And neither knows what the other will do.

The table that sums it up

With concrete prison terms (fewer years is better), the situation looks like this:

	Your partner stays silent	Your partner confesses
You stay silent	1 year each	You 10 years, he walks free
You confess	You walk free, he gets 10 years	5 years each

In the language of game theory, staying silent is “cooperating” (with your partner, not with the police) and confessing is “betraying” or “defecting.”

The reasoning that leads into the trap

Put yourself in one of the cells and reason coldly:

• If my partner stays silent, I am better off confessing: I walk free instead of serving a year.
• If my partner confesses, I am also better off confessing: five years instead of ten.

Whatever the other does, confessing is always better for me. Game theorists call this a dominant strategy.

The problem is that the other prisoner reasons exactly the same way. Both confess. Both serve five years.

And there is the paradox: if both had stayed silent, they would have served only one year each. Two individually logical decisions produce a collectively absurd result.

Two individually logical decisions produce a collectively absurd result.

What makes a game a prisoner’s dilemma

The prisoner story is just a costume. What defines the dilemma is the order of the outcomes, which the literature sums up with four letters:

• T (temptation): what you get if you betray while the other cooperates.
• R (reward): what you both get if you both cooperate.
• P (punishment): what you both get if you both betray.
• S (sucker): what you are left with if you cooperate and the other betrays you.

It is a prisoner’s dilemma when T > R > P > S. And when the game is repeated, one more condition applies: cooperating must always pay more than taking turns betraying (2R > T + S).

This matters because it means the dilemma shows up in any situation with that structure, whether or not there are jails, police or crimes. Two companies deciding whether to cut prices. Two countries deciding whether to arm. Two roommates deciding whether to clean the kitchen.

The equilibrium nobody wants

Mutual betrayal is what the mathematician John Nash called an equilibrium: a point at which no player gains anything by changing strategy alone. Nash formulated it in 1950, and that work eventually earned him the Nobel Prize in Economics in 1994.

What is disturbing about the prisoner’s dilemma is that this equilibrium is worse for both than another available option (mutual cooperation). Economists say it is not “Pareto efficient”: there is an outcome in which both would be better off and nobody worse off.

In other words, the dilemma does not pit the rational individual against the irrational one. It pits individual rationality against collective rationality. And in this game, they do not line up.

How it began: a January afternoon in 1950 at RAND

A lab for thinking about nuclear war

The dilemma was not born in a philosophy book but at the RAND Corporation, the strategic analysis center that the US Air Force set up after World War II in Santa Monica, California. It became an independent nonprofit in 1948.

It employed mathematicians, economists and physicists with an unusual mission: applying game theory to military strategy, in the middle of the nuclear race with the Soviet Union.

Game theory had just been born as a discipline with Theory of Games and Economic Behavior (1944), by John von Neumann and Oskar Morgenstern. Its great result, the minimax theorem, offered a rational solution for zero-sum games: those in which whatever one player wins, the other loses.

But real life is rarely zero-sum. That is where two RAND researchers came in.

The Flood and Dresher experiment

In January 1950, Merrill Flood and Melvin Dresher designed a simple game with the structure we know today and asked two colleagues to play it a hundred times in a row: the economist Armen Alchian, of the University of California, Los Angeles, and the mathematician John D. Williams, of RAND itself.

According to the theory, both should have betrayed in every round. They did not. They cooperated mutually in 60 of the 100 games and landed on mutual betrayal in only 14.

Flood and Dresher also collected the notes the players wrote during the game. They reveal something no table captures: one player was trying to “teach” the other to cooperate, the other grew impatient when he did not follow along, and both punished their partner’s betrayals in the next round.

Nash’s objection

John Nash, who had already formulated his equilibrium concept, raised an objection to the experiment: repeated a hundred times, it was not the same game as a single round. In a repeated game, the rational move can be different.

He was right, and that objection anticipated one of the most fertile ideas of the following decades: that cooperation can emerge from repetition, even among purely self-interested players. We will come back to it with Robert Axelrod.

Why people “don’t play right”: Flood’s other experiments

Flood had long been interested in how real people make decisions, not idealized players. William Poundstone, in his history of the dilemma, recounts two of his anecdotes that help make sense of the rest.

The first was the sale of a used Buick between Flood and an acquaintance. Both knew what the dealer would pay for it and what the dealer would sell it for. Cutting out the middleman left a gain of about 300 dollars to split, but nothing in formal logic said how to split it: either of them could veto the deal by being stubborn.

The second was a small experiment with two secretaries, who were offered a cash prize they would only collect if they agreed on how to divide it. The formal analysis allowed unequal splits. They divided it equally.

Flood’s conclusion was that the relationship between people changed the outcome far more than theory predicted. It is an observation that seventy years of research have only confirmed.

Tucker gives it a name (and a jail)

Flood and Dresher’s game was a table of numbers. It lacked a story.

Albert W. Tucker supplied one. He was a Princeton mathematician and Nash’s thesis adviser. In May 1950, during a stay at Stanford, he was asked to give a talk on game theory to the Psychology department. To make it accessible, he invented the story of the two prisoners in separate cells and titled it “A Two-Person Dilemma.”

The story worked so well that the game has been called the “prisoner’s dilemma” ever since. It is a good reminder that in science, how you tell things matters too.

The dilemma in the thick of the Cold War

The arms race as a prisoner’s dilemma

In August 1949 the Soviet Union tested its first atomic bomb. The United States was no longer the only nuclear power. For many analysts at the time, the situation had exactly the shape of the dilemma:

• If both countries disarm, both save money and live more safely (mutual cooperation).
• If one disarms and the other does not, the one that disarms is at the other’s mercy (the “sucker”).
• If both arm, both spend fortunes and live in fear (mutual betrayal).

Each, on its own, had reasons to arm. Both armed.

Von Neumann and preventive war

The logic of the dilemma led some of the century’s most brilliant minds to chilling conclusions. John von Neumann, the very man who founded game theory, argued for a time in favor of a preemptive strike on the Soviet Union before it could expand its arsenal.

Life magazine, in its 1957 obituary, reported a line attributed to him: if asked “why not bomb them tomorrow?”, he would reply “why not today?” It is the dilemma’s reasoning taken to the extreme: if mutual betrayal is inevitable, betraying first looks like the only sensible move.

The case of Bertrand Russell

More surprising is the case of the philosopher Bertrand Russell. Shortly after Hiroshima, Russell argued for pressuring the Soviet Union, even with the threat of war, to accept international control of atomic weapons while the United States still held the monopoly.

Years later, having become one of the great voices of nuclear disarmament, he denied ever having advocated preventive war. Poundstone documents that his own writings from the late 1940s said otherwise.

The story is uncomfortable, but useful: it shows that the dilemma’s reasoning is not the preserve of cynics. Thinkers with demanding ethics adopted it too, when fear of the other side’s betrayal dominated the analysis.

What the dilemma didn’t see

In hindsight, the Cold War did not end in the total betrayal the dilemma’s logic seemed to foretell. There were arms limitation treaties, mutual inspections, hotlines.

In other words, the two players found ways to change the structure of the game (communicating, verifying, repeating the interaction) instead of resigning themselves to it. That is, at bottom, the dilemma’s most important lesson, and the one that connects with everything that follows.

What people actually do when they play

The theory says that in a single round everyone should betray. The experiments, since 1950, say something very different.

Far more cooperation than predicted

In the 1950s, a group of researchers at Ohio State University (Scodel, Minas, Ratoosh and Lipetz, 1959) had students play many rounds of the dilemma. They found two things.

First, many players betrayed more than was good for them, because they seemed more concerned with not falling behind the other than with winning as much as possible. They preferred a relative gain over a larger absolute gain.

Second, there were stable differences between people. Poundstone reports a finding from the same line of studies: participants with more internationalist attitudes cooperated more than the more nationalist ones.

Even so, the body of later research shows a clear pattern: in anonymous one-shot games, a sizable share of people cooperate even though the selfish model predicts the opposite.

The name of the game changes almost everything

One of the most revealing experiments is by Varda Liberman, Steven Samuels and Lee Ross (2004). They had students play the same dilemma, with the same rules and the same payoffs. The only thing that changed was the name.

• When it was called “the Wall Street Game,” about two thirds of players competed.
• When it was called “the Community Game,” only a third did.

Before the experiment, the researchers had asked dorm advisers to predict who would cooperate and who would compete. That reputation predicted behavior far worse than the name of the game did.

The takeaway is blunt: what we think of as “character” depends, to a large extent, on how we interpret the situation we are in.

When there is a lot of money at stake

And if the payoffs are serious? Martijn van den Assem, Dennie van Dolder and Richard Thaler (2012) analyzed 287 episodes of the British game show Golden Balls, whose final round was a prisoner’s dilemma with prizes averaging around 20,000 dollars: each contestant secretly chose to “split” or “steal.”

52.8% of contestants chose to split. In 31% of episodes both split, in 44% one split and the other stole, and in 25% both stole.

The study found interesting nuances: men cooperated less than women, but the gap reversed with age, because men became more cooperative as they got older. And cooperation was surprisingly high for amounts that would be very significant outside the show but seemed small there next to the total jackpot.

Real prisoners playing the prisoner’s dilemma

The most ironic test was run by Menusch Khadjavi and Andreas Lange (2013): they took the dilemma into a women’s prison in Germany and compared the inmates with female university students.

The result contradicts the stereotype: in the simultaneous version of the game, the inmates cooperated more than the students. And a similar share of both groups showed social preferences, meaning they also cared about what happened to the other person.

Is cooperating the intuitive choice?

In 2012, David Rand, Joshua Greene and Martin Nowak published a widely cited paper in Nature. They used a public goods game, a relative of the dilemma with four players who contribute (or not) to a common pool, and found that those who decided quickly, under time pressure, cooperated more than those who had to think it over. They read this as evidence that cooperation is, for many people, the intuitive response, learned in an everyday life where cooperating usually pays off.

The idea is appealing, but it should be held with caution. A registered replication across 21 labs (Bouwmeester and colleagues, 2017) reproduced the pattern only among those who complied with the time instruction; when all participants were included, the effect disappeared. The question remains open.

Why talking helps (even if it doesn’t solve it)

What the meta-analyses say

The classic story isolates the prisoners in separate cells. What happens if they are allowed to talk?

Two meta-analyses answer fairly clearly. David Sally (1995) reviewed experiments published between 1958 and 1992 and concluded that communication raised cooperation very substantially. Daniel Balliet (2010), analyzing 45 effects, found a large effect (d = 1.01), bigger for face-to-face conversation (d = 1.21) than for written messages (d = 0.46).

Why, in theory, it shouldn’t work

From pure logic, talking changes nothing. If the other promises to cooperate, betraying is still more profitable. And if there is no way to make them keep their word, the promise is what economists call “cheap talk.”

That is why theorists insist on a caveat: if communication allows the players to sign a binding agreement that can be enforced, the game stops being a prisoner’s dilemma. It becomes a different game.

Why, in practice, it does work

That it works so well says a lot about us. Talking creates commitment, reduces uncertainty about the other person, activates norms of reciprocity and makes it psychologically more costly to betray someone you have looked in the eye.

Poundstone sums it up with a simple idea: a world of perfect communication and perfect honesty would be a world without prisoner’s dilemmas. It is not our world, but every conversation brings us a little closer.

When the game repeats: Axelrod and the simplest strategy

The 1980 tournament

Nash’s objection (the repeated game is a different game) hung in the air for thirty years. It was taken up by the political scientist Robert Axelrod, of the University of Michigan, with a brilliant idea: organizing a computer tournament.

He invited game theory experts to submit strategies for playing the repeated dilemma. He received 14 programs, which played every other program in games of 200 rounds, a length set in the tournament rules. Axelrod also added a program that played at random.

The winner was the shortest strategy of all: TIT FOR TAT. It was submitted by Anatol Rapoport, a mathematician and psychologist.

Its rule fits in one line: cooperate on the first move and, from then on, repeat whatever the other player did on the previous move.

The second tournament

Axelrod published the results and called a second tournament. This time everyone knew tit for tat had won and could design strategies to beat it. On top of that, the length of the games was no longer fixed or known, so that nobody could exploit the last round.

62 strategies from six countries arrived, from professors of computer science, physics, economics, psychology and mathematics, and even from a ten-year-old programming enthusiast. Rapoport submitted exactly the same strategy again. It won again.

The four properties of the winning strategies

Analyzing the results, Axelrod found that the best-performing strategies shared four traits:

1. Nice: they were never the first to betray.
2. Retaliatory: they responded to a betrayal right away, so as not to be exploited.
3. Forgiving: after punishing, they went back to cooperating if the other did, without holding a grudge.
4. Clear: they were so simple that the opponent quickly understood how they worked and found it worthwhile to cooperate.

There is a detail that often goes unnoticed: tit for tat never wins an individual game. At best it ties its opponent. It wins the tournament because it racks up many games of mutual cooperation, while the aggressive strategies tear each other apart.

Axelrod turned this into recommendations he published in The Evolution of Cooperation (1984): don’t be envious, don’t be the first to defect, reciprocate both cooperation and defection, and don’t be too clever.

The shadow of the future

Why does reciprocity work? Because in a repeated game the future weighs on the present. Axelrod called it “the shadow of the future”: betraying today gives you a quick gain, but it costs you tomorrow’s cooperation.

The more likely you are to meet again, and the more those future encounters matter, the more sense cooperating makes. That is why people behave differently with a lifelong neighbor than with a stranger in an airport.

The paradox of the known ending

There is a logical trap. If both players know the game has exactly a hundred rounds, in the last one there is no future left to protect: the rational move is to betray. But if both will betray in round one hundred, there is no reason to cooperate in round ninety-nine. And so on backward, all the way to the first.

This is the so-called backward induction paradox. Its formal conclusion is that two perfectly rational players would betray from the start. Observed behavior, once again, is different: most people cooperate for much of the game and only betray near the end.

Cooperation in the trenches

Axelrod illustrated his theory with an extraordinary historical case, documented by the historian Tony Ashworth: the “live and let live” system in the trenches of World War I.

On some stretches of the Western Front, enemy units that faced each other for months developed tacit agreements: firing at predictable times, deliberately aiming high, not attacking during meals. And they responded harshly when the other side broke the truce.

It was, according to Axelrod, a repeated prisoner’s dilemma solved through reciprocity. This deserves a caveat: political scientists such as Joanne Gowa and Andrew Gelman have questioned whether the case truly fits that model, and Ashworth himself gives a broader explanation, with factors that go beyond the game. But as an illustration of how cooperation arises among people forced to keep meeting, it is still hard to beat.

The limits of tit for tat

The echo problem

Tit for tat has a serious weakness: it does not forgive misunderstandings. If one player betrays by mistake, or if the other misreads a move as a betrayal, a chain of retaliation begins that may never end.

Poundstone calls it the echo effect, and applies it to international relations: two countries that do not share the same account of who started it can punish each other indefinitely, each convinced it is only responding.

More generous strategies

The theoretical biologists Martin Nowak and Karl Sigmund studied what happens when there are errors and strategies evolve. In 1992 they showed that tit for tat works as a pivot that opens the way to cooperation, but is eventually displaced by a more generous version that randomly forgives some betrayals.

In 1993 they described an even more robust strategy, known as “Pavlov” or “win-stay, lose-shift”: if the previous move went well, you repeat it; if it went badly, you switch. It corrects errors easily and, on top of that, takes advantage of players who always cooperate unconditionally. According to Nowak and Sigmund, that toughness toward unconditional cooperators is part of its strength: it keeps them from spreading and opening the door to players who always betray.

The underlying lesson is valuable: in a noisy world, strict reciprocity is not enough. Some forgiveness is needed for cooperation to survive mistakes.

Cooperation in nature

Selfish players who cooperate

For a long time, cooperation between unrelated animals was a puzzle for evolutionary biology. If natural selection rewards whoever reproduces most, why help another at your own expense?

In 1981, Axelrod and the biologist William D. Hamilton published a paper in Science that connected both worlds: cooperation based on reciprocity can emerge and persist among selfish players if encounters repeat often enough.

The phrase “survival of the fittest” is often misunderstood. Evolution does not select the strongest in the abstract, but the strategies that remain stable in a population. And sometimes the stable strategy is conditional cooperation.

The fish in front of the mirror

One of the most elegant demonstrations came from the biologist Manfred Milinski (1987), with sticklebacks, small fish that approach a predator in pairs to inspect it. Approaching is risky; letting your partner take the risk alone is the “betrayal.”

Since a real pair could not be controlled, Milinski used mirrors: one parallel to the tank simulated a partner that kept pace with the fish on every advance; another, angled, simulated a partner that hung back.

With the “cooperative” partner, the fish dared to get much closer to the predator. With the “defector,” they held back. Their behavior resembled tit for tat.

The dilemma’s relatives

Not every conflict situation is a prisoner’s dilemma. Changing the order of outcomes produces other games with very different logic, and they are worth telling apart.

Game	Worst outcome for you	What makes it hard
Prisoner’s dilemma	Cooperating while the other betrays	Betraying always pays, no matter what
Chicken	Both of you standing firm	Nobody wants to give way, but neither giving way is the disaster
Stag hunt	Cooperating while the other goes it alone	Cooperating is best for both, if you trust the other
Deadlock	Cooperating while the other betrays	Both prefer to betray: there is no real dilemma

Chicken and the missile crisis

In the game of chicken, two drivers speed toward each other down a road. The one who swerves first is the “chicken.” If neither swerves, they crash.

Bertrand Russell used this image in 1959 to describe nuclear policy, and the Cuban Missile Crisis (October 1962) is often analyzed in these terms: two powers that each need the other to back down without looking weak themselves. The way out depended largely on finding a way to back down without losing face.

The political version of the game is the so-called “madman theory,” associated with Richard Nixon: seeming unpredictable enough that the other side prefers to swerve.

The stag hunt: a problem of trust

The stag hunt comes from an image by Jean-Jacques Rousseau. Two hunters can hunt a stag together (a big reward, but it requires both to work together) or each go after a hare (a small reward, but a sure one).

Unlike the prisoner’s dilemma, here cooperating is best for both if the other also cooperates. The obstacle is not temptation but fear: if I doubt the other will stay, I go after my hare.

Many real-life problems described as prisoner’s dilemmas are actually stag hunts: they fail not out of greed but out of lack of trust. And that changes the solution, because trust can be built.

The dollar auction: the escalation trap

The economist Martin Shubik proposed a fiendishly simple game in 1971: a one-dollar bill is auctioned off. The highest bidder wins, but the second-highest bidder also pays what they bid, and gets nothing.

Bidding starts low. But when someone offers 99 cents and the other offers 1 dollar, the first bidder is better off bidding $1.01 rather than losing 99 cents. And the escalation continues, often far above the value of the bill.

It is the logic of price wars, of certain labor disputes, of sitting through a bad movie because you already paid for the ticket, or of staying on hold because you have already waited twenty minutes. The link to the dilemma is clear: each bidder mirrors the other’s defection, and together they sink.

The tragedy of the commons

When the dilemma involves many people, it becomes what the ecologist Garrett Hardin called, in 1968, the tragedy of the commons: a shared pasture where each herder benefits from adding one more sheep, until the pasture is exhausted for everyone.

For years it was thought there were only two ways out: privatizing the resource or having the state manage it. Elinor Ostrom showed there was a third. In Governing the Commons (1990) she documented communities that had managed pastures, forests and irrigation systems for centuries without depleting them, using their own rules, mutual monitoring and graduated sanctions. She received the Nobel Prize in Economics in 2009, the first woman to do so.

The dilemma in everyday life

Where it shows up without you noticing

Once you learn to recognize the structure, the dilemma turns up everywhere:

• At home: each roommate is better off if the other one cleans. If nobody cleans, everyone lives worse.
• At work: sharing information with a colleague benefits the team, but keeping it to yourself can give you an individual edge in the next review.
• In sports: if nobody dopes, they compete on equal terms; if one dopes, that athlete gains an edge; if everyone dopes, they are back to equal terms, but with risks to their health.
• In markets: two companies would earn more by holding prices, but each is tempted to cut them to steal the other’s customers.
• In couples: giving in during an argument can feel like losing, and if neither gives in, the conflict festers.

On this last front, John Gottman’s research on bids for connection in couples describes something very much like a repeated game: each small response to the other person feeds, or erodes, the future willingness to cooperate.

Not everything is a prisoner’s dilemma

It is also wise not to see dilemmas everywhere. Many situations look like conflicts but do not have that structure:

• If you genuinely care what happens to the other person, their losses count as yours and the order of outcomes changes. It is no longer a prisoner’s dilemma.
• If there is an enforceable agreement (a contract, a rule with consequences), it is not one either.
• If what is missing is not will but trust, it is probably a stag hunt.

Poundstone stresses this in his book: the game’s payoffs are utilities, not years in prison or dollars. If what you value includes the other person’s well-being, the game you are playing is a different one.

What changes the outcome, according to the evidence

Research does not offer a recipe, but it does identify, fairly consistently, the factors that make cooperation more likely. They all have something in common: they do not ask people to be better people. They change the situation.

1. Repetition and a long horizon. The more likely you are to meet again, the more cooperating pays (Axelrod). That is why lasting relationships withstand conflict better than one-off encounters.
2. Communication, especially face to face. It is one of the most robust effects in the entire literature (Sally, 1995; Balliet, 2010).
3. Reputation and visibility. When actions are seen and remembered, betrayal carries a cost that goes beyond the game.
4. Changing the payoffs. Contracts, guarantees, penalties or shared incentives turn betrayal into a bad option. That is what institutions do.
5. The framing of the situation. The same game produces twice as much cooperation when it is understood as a community matter rather than a competition (Liberman, Samuels and Ross, 2004).
6. Their own rules and graduated sanctions. Communities that successfully manage shared resources set their own norms, monitor each other and punish proportionately (Ostrom, 1990).
7. Reciprocity with room for error. Responding to betrayal while leaving the door open to cooperate again protects better than strict revenge (Nowak and Sigmund, 1992).

In negotiation, several of these ideas overlap with what the Harvard method calls having a good BATNA, the best alternative if there is no agreement: knowing what happens if the other side betrays reduces the fear that pushes people to betray first.

Our reading: what the prisoner’s dilemma doesn’t say

After going through seventy years of theory and experiments, some conclusions strike us, as an editorial team, as solid and others deserve to be taken apart.

It is not proof of human selfishness

The dilemma has often been used as an argument that people are, deep down, calculating egoists. The data say the opposite: from Alchian and Williams in 1950 to the German inmates in 2013, people cooperate far more than the model predicts.

What the dilemma does show is that good will is not enough when the structure rewards betrayal and punishes whoever trusts. It is an argument about situations, not about people.

The trap lies in being only half “rational”

The cognitive scientist Douglas Hofstadter, in his 1983 Scientific American column, put forward a provocative idea: “superrationality.” If both players are equally rational and know it, they should reach the same conclusion. And between “we both cooperate” and “we both betray,” the rational choice is obvious.

Many game theorists reject it, and they have their arguments. But Hofstadter’s intuition captures something real: the reasoning that leads to mutual betrayal treats the other as an obstacle, not as someone thinking the same thing you are.

The dilemma is a tool, not a destiny

There is something dangerous about learning the prisoner’s dilemma: it can become a prophecy. Whoever sees dilemmas everywhere and always expects the other to betray tends to betray first, and so brings about exactly what they feared. The Ohio State studies were already picking up players more concerned with not falling behind than with winning.

The value of the model, as Poundstone points out, is that it lets you see clearly the principles at work in a situation. Once you see them, they can be changed.

The conclusion we keep

The prisoner’s dilemma has a way out, but it is not inside the cell. It is not solved by choosing better between cooperating and betraying, but by changing the conditions of the game: talking, repeating, making actions visible, creating fair rules, forgiving mistakes.

Poundstone closes his book with an idea that sums all this up well: the paradox of the prisoner’s dilemma will never be resolved, or it is already resolved, because it does not really exist. What exists is a real conflict between what suits each person and what suits everyone. And managing that conflict is probably one of civilization’s oldest tasks.

What kind of player are you?

The way someone faces a dilemma like this has a lot to do with their style in negotiating and handling conflict: some prioritize the outcome, some protect the relationship and some look for solutions that work for both sides.

If you are curious where you land, you can take the negotiation style quiz or the quiz on how you handle conflict. And if what intrigues you is your relationship with uncertainty (how hard it is for you to trust when you don’t know what the other will do), the quiz on how you approach risk rounds out the picture.

To go deeper into the profiles, the articles on the collaborative negotiator, the competitive negotiator and the collaborating conflict style describe how each one behaves when cooperation and competition pull against each other.

Disclaimer: This article is for informational purposes only. It summarizes research in game theory, experimental economics and social psychology, and does not constitute professional advice on financial, legal or negotiation decisions.

Sources

• Poundstone, W. (1992). Prisoner’s Dilemma: John von Neumann, Game Theory, and the Puzzle of the Bomb. Doubleday.
• Flood, M. M. (1952). Some Experimental Games. Research Memorandum RM-789-1. RAND Corporation.
• Tucker, A. W. (1950). A Two-Person Dilemma. Talk notes, Stanford University, May 1950.
• Von Neumann, J. and Morgenstern, O. (1944). Theory of Games and Economic Behavior. Princeton University Press.
• Nash, J. F. (1950). Equilibrium points in n-person games. Proceedings of the National Academy of Sciences, 36(1), 48-49.
• Scodel, A., Minas, J. S., Ratoosh, P. and Lipetz, M. (1959). Some descriptive aspects of two-person non-zero-sum games. Journal of Conflict Resolution, 3(2), 114-119.
• Axelrod, R. (1980). Effective choice in the prisoner’s dilemma. Journal of Conflict Resolution, 24(1), 3-25.
• Axelrod, R. and Hamilton, W. D. (1981). The evolution of cooperation. Science, 211(4489), 1390-1396.
• Axelrod, R. (1984). The Evolution of Cooperation. Basic Books.
• Ashworth, T. (1980). Trench Warfare 1914-1918: The Live and Let Live System. Macmillan.
• Milinski, M. (1987). TIT FOR TAT in sticklebacks and the evolution of cooperation. Nature, 325, 433-435.
• Nowak, M. A. and Sigmund, K. (1992). Tit for tat in heterogeneous populations. Nature, 355, 250-253.
• Nowak, M. A. and Sigmund, K. (1993). A strategy of win-stay, lose-shift that outperforms tit-for-tat in the Prisoner’s Dilemma game. Nature, 364, 56-58.
• Sally, D. (1995). Conversation and cooperation in social dilemmas: A meta-analysis of experiments from 1958 to 1992. Rationality and Society, 7(1), 58-92.
• Balliet, D. (2010). Communication and cooperation in social dilemmas: A meta-analytic review. Journal of Conflict Resolution, 54(1), 39-57.
• Liberman, V., Samuels, S. M. and Ross, L. (2004). The name of the game: Predictive power of reputations versus situational labels in determining prisoner’s dilemma game moves. Personality and Social Psychology Bulletin, 30(9), 1175-1185.
• Van den Assem, M. J., van Dolder, D. and Thaler, R. H. (2012). Split or steal? Cooperative behavior when the stakes are large. Management Science, 58(1), 2-20.
• Khadjavi, M. and Lange, A. (2013). Prisoners and their dilemma. Journal of Economic Behavior & Organization, 92, 163-175.
• Rand, D. G., Greene, J. D. and Nowak, M. A. (2012). Spontaneous giving and calculated greed. Nature, 489, 427-430.
• Bouwmeester, S. et al. (2017). Registered Replication Report: Rand, Greene, and Nowak (2012). Perspectives on Psychological Science, 12(3), 527-542.
• Shubik, M. (1971). The Dollar Auction game: A paradox in noncooperative behavior and escalation. Journal of Conflict Resolution, 15(1), 109-111.
• Hardin, G. (1968). The tragedy of the commons. Science, 162(3859), 1243-1248.
• Ostrom, E. (1990). Governing the Commons: The Evolution of Institutions for Collective Action. Cambridge University Press.
• Hofstadter, D. R. (1983). Metamagical Themas [column]. Scientific American, May and June 1983 issues.