Two hunters set out at dawn. A stag would feed both families for a week, but it takes two people to bring one down. A hare can be caught alone: enough for today, and no more.
They split up to stalk, out of sight of each other, and each has to decide alone.
A share of the stag is worth 4. A hare is worth 3, whatever the other does. Go for the stag alone and you come home with 0.
You are Hunter A
Watch one round, or choose now.
Hunter A
or
You
Hunter A
?
You
Hunter B
or
The other hunter
Hunter B
?
The other hunter
The other hunter picks at random here, like a coin toss.
The whole game on one grid
Every pair of choices and what each side gets: yours first, then theirs. Each number is a payoff.
Step 0 of 3
The cards are face down. Turn them, then step through which squares are stable.
What do you think they'll do?
Your best move depends on what you expect from the other hunter. Drag to set how likely you think it is they pick Hunt stag.
What it teaches
When cooperation is not dominated, the obstacle is not greed but doubt about the other.
The usual mistakeReading it as a prisoner's dilemma, concluding that cooperation is irrational and that enforcement is needed, when in fact cooperating is a best reply as soon as you expect the other to cooperate.
The checkAsk what you would do if you were certain the other hunts stag. If the answer is hunt stag, the remedy is reassurance and not enforcement.
What people actually do
The name of the game changes how it is played
The theory says
A game is defined by its players, actions and payoffs. Two presentations with identical payoff matrices are the same game, so labels, descriptions and irrelevant context should not change behaviour.
People actually
A one-shot prisoner's dilemma labelled the Community Game draws roughly twice the cooperation of the same game labelled the Wall Street Game, and labels predict individual behaviour better than observers' judgements of who is cooperative. In paired treatments where a payoff parameter changes without moving the equilibrium, behaviour tracks the parameter rather than the prediction, as with claims in the travellers' dilemma moving with the size of the penalty.
Why is still argued over.
In the real world
Real world
Two people on a shared deadline
Hunter A
you
Hunter B
your colleague
Hunt stag
commit to the ambitious plan
Hunt hare
do the safe minimum
Fiction: One Piece
The raid on Onigashima
The allies against Kaido spend years in hiding and agree to attack together at one place and time, because a group that moves alone is destroyed while a group that stays hidden survives in reduced circumstances. A traitor inside the alliance passes the plan on, the rendezvous is compromised, and the whole thing has to be rebuilt around a signal each group can see that the others have received. The raid only goes ahead once that common signal is back in place.
Rising together is the best outcome for everybody, rising alone is the worst, and hiding is a tolerable third, so nobody has a dominant strategy and the binding constraint is confidence that the others will turn up. Calling it a prisoner's dilemma would wrongly suggest the allies were tempted to sell each other out for gain.
Real worldContested
Verified limits on Iran's nuclear programme
The 2015 agreement capped Iranian enrichment and gave the IAEA extensive access in exchange for sanctions relief, and the inspection regime existed precisely because neither side would take the other's word. The United States left the agreement in 2018 and reimposed sanctions, Iran stepped past the limits from 2019, and after its nuclear sites were struck in June 2025 it suspended cooperation with the inspectors, so that by early 2026 the agency was reporting that it had no access to any of the declared enrichment facilities. In June 2026 the two sides signed a memorandum in which Iran reaffirmed that it would not acquire nuclear weapons, and opened a further round of talks.
Neither side has a dominant strategy, because restraint met with restraint is better for each of them than the mutual hard line, and what blocks it is doubt about the other rather than a standing temptation to cheat on an agreement both would otherwise keep, which is why the whole apparatus is verification rather than punishment. The inference, and it is inference rather than fact, is that Tehran has been optimising for preserving the option rather than exercising it, and Washington for verifiable limits rather than collapse; read instead as a prisoner's dilemma, cheating would be dominant and inspections would be beside the point.
Assurance gameCooperation is risk dominant as well as payoff dominant, so the belief needed to sustain it is below one half. Not on the wall yet.
Bank run gameThe safe action is strictly better than nothing rather than merely adequate, and the off-diagonal payoffs run at the other player's expense, so waiting takes a two-thirds belief in the other to justify. Not on the wall yet.