The payoff that makes betrayal tempting

The prisoners’ dilemma is built around a mismatch between what is safest for one person and what is best for the pair. Imagine two suspects questioned separately, unable to coordinate. Each can stay silent or implicate the other. The rules are arranged so that, whatever the other suspect does, implicating them gives you the better personal result: if they stay silent, you may earn leniency while they take the blame; if they implicate you, you avoid being the only one who stayed silent.

That comparison is the key. A suspect does not need to be cruel or irrational to confess. They only need to reason from the options available to them, assuming the other person might betray them too.

Why the individually safer move harms both

Because the same incentive applies to both suspects, each has reason to implicate the other. If both follow that logic, both receive a worse outcome than they would have received had they both stayed silent. Silence is better for the pair, but it is risky for either person acting alone: a silent suspect can be left carrying the blame if the other confesses.

In game theory, a choice that remains best for a player no matter what the other player chooses is called a dominant strategy. In the classic dilemma, confessing is dominant for each suspect, even though the result of both confessing is collectively worse. The trap is not that either person misunderstands the arithmetic; it is that each person’s safest response helps produce the outcome neither would have chosen together.

When the dilemma changes

The classic setup assumes a one-time decision, no communication, and payoffs that make betrayal individually advantageous in either case. Change those conditions and cooperation may become more plausible. If the same people interact again, today’s choice can affect whether the other person trusts or helps you tomorrow. A reliable way to coordinate—or a consequence for breaking an agreement—can also alter the incentives.

That is why the dilemma is useful beyond a police interrogation. It shows how a bad shared result can emerge without either person wanting it: when each protects against being the one who cooperates alone, both can end up worse off.