The Prisoner’s Dilemma: Why Cheating on a Test (Sometimes) Doesn’t Pay Off

(Or: the math trap that explains why trust is worth so much)

Imagine you and your desk partner have a hard exam tomorrow. Neither of you studied enough. You could agree to help each other cheat during the exam. Sounds perfect… until you realize the teacher spreads out the desks and now each of you has to decide alone, without knowing what the other one will do. Welcome to the prisoner’s dilemma, one of the most famous concepts in game theory.

The original story (with criminals, not exams)

The name comes from a classic scenario: two suspects are arrested and separated. The police don’t have enough evidence to convict them of the serious crime, so they offer each one, separately, the same deal:

• If neither confesses, both get a short sentence for a minor offense.
• If one confesses and the other doesn’t, the one who confesses goes free and the other gets a long sentence.
• If both confess, both get a medium sentence.

Each one, thinking only about themselves, reasons: “if I confess, worst case I get the medium sentence, best case I go free. If I don’t confess, worst case is awful.” So both end up confessing, even though they both would have been better off if neither had. That’s the heart of the dilemma: the individually “rational” decision doesn’t always lead to the best collective outcome.

Back to the exam

“If I don’t cheat and my partner does, they get a better grade than me without deserving it. If I cheat and they don’t, I gain the advantage. If we both cheat, we both risk getting caught.”

Each one, without knowing what the other will do, has an incentive to “betray” the joint strategy (cheating together) even though they’d both be better off agreeing to it and sticking with it. This exact same pattern shows up in situations far more serious than an exam.

Where else this dilemma shows up (in the real world)

• Companies and prices: two companies would earn more if both kept prices high, but each one has an individual incentive to cut its price and steal the other’s customers — until both end up competing so hard they earn less than they could have.
• Countries and the environment: every country would be better off cutting emissions together, but each individual country gains a short-term economic edge by not doing it while the others do.
• Arms races: both countries would be better off not spending on weapons, but each fears being at a disadvantage if the other does anyway.

Why does this matter so much in economics?

Game theory studies exactly this type of situation: decisions where the outcome doesn’t depend only on what you do, but on what everyone else does at the same time. Understanding this helps you predict why “rational” people, companies, or countries so often end up in outcomes that are bad for everyone, simply because nobody can trust the other to cooperate.

Is there a way out?

Yes, and it’s called repeated cooperation. When the same situation repeats many times (it’s not a one-shot game), people start building trust, reputation, and punishment for whoever betrays the deal. That’s why companies in the same industry sometimes manage to hold implicit agreements, and why you and your desk partner, if you’re going to share a classroom all year, will probably coordinate better than two strangers who only meet once.

“The prisoner’s dilemma doesn’t say people are naturally selfish. It says that, without trust or communication, even well-intentioned people can end up making the worst possible decision for everyone.”

So next time you see two people (or two countries) unable to agree on something that would clearly benefit both, now you know what’s happening: a classic prisoner’s dilemma, playing out in real time.