The Prisoner's Dilemma
Three prisoners , , and have applied for parole. The board will release exactly two of the three, and each prisoner knows this, but not which two.
Prisoner asks a warder (who knows the decision) to name one prisoner other than himself who will be released. The warder says: " will be released."
Prisoner reasons: "Before asking, my chance of release was . Now I know it's either or , each equally likely — so my chance has dropped to . I should not have asked!"
Is 's reasoning correct? What is 's actual probability of release after the warder's answer?
(Assume: if both and are to be released, the warder picks uniformly at random between naming or naming .)
Answer: The Prisoner's Dilemma
Key Idea / Intuition
The mistake is subtle: forgets that the warder's choice of words is itself random data. When and are both to be released, the warder must say "" (he can't name ). But when and are both to be released, the warder only says "" with probability . So hearing "" is more likely under the scenario than under , which tilts the posterior back in 's favor. The answer is: 's probability of release is still — asking changes nothing!
Formal Proof / Solution
Step 1 — Prior sample space.
The three equally likely release pairs are:
Step 2 — Warder's behavior (the key modeling step).
The warder must name someone other than who will be released.
| Event | Probability of event | Prob. warder says "" given event | |---|---|---| | released | | (only is available) | | released | | (must say "") | | released | | (names or equally) |
Step 3 — Joint probabilities.
Step 4 — Bayes' theorem.
Conclusion. 's probability of release is still , exactly as before asking. The information " will be released" tells nothing new about his own fate — but it does update the relative probability between and in exactly the right way to cancel 's naive worry.
This is a cousin of the Monty Hall problem: the host/warder's protocol for choosing what to reveal is essential to the calculation.
Source: Fifty Challenging Problems in Probability with Solutions, Frederick Mosteller, Problem 13