Boy or Girl Probability Puzzle

This classic quant interview puzzle is about updating beliefs when you see a partial outcome in a simple, concrete setting. You start with uncertainty about an unknown characteristic, then you observe something correlated with it, and must revise the probability that the unknown has a given value. The setup is deliberately simple so the focus stays on conditional probability rather than algebraic complexity.

It trains your intuition for conditional probability, the law of total probability, and Bayesian-style updating. You learn to enumerate underlying scenarios, weigh them correctly, and reinterpret an observed event as information that reshapes prior beliefs. It also strengthens your ability to translate a verbal description of uncertainty into a clean probabilistic model.

For quant prep, this matters because top trading firms use such questions to test whether you can reason probabilistically under uncertainty, avoid intuitive traps, and think rigorously about inference in trading and risk situations.

What it tests

This problem class is governed by conditional probability and the law of total probability, particularly in settings where an observed outcome is used to update beliefs about an unobserved event. The core structure is that you have two (or more) hypotheses about an unknown, and you observe an event whose likelihood depends on which hypothesis is true. Bayes' theorem formalizes how to reverse the direction of conditioning: instead of asking 'what is the chance of seeing this evidence if the hypothesis is true?', you ask 'what is the chance the hypothesis is true given this evidence?' The pattern holds because probabilities must be consistent across all possible worlds: the chance of the evidence is the weighted sum of its chance under each hypothesis, and the posterior probability is the fraction of that evidence attributable to each hypothesis. This is the backbone of inference wherever you see 'given that we observed X, what is the chance of Y?'