Boy or Girl Paradox Calculator

The Boy or Girl Paradox, formulated by Martin Gardner in 1959, involves a family with two children of unknown genders. When told that at least one child is a girl, most people intuitively guess there's a 1/2 chance both are girls — but the correct answer is 1/3. The paradox illustrates how the phrasing of a condition dramatically affects conditional probability.

With two children, there are four equally likely combinations: BB, BG, GB, and GG. Knowing "at least one is a girl" eliminates BB, leaving three possibilities: BG, GB, GG. Only one of those three has both girls, so the probability is 1/3, not 1/2. The intuitive 1/2 answer incorrectly treats the remaining child as an independent coin flip.

If you know specifically that the older child is a girl, that eliminates BB and BG, leaving only GB and GG. Now exactly one of two possibilities has both girls, so the probability becomes 1/2. This is why the wording matters enormously — 'the older child is a girl' is a much more specific condition than 'at least one is a girl.'

A famous extension asks: 'A family has two children; at least one is a girl born on a Tuesday. What is the probability both are girls?' Surprisingly, the answer is 13/27 — closer to 1/2 than 1/3. Adding the day-of-week detail changes the sample space, showing that even seemingly irrelevant information can shift probabilities.

It matters in the sense that we count BG (older boy, younger girl) and GB (older girl, younger boy) as distinct outcomes, giving four equally likely combinations. This is the standard mathematical treatment. If you ignore birth order and only count unordered pairs (BB, BG, GG), the probability changes — which is part of why the paradox generates so much debate.

The two-child problem is another name for the Boy or Girl Paradox. It refers to the broader class of conditional probability puzzles involving a family with two children where partial gender information is given. The key insight is that the answer depends critically on how the information was obtained — randomly versus by deliberate selection.

If the probability of a girl is p (not necessarily 0.5), the four combination probabilities become unequal: BB = (1−p)², BG = p(1−p), GB = (1−p)p, GG = p². The conditional probability of both girls given 'at least one is a girl' becomes p² / (1 − (1−p)²). You can explore this using the probability slider in this calculator.

The paradox is a powerful reminder that how a question is asked shapes the answer in statistics and science. It has real implications for interpreting medical test results, survey data, legal evidence, and scientific studies. Any time a condition filters or selects observations, the framing of that filter changes the resulting probabilities.