Seeing a gold coin makes a second gold coin more likely than you think. The Bertrand’s box paradox calculator takes the number of gold-gold, mixed and silver-silver boxes and gives the chance that the second coin is gold, the chance of starting with gold, the gap with the naive 50 % and the expected number of draws of each kind over many repetitions.
How the calculation works
- Enter how many boxes hold two gold coins, one gold and one silver coin, and two silver coins. The classic puzzle has one box of each kind.
- Pick a box at random and a coin at random from it. Suppose the coin you see is gold. The calculator gives the chance that the other coin in the same box is gold too.
- The answer is 2a ÷ (2a + b), where a is the number of gold-gold boxes and b the number of mixed boxes. With one box of each kind it is 2/3, not the 1/2 that most people expect.
- Enter a number of repeated draws to see how many of them start with gold, and how many of those end with a second gold coin.
Example
Three boxes hold two gold, one gold and one silver, and two silver coins, and you repeat the draw 300 times. Half of the draws, 150, start with a gold coin, because 3 of the 6 coins are gold. Those 3 gold coins are the two from the gold-gold box and one from the mixed box, so the second coin is gold in 2 out of 3 cases, that is 66.67 %, or 100 draws out of 150.
Why it matters
The puzzle is a classic way to teach conditional probability. It shows why counting coins, not boxes, is the right way to reason, and the same idea applies to medical tests, quality control and any case where you learn something about an item before judging the rest.
Good to know
The result assumes that the box is chosen at random and then a coin is chosen at random inside it, with every coin equally likely. If the coin were chosen in a different way, for example by always showing the gold one first, the probabilities would change. The repeated draws give expected numbers, not exact outcomes.
Related tools
The Monty Hall calculator works through another famous counterintuitive puzzle, and the Bayes’ theorem calculator updates a probability after new evidence.