Bertrand Box Paradox Calculator - Conditional Probability Puzzle

The bertrand box paradox calculator resolves a classic conditional probability puzzle that has misled sharp mathematicians since Joseph Bertrand published it in 1889. The setup is three boxes of two coins each: one box holds two gold coins, one holds two silver coins, and one holds one gold and one silver. You choose a box at random and draw one coin without looking at the other.

• • Students learning conditional probability: Replace a hand-wavy explanation with a live number that updates as the box counts change.
• • Teachers demonstrating Bayes rule: Show why counting coin faces, not boxes, is the step that fixes the intuition.
• • Puzzle solvers checking a hunch: Test the classic setup and then stretch it to odd box combinations to see the rule hold.

Most people guess that if the coin you see is gold, the box is equally likely to be the all-gold box or the mixed box, so the hidden coin is gold with one-half probability. The correct answer is two-thirds, because a gold draw is twice as likely to have come from the all-gold box than from the mixed box. This tool works the conditional probability out for any mix of boxes you set, so you can watch the two-thirds result emerge from the classic setup and then test stranger combinations.

The trap is that we tend to compare boxes when we should compare coin faces. A randomly chosen box and a randomly chosen side make each individual coin face equally likely to appear, and three of the six faces are gold. Counting faces instead of boxes is the single move that turns the wrong one-half answer into the right two-thirds, and the calculator makes that counting visible rather than abstract.

The puzzle is not a trick about trick coins. Every box and every side is chosen fairly; the surprise comes purely from how the evidence, seeing a gold coin, reweights the possibilities. Once you accept that a gold draw rules out three of the six faces, the rest is just careful counting, which is exactly what the result panel does for you.

Before trusting the two-thirds answer, review the basic rules of chance with the probability calculator.

The calculator treats every coin face as equally likely, because a randomly chosen box and a randomly chosen side give each of the six faces the same shot at being the one you see. Three of those faces are gold: two in the all-gold box and one in the mixed box. The all-silver box contributes no gold faces at all, so it drops out of the conditional calculation completely.

Once a gold face is showing, the question is which box it came from. Two of the three gold faces belong to the all-gold box, so the chance the hidden coin is gold is two out of three. The mixed box contributes only one gold face, so it accounts for the remaining third. The same logic scales to any counts you enter: with a all-gold boxes, b all-silver boxes, and c mixed boxes, the all-gold box owns 2a of the (2a + c) gold faces.

Writing the rule as a fraction makes the scaling obvious. The numerator counts gold faces that came from all-gold boxes, and the denominator counts every gold face across all boxes. Divide them and you have the posterior probability that the drawn-gold box was the all-gold box, which is exactly the chance the companion coin is also gold.

The overall draw rates are a separate question. Before you look at any coin, the chance of seeing gold is the total number of gold faces divided by all faces, or (2a + c) over 2(a + b + c). That number answers how often a gold coin appears, whereas the conditional result answers what you learn after one already shows up.

According to Wikipedia, Bertrand's box paradox was posed by Joseph Bertrand in 1889 to show how conditional probability overturns first instincts.

A single random draw builds intuition, so try the coin flip probability calculator to compare one-coin odds against this puzzle.

Three ideas carry the whole puzzle: equally likely faces, conditioning on what you observed, and the gap between box odds and coin-face odds. Keep these straight and the two-thirds answer stops feeling strange.

The conditioning step is the part people skip. Before you see any coin, each box is one-third likely, but the moment a gold coin appears you throw away every outcome where a silver face showed up. That single act of discarding outcomes is what reweights the boxes, and it is the same step behind every Bayes-rule update in medicine, spam filtering, and forensic science.

Box odds and coin-face odds are not the same thing, and confusing them is the heart of the paradox. Choosing a box uniformly gives each box equal prior weight, but drawing a coin then favors the box that supplies more of the observed face. The calculator keeps those two stages separate so the reweighting is visible instead of hidden.

When you repeat the box draw many times, the binomial distribution calculator models how often a gold coin appears across trials.

Enter your box counts and read the result panel. The defaults reproduce the classic three-box setup, so you can confirm the two-thirds answer before experimenting.

1. 1 Enter all-gold boxes: Type how many boxes hold two gold coins; use 1 for the classic setup.
2. 2 Enter all-silver boxes: Type how many boxes hold two silver coins; use 1 for the classic setup.
3. 3 Enter mixed boxes: Type how many boxes hold one gold and one silver coin; use 1 for the classic setup.
4. 4 Press Calculate: Read the chance the other coin is gold, the chance the box is mixed, and the overall gold and silver draw rates.
5. 5 Vary the counts: Change the numbers to test lopsided setups, such as several all-gold boxes and no mixed boxes.

With 2 all-gold boxes, 1 all-silver box, and 0 mixed boxes, every gold draw must come from an all-gold box, so the other coin is gold 100% of the time and the overall gold-draw rate rises to about 66.67%.

The paradox hinges on multiplying joint events, so the AND probability calculator clarifies the multiplication step in Bayes rule.

Working the paradox by hand is tempting to get wrong because the naive one-half answer is so tempting and the arithmetic is fiddly once you leave the classic three-box case.

• • Removes arithmetic errors: The result panel does the face-counting so attention stays on the conditioning logic.
• • Supports live experimentation: Changing box counts shows how the two-thirds answer generalizes, not just the classic case.
• • Builds transferable intuition: The same counting habit applies to screening tests and other conditional-probability problems.

The bertrand box paradox calculator removes that arithmetic so you can focus on the reasoning and on seeing how the result moves as the box mix changes, rather than on whether you divided by the right number. It turns a memorable but slippery fact into something you can verify in seconds.

It also makes the lesson portable: a teacher can show the classic case, then immediately test a student's invented setup and watch the conditional probability respond in real time. That feedback loop is where the intuition actually sticks, because the learner sees the two-thirds answer survive every reasonable variation instead of memorizing one number.

Beyond the classroom, the same structure shows up whenever a positive result must be turned into a probability about the underlying state, from medical screening to quality control on a production line. Practicing on coins builds the habit of counting the favorable cases among the observed cases, which is the habit the harder problems also reward.

Because the paradox is one conditioning step inside Bayes rule, the joint probability calculator shows how the same draw splits into the coin you see and the coin you don't.

Three counts drive every number on the panel. Adjust them and you can see exactly which part of the puzzle each one controls.

• • The two-thirds result assumes a fair random box and coin draw; a biased pick changes the answer.
• • The conditional step only applies once a gold coin is observed; knowing only that a box was chosen reverts the odds to the box mix.

The all-gold and mixed counts set the conditional result, because they are the only boxes that can produce a gold draw. The all-silver count never contributes a gold face, so it changes only the overall draw rates, never the answer to what the other coin is once gold is showing.

When you add all-gold boxes, the other-coin-gold chance climbs toward 100%, because more gold faces come from boxes where the partner is also gold. Adding mixed boxes pushes it down toward 50%, since each mixed box adds one gold face whose partner is silver.

According to Wikipedia, the paradox updates the prior box probabilities by the likelihood of drawing gold from each box type.

When the result surprises you, the expected value calculator shows the average coin you would expect across many repeated draws.