Problem: The Birthday Wheel
How many spins before some number repeats? Far fewer than most people guess.

The birthday problemmathBirthday problemThe probability that any two observations match. On independent fair 37-pocket spins, eight spins give a better-than-even chance of a repeat.Read full definition → on a 37-pocket wheel: after just 8 spins there is a better-than-even chance that some number has already appeared twice. Players read repeats as patterns; they are the expected behavior of a fair wheel.
The problem
Assume independent spinsmathIndependent spinsSpins for which learning an earlier outcome does not change the probabilities assigned to a later one.Read full definition → on an unbiased 37-pocketequipmentPocketA compartment on the rotor in which the ball can settle. The number of pockets depends on the wheel.Read full definition → wheel. How many spins until the probabilitymathProbabilityA measure of how likely an event is, from zero to one. Percentages express the same scale from 0% to 100%.Read full definition → that at least two spins have shown the same number exceeds one half? The answer is 8: any pair can match, so you do not need to wait for a particular pocket.
The solution
Compute the probability of no repeat and subtract from one. After k spins, p(all different) = (37/37)(36/37)(35/37)…((38 − k)/37). For k = 8: 1 × 0.973 × 0.946 × 0.919 × 0.892 × 0.865 × 0.838 × 0.811 ≈ 0.44, so p(some repeat) ≈ 0.56. For k = 14 the repeat probability is about 0.941; for k = 20 it is about 0.9983. The intuition fails because the number of pairs of spins grows as k², and any pair can match.
Why it matters at the table
Under this model, a results board of 20 numbers shows a repeat about 99.8% of the time. Players see “17 came twice in ten spins” and conclude 17 is hot; the birthday arithmetic says a repeat within ten spins happens 74% of the time and the repeated number is random. A repeat alone does not show that the repeated pocket is more likely next time.
The follow-up problem
How many spins until a specific number, say 17, has appeared twice? That is a different question with a much larger answer: the expected wait is 74 spins, and the probability of two 17s within 20 spins is about 10%. These figures use different summaries: 8 is the median wait for any repeat, while 74 is the mean wait for a named pocket to appear twice. The mean wait for any repeat is only about 8.3 spins. Specifying the number in advance changes the question.
Try it
Run the Demo Game for ten spins and look at the history. There will usually be a repeat. Run it again. The repeated number may change or happen to be the same. Repeating this experiment estimates the model probability; a small sample can differ substantially.
Formula
Questions
Exactly, with 37 pockets instead of 365 days. 23 people give 50% for birthdays; 8 spins give 56% for the wheel.
The probability of some pocket appearing at least three times first exceeds 50% at 21 spins: about 53.6%. At 20 spins it is about 48.5%, under the independent uniform model.
One ordinary repeat is weak evidence because repeats are common on an unbiased wheel. Detecting bias requires a defined test and enough observations for the size of bias you are trying to detect; no fixed number of spins automatically proves it.
Fun with Math
Try these yourself first. Each solution is folded; open it when you are ready.
A results board beside an American wheel shows the last 16 spins. How likely is it that no number appears twice on it?
Show the solution
Answer About 2.5%: roughly 39 boards in 40 show at least one repeat.
Go spin by spin. The first spin can be anything. The second must miss it: 37 chances in 38. The third must miss both: 36 in 38. The sixteenth must miss fifteen numbers: 23 in 38.
Multiply: 38/38 × 37/38 × 36/38 × … × 23/38 ≈ 0.0246.
A board with no repeat is the rare one. A repeat on such a board is common under the independent uniform model, so it is weak evidence of bias by itself.
In ten spins of a European wheel, what is the chance that 17 comes up exactly five times? That some number, any number, comes up five or more times? And that some number comes up at least three times in four spins?
Show the solution
Answer 17 exactly five times: 1 in 315,576. Some number five or more times: about 1 in 8,333. Some number three or more times in four spins: 1 in 349.
17 five times: choose which five of the ten spins, 252 ways, each with chance (1/37)⁵ × (36/37)⁵. Together, 1 in 315,576.
Some number: each of the 37 numbers has a chance of 1 in 308,325 of five or more hits, which gives 37 in 308,325, or 1 in 8,333. The rare case of two numbers with five hits each is counted twice, but taking it out changes nothing at this precision.
Three times in four spins: for one number, 4 × 36 + 1 = 145 of the 37⁴ possible results; for any of the 37 numbers, 145/37³ = 1 in 349.
For a double-zerowheelZeroThe 0 outcome, conventionally shown in green. Ordinary color, parity, high/low, dozen and column bets do not include it.Read full definition → wheel, the corresponding figures are 1 in 359,275, 1 in 9,243 and 1 in 368. The gap between “17” and “some number” is the birthday problemmathBirthday problemThe probability that any two observations match. On independent fair 37-pocket spins, eight spins give a better-than-even chance of a repeat.Read full definition → again: there are 37 ways to be surprised.
Comic relief

In eight independent spins on a fair 37-pocket wheel, the chance of any repeated number is about 56%.
Sources
- Feller, W. An Introduction to Probability Theory and Its Applications, Vol. 1.
- Wizard of Odds. “Ask the Wizard”, roulette questions.