Roulette Online PRO
Mathematics · Math problem · 7 min

Problem: How Long Until Zero

Zero hits one spin in 37 on average. How long do you actually wait, and what does that do to an even-money bettor?

Short answer

The wait for zero is geometric: mean 37 spins, median 25, and one wait in twenty exceeds 110 spins. Over a 200-spin session an even-money bettor expects 5.4 zeros and will see anywhere from 1 to 10.

mean wait
37 spins
median wait
25 spins
p(no zero in 100 spins)
6.5%
zeros in 200 spins
5.4 ± 2.3

The problem

ZerowheelZeroThe single green pocket that no even-money, dozen or column bet covers; the source of the house edge. Also: single zero, 0.Read full → has probabilitymathProbabilityChance of an outcome, from 0 to 1 or 0% to 100%. One number: 1/37 = 2.70%.Read full → 1/37 per spin. (a) What is the probability of no zero in the next 37 spins? (b) What is the median wait? (c) Over a 200-spin even-money session, how many zeros should a player expect, and what is the range?

Solutions

(a) (36/37)³⁷ = 36.3%. More than a third of the time, 37 spins pass with no zero at all, which is the same arithmetic as “at least one hit on a straight upbetStraight upA single number, pays 35:1.Read full → in 37 spins” being only 63.7%. (b) The wait W is geometric; p(W > k) = (36/37)ᵏ; the median is the k where this crosses 0.5, which is k ≈ 25. Half the waits are 25 spins or shorter and half longer; the mean of 37 is pulled up by the long tail. (c) Zeros in 200 spins are binomial with mean 200/37 = 5.4 and standard deviationmathStandard deviationSquare root of variance; the typical distance of a result from the expected value.Read full → √(200 × (1/37) × (36/37)) ≈ 2.3. A 95% range is roughly 1 to 10.

wait for zeroprobability
within 10 spins24.0%
within 25 spins49.6%
within 37 spins63.7%
within 50 spins74.6%
within 100 spins93.5%
more than 110 spins≈ 5%
more than 200 spins0.4%
p(W ≤ k) = 1 − (36/37)ᵏ. Mean 37, median 25.

What zero costs an even-money bettor

Each zero costs one stake, and zeros are the whole edge on redbetRedEven-money bet on the 18 red numbers. French: rouge.Read full →: over 37 average spins the 18 wins and 18 losses cancel, and the one zero is the −1/37. In a 200-spin session at $10, the expected 5.4 zeros cost $54, which is exactly 200 × $10 / 37. A session with 2 zeros has had $34 of good luck; a session with 9 zeros has had $36 of bad luck, and will feel rigged. On a la partageruleLa partageFrench rule: when zero hits, half of an even-money stake is returned. Halves the house edge on those bets to 1.35%.Read full → table each zero costs half a stake and the figures halve.

Zeros in a 200-spin session% of sessions
0–1
2.5%
2–3
18%
4–5
34%
6–7
28%
8–9
13%
10 or more
4.5%
% of sessions
01734%
Binomial(200, 1/37). The $10 red bettor pays $10 per zero; the spread in zeros is the spread in the evening’s price.

La PartageMyth: Online Roulette Is Rigged

The follow-up

What is the probability of two zeros in a row? (1/37)² = 1/1,369. In a casino running 20 tables at 50 spins an hour, that is about one double zerovariantDouble zeroThe 00 pocket of the American wheel. Its presence raises the house edge from 2.70% to 5.26%.Read full → somewhere every 1.4 hours. Players who see it conclude the wheel is broken; the arithmetic says it is Tuesday.

Try it

The Session Simulator counts zeros per session. Run 100 spins a few hundred times and plot the count: the histogram above will appear, with its long right tail. The tail is where the stories come from.

Formula

p(W > k) = (36/37)ᵏ
E[W] = 37 · median ≈ 25
zeros in N spins ~ Binomial(N, 1/37)

Questions

Is zero more likely on some wheels?

Not on a fair one. On a biased wheellabBiased wheelA wheel whose physical defects make some pockets more likely than 1/37. Rare, detectable only with thousands of recorded spins.Read full → any pocketequipmentPocketOne of the 37 numbered compartments on the rotor. Also: slot, cup.Read full → could be, and it would take thousands of spins to tell.

Should I always put a chip on zero as insurance?

It is a 2.70% bet like every other. It changes the shape of the session (a zero becomes a win) and not the cost.

Why is the median less than the mean?

Because the distribution of waits is skewed: many short waits, a few very long ones.

Comic relief

Mean 37, median 25, and a long tail where the legends live.

Mean 37, median 25, and a long tail where the legends live.

Sources

  1. Feller, W. An Introduction to Probability Theory and Its Applications, Vol. 1. Wiley, 3rd ed., 1968. Gambler’s ruin, chapter XIV.
  2. Ethier, S. N. The Doctrine of Chances: Probabilistic Aspects of Gambling. Springer, 2010. Chapter 13, roulette.
reviewed 2026-09-28model: European, single zeromethodology ↗sources ↗