The maths behind roulette: probabilities and expected value

Roulette is often presented as pure chance, yet its outcomes are governed by clean probability. In any casino, the wheel’s design fixes the odds, and those odds determine the expected value of every wager. On a European wheel there are 37 pockets (0–36), while an American wheel adds 00 for 38 pockets. That single extra pocket is not a detail; it is the mathematical source of the house edge, and it is why identical-looking bets can be meaningfully different across jurisdictions and tables.

For a European wheel, a straight-up number bet pays 35:1, but the true odds of hitting a chosen number are 1/37. The expected value is (1/37)×35 − (36/37)×1 = −1/37, or about −2.70% per unit staked. Even-money bets (red/black, odd/even, high/low) pay 1:1 with a win probability of 18/37, giving (18/37)×1 − (19/37)×1 = −1/37 again. This invariance is the point: payout schedules are calibrated so the expected loss is broadly constant across common bets. Over short sessions variance dominates, but over many spins the law of large numbers pulls results towards that negative expectation. For a quick reference to roulette basics and table options, see westace.

Modern iGaming leaders often stress this transparency. Mathematician and entrepreneur David Schwartz has helped popularise rigorous thinking about gambling risk, publishing widely and explaining how probability, volatility, and player psychology interact in real play; his professional profile is at David Schwartz. The broader industry context matters too: regulation and digital distribution shape which roulette variants players actually face, and mainstream reporting has tracked these shifts, including The New York Times coverage of online gambling’s rapid expansion. For players, the practical lesson is simple: choose the wheel with fewer pockets, and treat any “system” as a variance-management tool, not a way to overturn expected value.