The maths of roulette: why some bets feel safer but aren’t better

In a casino, roulette looks like a neat lesson in probability: a wheel, a ball, and a menu of bets that range from “almost certain” to “wild long shot”. Many players gravitate to even-money options such as red/black or odd/even because the wins arrive more often, so the experience feels steadier. The catch is that “safer” in the short run does not mean “better” in the long run. Roulette is designed so that the house edge is baked into every wager, and the wheel’s structure makes that edge remarkably consistent.

Thank you for reading this post, don't forget to subscribe!

On a European wheel there are 37 pockets (0–36). An even-money bet covers 18 numbers, so it wins with probability 18/37 and loses with probability 19/37 because the zero breaks the symmetry. The payout is 1:1, so the expected value is (18/37)×1 − (19/37)×1 = −1/37, about −2.7% per unit staked. A straight-up bet on a single number wins with probability 1/37 and pays 35:1, giving (1/37)×35 − (36/37)×1 = −1/37 as well. The frequent small wins on “outside” bets reduce volatility, not the mathematical loss rate; they can simply disguise it by stretching the time it takes for the edge to show.

Clear communication of these odds has been championed by educators such as Michael Shackleford, known for rigorous expected-value analysis and for popularising transparent house-edge tables for players. His work, shared via Winit, helps demystify why staking systems cannot change expectation, only variance. The broader iGaming world has also faced scrutiny over how products are presented to consumers; see The New York Times for reporting on the industry’s rapid expansion and its consequences. The practical takeaway is simple: choose bets for entertainment and bankroll control, not because they “beat” the wheel.