Back to news
History and probabilityFull storyJun 25, 2026 / 3 min readUpdated:

García-Pelayo and biased roulette wheels: what the data must show

Separate physical wheel bias from hot-number betting. Work through the break-even probability and learn why a fresh sample matters.

Share

A claim about a wheel, not a betting progression

The García-Pelayo story concerns a family looking for non-uniform outcomes on particular roulette wheels. That is a different claim from saying that a number is due, or that doubling a stake changes its chance of winning. A mechanical explanation must be specific to the equipment and conditions being observed. It cannot simply be carried over to a different wheel or an online random-number generator.

Work out the break-even probability first

A one-unit straight-up bet returns a net profit of 35 units when it wins and loses one unit otherwise. If the actual chance of winning is p, its expected net result is 35p − (1 − p) = 36p − 1. Break-even therefore requires p = 1/36. On an ideal European wheel the chance is only 1/37, giving an expected loss of 1/37 of the amount wagered. The calculation describes an underlying probability, not a guarantee about the next spin.

Why the hottest number is not enough evidence

If you inspect all 37 numbers and choose the one with the most hits, you are selecting a winner after seeing the data. Random samples also have a highest count. A useful test sets its rule in advance, records all outcomes and checks the proposed effect in a separate sample. Mixing different wheels or changing the rule halfway through prevents a clean comparison. A historical account is not a substitute for the original complete dataset, including losing sessions.

Key Takeaways

  • Changing the stake cannot create mechanical wheel bias.
  • An observed frequency is an estimate, not the true probability.
  • Independent verification matters more than a memorable winning session.
Roulette simulator lesson

Pick a frequent number from one random session, then test it in a fresh session without choosing again. Record both successes and failures. This demonstrates selection bias; it does not reproduce a physical defect.

Also read

Joseph Jagger and the wheel that broke Monte Carlo

Joseph Jagger was a Yorkshire textile man who understood spinning machinery. Around 1881 he went to Monte Carlo, watched roulette tables, and looked for imperfect balance. When he found numbers appearing more often than expected, he bet on them and reportedly ...

Read full story
Interesting strategy

Fibonacci

A sequence-based progression that moves forward after losses and back after wins.

Fibonacci