Roulette: Probability and Independent Results

Start with the wheel, not the color labels

Roulette is a wheel-and-ball game whose numbered pockets define a set of possible results. A single-zero wheel includes the numbers 1 through 36 and zero. A double-zero wheel includes an additional pocket. Pennsylvania’s equipment rules document both arrangements; they are useful examples, not a description of every online product.

For the calculations below, assume a single-zero wheel with equally likely pockets and independent spins. Those assumptions matter. A rendered wheel on a screen does not prove a particular probability distribution, and a photograph of physical equipment cannot establish its fairness.

Sources: Pennsylvania Code — 58 Pa. Code § 617a.1

Why two colors do not make a 50–50 event

The familiar red-versus-black framing omits zero. There are 18 red numbers and 18 black numbers, but the denominator is 37. Dividing 18 by 37 gives about 48.65%, rather than 50%. This is a counting exercise, not a recommendation about a wager.

Zero is one outcome out of 37, or about 2.70%. Adding the three unrounded fractions gives one: 18/37 + 18/37 + 1/37 = 37/37. If a description removes an outcome from the denominator without removing it from the actual process, its probability claim is incomplete.

Sources: Pennsylvania Code — 58 Pa. Code § 617a.1

The next spin does not remember the last one

In the independent model, conditioning on a run of earlier red results does not change the next red probability: it remains 18/37. The same is true after alternating colors or a zero. Independence is a statement about the random process, not a claim that short records must look balanced.

Imagine writing a sequence on paper before observing ten spins. Any one exact sequence of ten numbered pockets has probability (1/37) to the tenth power under this model. A sequence containing repetition and a sequence without repetition have the same probability when each is specified exactly in advance.

That does not mean all broad categories of sequences are equally common. There are many more ways to form some patterns than others. Confusing one exact sequence with a category containing many sequences is another way that intuitive probability arguments go wrong.

What a history display can and cannot show

A results panel can document recorded outcomes. It cannot, by itself, show that a color is hot, a number is due or a cycle is about to restart. A claim of prediction would need a defined model and evidence that works on data not used to invent the claim.

For remote products within its jurisdiction, the UK Gambling Commission requires simulated physical devices to reflect the corresponding probabilities. That provides a technical example of what faithful simulation means. It does not certify a sweepstakes-branded wheel or establish its regulatory status.

Sources: UK Gambling Commission

Probability is not a recovery plan

Knowing the denominator does not remove financial or emotional risk. WHO describes gambling harm that extends beyond a clinical diagnosis. A belief that the next result must reverse a loss is not supported by independence.

A percentage describes uncertainty under stated assumptions. It does not promise a sequence, guarantee fairness, or measure whether spending is affordable. Those limits remain even when the arithmetic itself is correct.

Sources: World Health Organization

Questions & answers