The Wheel Has No Memory
A player watches red come up five times in a row and pushes a larger stack onto black. The reasoning feels solid: the wheel has to balance out. Red cannot keep winning forever, so black is due. That reasoning is the law of averages, and on a roulette wheel it has no mechanical basis at all.
The law of averages is a popular intuition, not a mathematical principle. It suggests that outcomes in a random process will even out over time, and that short-term deviations from the average will be corrected by opposite deviations soon after. In games of independent chance, that correction never happens. The wheel does not track what it landed on last spin. It has no mechanism for remembering red, no obligation to produce black, and no way to know what a player bet on.
What evens out over many spins is the proportion of outcomes, not the absolute difference between them. That distinction matters more than most players realise.
What the Law of Averages Claims, and What Probability Says
The law of averages is often confused with the law of large numbers. The two are not the same thing, and the difference is the entire answer to this question.
The law of large numbers is a genuine statistical theorem. It says that as the number of independent trials grows, the observed frequency of an outcome tends toward its theoretical probability. On a European wheel, red is expected on a fixed proportion of spins across a very large sample. The theorem does not say when that convergence will happen, how quickly, or that a run of one colour makes the other colour more likely next spin.
The law of averages, as most players use it, goes further. It claims that deviations must be corrected, and soon. That is false. A roulette spin is an independent event. The probability of red on the next spin is the same whether the last spin was red, black, or green. Nothing accumulates, nothing is owed, and no outcome becomes more likely because it has been absent.
| Concept | What it claims | What it means for roulette |
|---|---|---|
| Law of large numbers | Observed frequency approaches theoretical probability over many trials | Over a very large number of spins, the share of red outcomes drifts toward its expected proportion |
| Law of averages (popular version) | Short-term deviations must be corrected soon | False on a wheel; no spin changes the odds of the next |
| Gambler's fallacy | A specific outcome is due after a streak | The wheel has no memory and no balancing mechanism |
The House Edge Is Built Into Every Single Bet
Roulette does not need streaks, memory, or the law of averages to make money for the house. The advantage is built into the payout structure on every spin, before the ball is released.
On a European wheel there are 37 pockets: the numbers 0 through 36. A bet on a single number pays 35 to 1. If the game paid true odds, a single-number bet would pay 36 to 1, because there are 36 losing numbers and one winning number. The gap between the payout and the true odds is the house edge. It exists on every spin, regardless of what happened before.
An American wheel has 38 pockets because it adds a double zero. The extra pocket increases the house edge on the American layout compared with the European layout, because the payout for a winning number stays the same while the number of losing pockets grows.
Return to Player, or RTP, is the complement of the house edge for the same bet under the same rule set. If a bet carries a house edge of E percent, its RTP is 100 minus E percent. RTP is a long-run statistical average, not a per-session guarantee. A player can win for an evening while the RTP stays exactly where it is, because RTP describes the average over a very large number of spins, not the result of any particular sequence.
Why Short Sessions Look Nothing Like the Long-Run Average
Roulette outcomes are volatile. A player who sits down for fifty spins is experiencing a tiny sample of a process that only reveals its average over tens of thousands of spins or more. In a short session, almost anything can happen.
Suppose a player bets one unit on red for 100 spins on a European wheel. Red is expected to appear on a fixed proportion of spins in the long run, but in any given 100-spin sample the actual count can sit well above or well below that expectation. The player might finish ahead, behind, or roughly even. None of those outcomes tells them anything about whether the wheel is fair, whether the law of averages is working, or what will happen in the next 100 spins.
Variance is the term for how results are distributed around the average. High-variance games produce long losing stretches punctuated by large wins. Roulette is a high-variance game compared with something like a coin flip, because a single winning spin can pay 35 times the stake. That variance is what makes short-term results so disconnected from the long-run average. It is also what makes progression systems feel like they should work.
What Decisions Per Hour Do to Total Exposure
Every spin costs the player the house edge, on average. The faster the game, the more spins per hour, and the more total money is put through the wheel. That total is called coin-in or handle, and it is the number that matters for expected loss.
Expected loss is not a prediction of what will happen in any given session. It is the product of three things: the amount wagered per spin, the house edge on the chosen bet, and the number of spins. If a player bets one unit per spin on a bet with a house edge of E percent, and plays N spins, the expected loss is one unit times E percent times N. That arithmetic uses hypothetical inputs only; the actual edge on any given roulette bet is printed in the game's information panel or the operator's rules page.
Playing faster does not change the edge. It changes how many times the edge is applied. A player who makes 60 spins an hour exposes themselves to the edge twice as often as a player who makes 30 spins an hour. The edge itself stays fixed. What changes is the total amount wagered, and therefore the expected loss over the session.
Bet selection changes the edge, and roulette offers a range of edges on the same table. A bet on a single number carries a different expected loss than a bet on red or black, because the payout structure differs. The player's only genuine choices are which bets to make, how much to wager on each, how fast to play, and when to stop. None of those choices removes the edge. They only change how much of it the player pays per hour.
Why Progression Systems Redistribute Outcomes Instead of Removing the Edge
A progression system is any betting pattern that changes the stake based on previous results. The most famous is the martingale: double the bet after every loss, return to the base stake after a win. The logic is that a win must eventually arrive, and when it does, it recovers all previous losses plus one unit.
The logic is arithmetically sound in a world without table limits, without a finite bankroll, and without a house edge. None of those conditions hold on a real roulette table.
First, the house edge applies to every spin, including the winning one. The player is not paid true odds, so even a winning spin returns slightly less than the amount that would make the progression whole. Second, losing streaks long enough to exhaust a bankroll or hit the table maximum are not rare; they are guaranteed to occur eventually in any long sequence of spins. When they do, the progression collapses and the player loses the accumulated stake.
What a progression system does is reshape the distribution of outcomes. It creates many small wins and a small number of very large losses. The expected loss remains the same as flat betting, because every unit wagered still carries the house edge. The system does not remove the edge, reduce it, or make any spin more likely to win. It redistributes when the losses arrive.
| Approach | What changes | What stays fixed |
|---|---|---|
| Flat betting | Stake is constant every spin | House edge on every unit wagered |
| Martingale progression | Stake doubles after losses; many small wins, rare large losses | House edge on every unit wagered; expected loss per unit |
| Any progression system | Timing and size of wins and losses | House edge, RTP, and long-run expectation |
No betting pattern changes the built-in edge of a negative-expectation game. Progression systems, including doubling after losses and similar schemes, redistribute outcomes rather than remove the edge. A player who understands that can decide whether the reshaped risk profile is worth the same expected loss, but they cannot turn a negative-expectation game into a positive one by changing stake sizes.
What a Player Can Control
The honest list of what a roulette player controls is short. It includes bet selection, stake size, speed of play, session length, and stopping rules. It does not include the wheel's memory, because the wheel has none.
Bet selection matters because different bets on the same table carry different house edges. On a European wheel, the zero pocket is what gives the house its edge on the outside bets, and the payout gap is what gives the house its edge on the inside bets. The player can read the table layout and the game's information panel to see the payout for each bet, and compare it with the true odds implied by the number of winning and losing pockets.
Stake size and speed of play determine total exposure. A player who wants to limit expected loss can bet smaller, play slower, or both. Stopping rules matter because a session is a finite sample. A player who stops after a fixed number of spins, or after a fixed win or loss limit, is choosing the size of the sample they will experience. They are not changing the odds of any spin.
What the player cannot control is the edge itself. It is built into the payout table, and it applies to every spin, every bet, and every progression. The law of averages does not rescue a losing session, and it does not punish a winning one. It simply does not operate on a roulette wheel.
FAQ
Does the law of averages apply to roulette in the long run?
The law of large numbers applies in the long run: the observed proportion of outcomes tends toward the theoretical probability over a very large number of spins. The law of averages, as a claim that short-term deviations must correct soon, does not apply at all. The wheel has no memory and no balancing mechanism.
Can a roulette betting system overcome the house edge?
No. A betting system changes the size and timing of wagers, not the edge on each wager. Progression systems create many small wins and occasional large losses, but the expected loss per unit wagered stays the same. The house edge is built into the payout structure and applies to every spin.
Why do some players believe the law of averages works on roulette?
Because short sessions are noisy. A player who sees red win five times in a row feels that black is due, and occasionally black does win next. That occasional confirmation reinforces the belief, even though the probability of black was unchanged. The pattern is a cognitive bias, not a property of the wheel.
Play responsibly
Gambling should be treated as paid entertainment, never as a way to earn income or recover losses. 1-800-MY-RESET (1-800-697-3738)
18+ or 21+ depending on where you are; follow the minimum age that applies to you.
This article is general information about how these mechanics work. It is not legal advice and not a recommendation to gamble or to use any particular operator. Availability and legality differ by jurisdiction — check the rules that apply where you are.