The wheel does not remember the last spin
A roulette ball has no memory. The wheel, the ball, and the pockets do not store information about where the ball landed ten seconds ago or ten hours ago. Each spin is an independent event in the sense that the physical system is reset: the wheel is spinning one way, the ball is launched the other way, and the outcome is a product of speed, friction, the frets between pockets, and a cascade of tiny mechanical variations that no player can measure precisely enough to predict.
This independence is exactly what the game's mathematics assumes. The probability of any single number landing on a European wheel is one in thirty-seven on every spin, regardless of whether that number has just hit three times in a row or has not appeared for two hundred spins. The wheel does not become due to produce a red after a run of blacks, and a number that has just hit is not any less likely to hit again on the next spin than it was before. The past sequence tells you nothing about the next outcome.
What players often describe as a pattern is the brain finding structure in a random sequence. Ten reds in a row looks extraordinary when it happens, but a sequence of ten spins has more than a thousand possible colour combinations, and any specific sequence is exactly as likely as any other. The sequence red-black-red-black-red-black-red-black-red-black is no more probable than ten reds. What differs is only how memorable each sequence feels.
Two sources of randomness: physical wheels and digital generators
Where you play changes what produces the random outcome, but not the statistical structure of the game.
A physical wheel at a table is a mechanical randomiser. The croupier launches the ball, the wheel and ball move in opposite directions, and the ball bounces off the frets before settling. The outcome is deterministic in the strict physical sense that a complete knowledge of every force involved would, in principle, allow a prediction. In practice the system is chaotic: tiny differences in launch speed, wheel speed, humidity, ball wear, and pocket geometry produce wildly different outcomes. No player at the table has access to enough information to exploit this, and venues monitor wheels for bias and re-level or replace them when wear patterns appear.
An online or electronic roulette game uses a random number generator, the algorithm that determines game outcomes. The generator produces a stream of numbers that are then mapped onto the wheel's pockets. Independent testing laboratories are commonly used to certify RNG behaviour; certification covers game mechanics, not an operator's financial conduct. A certified generator is tested for statistical randomness, meaning its output over millions of simulated spins matches the distribution the wheel's geometry implies. This certification does not make the game fair in the sense of profitable for the player; it addresses whether the outcomes are being generated without a hidden pattern.
| Property | Physical wheel | RNG-based game |
|---|---|---|
| Outcome source | Ball and wheel physics | Algorithm mapped to wheel pockets |
| Independence of spins | Yes, in practice | Yes, by design |
| Predictable by a player | Not with available information | Not when the generator is properly certified |
| What certification addresses | Wheel balance and pocket geometry | Statistical distribution of the generator's output |
| What neither addresses | The built-in house edge on every bet | |
Random outcomes do not mean fair payouts
Here is the distinction that answers most of what people are really asking when they ask whether roulette is completely random. The outcome of each spin is random. The payout schedule is not. The two facts live in different layers of the game, and confusing them is how players end up believing a random game should be beatable.
A European wheel has thirty-seven pockets. A straight-up bet on a single number pays thirty-five to one. If the payouts were perfectly fair, a winning single-number bet would pay thirty-six to one, because there are thirty-six losing pockets and one winning pocket. The missing unit on every winning payout is the house edge made concrete. The ball is equally likely to land in any pocket, but the payout is calculated as if there were one fewer losing pocket than there really is.
The same structure appears on every bet. An American wheel has thirty-eight pockets because it adds a double zero alongside the single zero. The extra pocket increases the house edge on the American layout. The payout for a single number is still thirty-five to one, but now there are thirty-seven losing pockets. The wheel is still random. The house edge has simply been widened by adding one more pocket that pays out as if it were not there.
Return to Player (RTP) is the other side of the same coin. A theoretical percentage of total wagered money a game returns to players over a very large number of spins or hands. It is a long-run statistical average, not a per-session guarantee. House edge and RTP are complements: they sum to 100% for the same game. The house edge is the share of each wager the game's payout structure withholds on average; the RTP is the share it returns. Neither figure tells you what will happen on your next spin or your next hundred. Both describe what the mathematics implies when the number of spins grows very large.
Why a short session looks nothing like the long-run average
Suppose a European wheel returns an RTP of R%, which means the house edge is (100 minus R)%. Those are long-run figures. A player who sits down for an hour and makes sixty spins is not going to observe R% of their stake returned with any reliability. The variance around the average is enormous at small sample sizes.
As an example, suppose you bet one unit on red for one hundred spins. The expected number of red hits is the probability of red multiplied by one hundred. But the actual number will usually differ from that expectation by a margin that shrinks only as a proportion of the total, not in absolute terms. A session of one hundred spins can easily finish well above or well below the expected return. A session of ten thousand spins will finish closer to the expected return as a percentage of total wagered, because the random fluctuations average out over more trials. This is the law of large numbers doing its quiet work, and it is also why a player who wins on a short session has not beaten the game. They have experienced a favourable run of variance.
Volatility in roulette is lower than in a high-variance slot, but it is still present. High volatility means rarer but larger payouts; low volatility means more frequent but smaller ones. Volatility does not change a game's RTP. A straight-up number bet has high volatility because it pays thirty-five to one and hits rarely. An even-money bet has lower volatility because it hits close to half the time and pays one to one. Both bets carry a house edge, and no amount of switching between them changes the long-run expectation.
Decisions per hour change total exposure, not the edge
One thing a player genuinely controls is how much total money passes across the table. The house edge applies to each wager. The more wagers placed, the more total exposure to that edge. This is not a strategy for winning; it is arithmetic about how losses accumulate.
Suppose a player bets ten units per spin and plays forty spins an hour. That is four hundred units of total wager per hour. If the house edge on each bet is E%, the expected loss per hour is E% of four hundred units. Double the bet size or double the spins per hour and the expected loss doubles, because the edge is applied to a larger total. Slow the game down, bet smaller, or leave earlier, and the expected loss shrinks for the same reason. None of this changes the edge per bet. It changes how many times the edge is collected.
This is the honest version of bankroll management. Sizing bets and limiting session length are methods for controlling how much total money is exposed to a negative-expectation game. They are not methods for turning the game positive. A player who bets small and leaves after an hour has a smaller expected loss than a player who bets large and stays for six hours. Both players face the same house edge on every spin.
Why progression systems redistribute outcomes
Betting systems that increase the stake after a loss, like doubling, are popular because they appear to guarantee recovery. They do not. What they guarantee is a different distribution of outcomes: many small wins and occasional very large losses, instead of a smoother pattern of small wins and small losses. The expected loss remains the house edge multiplied by total wagered, because the system does not change the probability of any spin or the payout on any winning bet.
As an example, suppose a player doubles a losing even-money bet after each loss, starting at one unit. After five consecutive losses the next bet would be thirty-two units, and the player would have staked sixty-three units in total to chase a one-unit recovery. The probability of five consecutive losses on an even-money bet is not zero, and when it happens the required next bet grows quickly. Table limits exist in part to cap exactly this escalation. The system does not remove the house edge; it concentrates the same expected loss into a smaller number of catastrophic sessions and spreads small recoveries across many ordinary ones.
The wheel is random. The payouts are not. No sequence of bet sizes can change either fact.
What a player can check
If you want to verify that the game you are playing is random in the sense that matters, there are concrete places to look. For an online game, the game information panel or help screen usually states the RTP and the rule set, and the operator's terms page states which testing laboratory, if any, certifies the RNG. For a physical table, the wheel's layout and the posted table rules state the payout schedule, and the venue's own procedures govern wheel inspection and replacement. None of these documents will promise you a win. What they let you do is confirm the structure of the game before you decide whether to play it.
FAQ
Can a roulette wheel be biased?
A physical wheel can develop mechanical bias if a pocket, fret, or level is worn or damaged, which would make some pockets appear more often than pure chance would predict. Venues monitor for this and re-level or replace wheels when wear patterns appear. Detecting a usable bias before the venue corrects it is not something a casual player can do from a few observed spins.
Is online roulette rigged to make you lose?
An online roulette game does not need to be rigged to produce a house profit. The payout schedule already pays winning bets slightly less than the true odds, so the house edge is collected on aggregate across all players without any manipulation of individual spins. Certification by an independent testing laboratory addresses whether the RNG is producing a statistically random distribution; it does not remove the built-in edge.
Does the same number ever hit twice in a row?
Yes, and the probability is exactly what the wheel's geometry implies. On a European wheel, the chance that a specific number hits on the next spin is one in thirty-seven, regardless of what just happened. A number hitting twice in a row is uncommon but not evidence of a fault or a pattern, and it does not make any future outcome more or less likely.
If you or someone you know has a gambling problem, help is available. In the US, call or text the helpline printed in the footer below. In the UK, call 0808 8020 133. In Australia, call 1800 858 858. In Canada, contact your provincial helpline. Please gamble responsibly.
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.