A player staring at a roulette layout sees red and black split almost evenly, and the natural assumption is that a colour bet should be close to a coin flip. The number that keeps circulating—47 percent—comes from that exact comparison: it is what you get when you count the winning pockets for a colour bet on an American wheel and divide by the total pockets. The gap between that figure and 50 percent is not a rounding error. It is the entire business model of the game, and it is worth understanding before you put a chip down.
Where the 47 percent figure comes from
Roulette wheels come in two common layouts. A European wheel carries 37 pockets: the numbers 1 through 36 plus a single zero. An American wheel carries 38: the same 36 numbers plus a single zero and a double zero. The zero pockets are not red or black. They are green, and they belong to neither colour bet.
On an American wheel, a bet on red wins on 18 pockets and loses on the remaining 20. The winning share is 18 divided by 38. As a percentage, that is just under half—close to the 47 percent figure people quote. On a European wheel, the same bet wins on 18 of 37 pockets, which lands closer to 48.6 percent. Neither is 50 percent, and neither ever will be, because the green pockets sit on the wheel for exactly this purpose.
What the percentage measures
The 47 percent figure is a probability, not a payout rate. It says how often a colour bet wins, assuming the wheel is fair and every pocket is equally likely. It does not say how much you get back. A winning colour bet pays even money, so the payout is 1 unit for every 1 unit staked. If the bet won 50 percent of the time, the player and the house would break even over the long run. It wins less than 50 percent of the time, and that shortfall is the house edge.
House edge and Return to Player (RTP) are complements: they sum to 100 percent for the same bet under the same rule set. If a bet returns 94.7 percent over a very large number of spins, the house edge on that bet is 5.3 percent. The numbers are two ways of describing the same relationship. The player keeps, on average, a little less than they stake; the house keeps the difference. This is a long-run statistical average, not a session guarantee, and it applies to the specific bet, not to the whole table.
Colour bets versus other bets on the same table
Different bets on a roulette table carry different probabilities and different payouts, but the house edge on American roulette stays the same for most of the standard inside and outside bets. A single-number bet pays 35 to 1, which sounds generous, but it wins on only 1 of 38 pockets. The payout is short of what a fair game would pay on those odds, and that shortfall matches the edge on the colour bets. The table below compares what changes between a colour bet and a single-number bet on the same American wheel.
| Bet type | Winning pockets (American wheel) | Payout | What the payout implies versus what the wheel gives |
|---|---|---|---|
| Red or black | 18 of 38 | Even money | Payout assumes a 50 percent chance; the wheel gives less because of the two green pockets |
| Single number | 1 of 38 | 35 to 1 | Payout assumes a 1-in-36 chance; the wheel gives 1-in-38, so the payout is short by two pockets |
The European wheel removes one green pocket, which changes the probability on every bet and lowers the house edge across the standard bets. The difference between the two wheel types is not cosmetic. On a colour bet, the European layout gives the player a winning share closer to half; on the American layout, the second zero pushes it down toward the 47 percent figure that prompted the question.
Why short sessions look nothing like the long-run average
A 47 percent winning probability does not mean you will win 47 out of every 100 spins you play. It means that over a very large number of spins, the share of winning colour bets will tend toward that figure. In a session of 20 or 50 spins, the actual share can swing far above or far below it. You may win eight of ten colour bets and leave ahead. You may lose nine of ten. Neither outcome contradicts the underlying probability.
This is variance: the distribution of outcomes over time. Roulette is a high-variance game in short bursts because each spin is independent and the payout on a colour bet is flat. A run of reds does not make black more likely on the next spin. The wheel has no memory, and the pockets do not owe anyone a correction. The long-run average is a property of the game's structure, not a promise about your evening.
What decisions per hour do to total exposure
The house edge is applied to the amount you wager, not to the amount you brought. Every spin puts a fresh bet at risk, and the edge is taken on each one. The more decisions per hour, the more total coin-in, and the larger the expected loss for a given edge. A player making 60 colour bets an hour exposes 60 times their average bet to the same percentage. A player making 20 bets an hour exposes one-third of that.
This is why table speed matters as much as bet size. A crowded table moves slower. An online auto-spin setting moves faster. Suppose, as an example, a player wagers 5 units per spin on a colour bet. At 60 spins an hour, that is 300 units of total wagers. At 20 spins an hour, it is 100 units. The edge applies to the total wagered, so the slower session carries one-third of the expected loss of the faster one, all else equal. Nothing about the game changed; only the pace changed.
Why progression systems redistribute rather than remove the edge
The 47 percent figure invites a natural response: if the bet loses more often than it wins, why not raise the stake after a loss so that a win recovers everything? This is the logic behind doubling systems and their relatives. The problem is structural. The edge is attached to every individual wager. Raising the size of the next wager raises the amount exposed to that same edge. The system changes the shape of the outcomes—many small wins punctuated by rare, large losses—but it does not change the expected value.
As an example, suppose a player starts with a 1-unit colour bet and doubles after each loss, planning to stop after a win. The sequence of stakes is 1, 2, 4, 8, and so on. Most sequences end with a 1-unit gain. The rare sequence that runs long enough hits the table maximum or the player's bankroll limit, and the loss on that sequence is the sum of all the doubled stakes. The arithmetic works out so that the expected loss remains the product of total wagers and the house edge. The progression redistributes when losses land; it does not remove the built-in shortfall from the green pockets.
What the player can choose
What varies under the player's control is not the edge but the conditions around it. The wheel type is the clearest choice: a single-zero layout gives a better winning share on colour bets than a double-zero layout, because there is one fewer pocket that pays neither colour. The bet selection matters too, though on roulette the standard bets share a similar edge on the same wheel. The pace of play, the size of the stake, and the length of the session all multiply or reduce total exposure. A stopping rule—leaving after a set number of spins or a set loss—does not improve the odds, but it bounds the session.
No betting pattern changes the built-in edge of a negative-expectation game. The 47 percent figure is a description of the wheel's structure. The only decisions that change the expected loss are decisions about which wheel, which bet, how much, and for how long. Everything else is variance, and variance can be kind or cruel without ever changing the underlying arithmetic.
FAQ
Is 47 percent the same as the house edge?
No. The 47 percent figure is the probability of winning a colour bet on an American wheel. The house edge is the shortfall between that probability and what the payout implies. They are related, but one is a chance and the other is a rate of loss per unit wagered.
Does the 47 percent figure apply to every roulette bet?
No. It applies specifically to colour bets on a double-zero wheel. A single-number bet wins far less often, while a bet covering more pockets wins more often. The payout on each bet is set against its own probability, and the edge is the gap between the two.
Can a player overcome the 47 percent by watching for patterns?
No. Each spin is independent, and past results do not change the probability of the next spin. The winning share tends toward the long-run average over many spins, but no pattern in past spins predicts the next one, and no staking system removes the edge attached to each wager.
Gambling involves risk. Please only gamble with funds that you can comfortably afford to lose. While we do our best to provide good advice, we cannot be held responsible for any loss that may occur as a result of gambling. We do our best to make sure all the information that we provide on this site is correct. However, from time to time, mistakes will be made and we will not be held liable. Please check any stats or information if you are unsure how accurate they are. No comment on any of the brands we cover should be taken as advice that you should play at that brand. We accept no responsibility for any actions taken as a result of the information on this site. If you or someone you know has a gambling problem, call the National Gambling Helpline at the helpline printed in the footer below.
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.