The payout never matches the true odds
Every roulette bet is settled at odds that are slightly worse than the actual chance of the bet winning. That gap is the entire reason the game loses money over time. It is not a hidden trick, not a rigged wheel, and not something that changes when you switch from red to black or from a single number to a column. The gap is printed into the payout table itself.
Take a straight-up bet on a single number. On a European wheel there are 37 pockets, so the true odds against your number coming up are 36 to 1. The table pays 35 to 1. That one-unit shortfall on every winning spin is the house edge. The same pattern repeats across the layout. A split bet covers two numbers out of 37, so the true odds are 17.5 to 1, but the table pays 17 to 1. A street bet covers three numbers, true odds just under 11.67 to 1, and the table pays 11 to 1. Every payout is rounded down from the true odds, and the rounding always favours the operator.
On an American wheel the gap widens because the wheel carries 38 pockets: the numbers 1 through 36 plus a zero and a double zero. The true odds on a single number become 37 to 1 while the payout stays at 35 to 1. The extra pocket does not change what the table pays; it only makes every bet less likely to win. That is why the American layout carries a larger built-in advantage than the European one, even though the payouts look identical at first glance.
Expected value is the slow leak
Expected value describes what a bet is worth on average before the ball lands. It is not what any individual spin returns. A single spin returns everything or nothing, or some multiple of the stake. Expected value is the weighted average of all those possible outcomes, and on every roulette bet it is negative.
Here is the arithmetic in a labelled example. Suppose you bet 100 units on red on a European wheel. There are 18 red pockets and 19 pockets that are not red, so the probability of winning is 18 out of 37 and the probability of losing is 19 out of 37. A win pays even money, so you gain 100 units; a loss costs 100 units. The expected value of one spin is (18/37 times 100) plus (19/37 times minus 100). That works out to a small negative number. The same calculation on an American wheel uses 18 winning pockets out of 38, and the expected loss per unit wagered is larger.
That negative expected value is the house edge expressed in money terms. It is not a fee that appears on a receipt. It is the average amount that stays with the operator across a very large number of spins. On any given spin you might win, and on any given evening you might walk away ahead. But the arithmetic of the bet itself is negative before the ball is released, and no betting decision changes that sign.
RTP and house edge are two views of one rule set
Return to Player, or RTP, is the same relationship viewed from the player's side. Where the house edge is the operator's average share of each wager, RTP is the share that comes back to players over the long run. For the same bet under the same rule set, the two figures sum to 100%.
This pairing is specific to a single game and a single set of rules. It is not a property of the casino as a whole, and it cannot be averaged across different games to produce a meaningful number. A European roulette bet and an American roulette bet have different edges and therefore different RTPs, even though they are both called roulette. A slot machine's published RTP has nothing to do with the edge on a baccarat table. The complement relationship only holds when you are comparing the same bet under the same conditions.
What this means in practice is that the RTP figure on a game information panel is a long-run statistical average, not a promise about your session. A game with a high RTP still has a house edge, and that edge still applies to every spin. The RTP simply tells you how much of the total wagered amount the game is designed to return across an enormous number of rounds. It says nothing about which players receive that return or when.
Short sessions look nothing like the long run
The reason roulette can feel winnable is that the house edge is small relative to the variance of a single spin. Variance describes how much individual outcomes swing around the average. In roulette, the swings are large. A single number pays 35 to 1, which means one lucky hit can put a session deeply into profit even though the expected value of that bet was negative from the start.
Over a small number of spins, almost anything can happen. Ten spins on red can produce eight reds. Twenty spins can produce a run that doubles a bankroll. The edge is still there, but it is drowned out by the noise of short-run randomness. This is why losing players sometimes believe a system is working and why winning players sometimes believe they have found an edge. Neither belief survives contact with enough spins.
The long run is where the edge emerges from the noise. As the number of spins grows, the actual return per spin drifts toward the expected value. The drift is not smooth, and it does not happen on a schedule. But the more decisions you make, the more the accumulated result reflects the built-in arithmetic rather than the luck of any particular sequence. This is the core of why the game loses money over time: the edge is small enough to ignore in an hour and large enough to dominate over a year of regular play.
| What changes with more spins | What stays fixed |
|---|---|
| The spread of possible session outcomes narrows relative to the amount wagered | The house edge on each bet |
| The chance of being ahead after the session falls | The payout odds printed on the table |
| Total expected loss grows in proportion to total wagered | The true odds of each bet winning |
| Short-term luck fades as a share of the result | The negative expected value of every spin |
Decisions per hour multiply total exposure
Roulette is a fast game. A live table can produce dozens of spins an hour, and an online game can run faster still. Each spin is a separate bet with its own negative expected value, so the expected loss of a session is not the edge on one spin. It is the edge per spin multiplied by the amount wagered per spin multiplied by the number of spins.
This multiplication is why two players at the same table can have very different long-run experiences while playing the same game. A player who bets 10 units per spin for 50 spins has put 500 units through the wheel. A player who bets 50 units per spin for 200 spins has put 10,000 units through it. The second player's expected loss is twenty times larger, not because the game treats them differently but because the edge applies to every unit wagered.
Game speed matters for the same reason. A slower table with fewer spins per hour produces less total exposure than a fast online game with the same stake. The edge per spin is identical, but the number of times it is collected changes. This is the honest answer to why losing over time is so reliable in roulette: the game is designed to be played repeatedly, and the edge is collected on every repetition.
Progression systems redistribute outcomes, not edge
Betting systems that change the stake after wins or losses are popular because they seem to address the problem directly. The most famous is the doubling system, where a player doubles the stake after every loss so that the first win recovers all previous losses plus one unit. The arithmetic of the sequence works on paper. The problem is that it does not change the expected value of any individual spin.
What a progression system does is reshape the distribution of outcomes. Instead of many small independent results, the player experiences a long series of small wins punctuated by occasional large losses. The small wins are frequent because the system is designed to produce them. The large losses are rare but severe, because a losing streak forces the stake up quickly and eventually hits either the table maximum or the player's available funds.
The expected value of the whole progression is the sum of the expected values of its parts. Since every part is a roulette bet with a negative expected value, the sum is negative. The system does not remove the edge; it concentrates it into the tail of the distribution. A player using a progression is not betting against the wheel with better odds. They are betting that a long losing streak will not occur during their session. When it does, the loss is large enough to erase many prior wins. This is why systems can feel successful for long stretches and still lose money over time.
What the player controls
None of this means the player has no decisions. The decisions just do not include changing the sign of the expected value. What the player controls is the size of the edge they face, the speed at which they play, and the amount they put at risk.
The edge itself varies by wheel type and by bet type. A European wheel carries a smaller built-in advantage than an American wheel, because the extra double-zero pocket makes every bet on the American layout less likely to win. Within a given wheel, some bets carry the same edge as each other, but the bet type changes the variance. A straight-up number has the same edge as an even-money bet on a European wheel in most rule sets, but the straight-up bet produces far wilder swings. A player who wants a longer session with smaller swings is choosing a different variance, not a different edge.
Speed and stake are the two biggest levers on total expected loss. Fewer spins per hour and smaller stakes per spin reduce the total amount wagered, which reduces the total expected loss in direct proportion. Session length does the same. None of these choices makes the game positive-expectation, but they do change how much the edge costs over a given period. The player who understands this is not looking for a way to win. They are deciding how much they are willing to pay, on average, for the entertainment of watching the wheel spin.
FAQ
Is roulette a game of skill or pure chance?
Roulette is a game of pure chance in the sense that no decision made after the bet is placed affects the outcome. The ball lands where it lands. The only decisions are which bet to place, how much to stake, and when to stop. None of those decisions changes the probability of any pocket coming up.
Why do some players seem to win at roulette for years?
Variance produces winning streaks that can last a long time, especially for players who bet infrequently or at low stakes. A player who visits a few times a year and bets small amounts may never experience the long run at all. Their results are dominated by luck, and luck can run positive for a long time before the edge becomes visible.
Can watching the wheel for patterns reduce the house edge?
No. Each spin is independent of the ones before it, and the wheel has no memory. A number that has appeared five times in a row is exactly as likely on the next spin as any other number, assuming a fair wheel. Patterns in past results are real in the sense that they happened, but they carry no information about the next spin.
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.