The house edge is the percentage of each bet a casino expects to keep over time. It’s built into the math of every game, not added on top afterward. This page explains what that number means, how it’s calculated, why it changes across games and rule variations, and why your results in any single session can look nothing like the long-run figure. By the end, you’ll have a solid basis for comparing games and understanding what any advertised return figure actually tells you.
That single number determines how any advertised return figure should be read: as a long-run statistical expectation, not a prediction about any individual session. This article covers what the house edge is, how it’s produced, how it varies across game types and rule setups, and why short-term results routinely look nothing like it.
The Core Definition of the House Edge
The house edge is how a casino retains a portion of every wager placed over a large number of bets. It’s a property of the game’s structure, not of any individual session or outcome. This section defines the metric precisely, explains how it’s calculated, and separates it from two figures it’s often confused with.
The house edge is the ratio of the average loss to the initial bet, expressed as a percentage. “Average loss” means the expected loss per bet calculated across all possible outcomes, weighted by their probabilities. “Initial bet” is the wager placed before any game action that might change the total amount staked. In games where a player can raise their bet mid-hand, that extra money is excluded from the denominator.
Because the figure comes from probability-weighted outcomes across all possible results, it describes what happens on average across a very large number of trials, not what will happen on any single hand or spin. A published house edge of 0.5%, for example, means that for every €100 wagered at the outset, the expected long-run loss is €0.50. That’s a statement about the statistical average, not a guarantee about any one session.
The house edge and the Return to Player percentage (RTP) are two ways of expressing the same underlying math, and they always add up to 100%. A game with a 5% house edge has an RTP of 95%, meaning the casino keeps 5 cents of every euro wagered on average and returns the other 95 cents to players over time. An operator may display one figure without the other, but knowing either one gives you the second through simple subtraction. A high RTP and a low house edge aren’t separate properties; they’re two ways of saying the same thing.
Two metrics are regularly mixed up with the house edge in published casino data: the element of risk and the hold. Each uses a different denominator, which produces meaningfully different figures for the same game. In games where the final wager can exceed the initial bet, such as those with optional raises, the house edge and the element of risk can diverge significantly, so the choice of metric matters when comparing games.
| Metric | What It Measures | Denominator | Common Point of Confusion |
|---|---|---|---|
| House Edge | Expected player loss as a percentage of the opening wager | Initial bet | Often mistaken for the share of all money wagered that the casino retains, which is a different calculation |
| Element of Risk | Expected player loss as a percentage of all money wagered, including any raises | Total money wagered | Produces a lower figure than the house edge in games with optional raises, leading readers to underestimate the house edge or overestimate the element of risk |
| Hold | The percentage of chips purchased that the casino retains across a session or period | Chips purchased | Commonly confused with the house edge even among casino staff, though it reflects player behaviour and session length rather than the game’s fixed mathematical structure |
How the House Edge Is Produced Mechanically
The house edge isn’t a fee charged on top of a game’s natural outcome. It’s built into the game’s payout structure. Every casino game is designed so that the amount it pays out for a winning result is slightly less than what the true probability of that result would justify. That gap, applied consistently across every possible outcome, is where the house edge comes from.
Every outcome in a casino game has a mathematically precise probability: the true odds. Those odds describe exactly how often that outcome will occur over a large number of trials.
The game’s payout for that outcome is set slightly below what the true odds would require for a break-even result. On a fair coin flip, true odds are 1-in-2, so a break-even payout would be exactly 1-to-1. If a game paid out at 0.95-to-1 on an even-money bet with a true 50% chance of winning, the player would receive less than the fair value of the outcome every time they won.
That shortfall doesn’t disappear across losing outcomes; it accumulates. When the payout gaps across all possible outcomes are weighted by their probabilities and added up, the result is the house edge expressed as a percentage of the initial bet.
The house edge is encoded in the payout table itself. Reading a game’s payout schedule against its true probabilities reveals the edge directly, without any additional calculation.
The same game can carry very different house edges depending on its specific rule set. Adding a second zero pocket to a roulette wheel, increasing the number of decks in a card game, or removing a rule that returns part of a losing bet each changes the relationship between true probabilities and payout odds, and therefore changes the edge. The rule set of a specific game, not its category name, determines the actual long-run expectation a player faces.
These categories of rule variables can meaningfully shift a game’s house edge:
- Wheel configuration: the number of zero pockets on a roulette-style wheel changes the true probability of every non-zero outcome without changing the payout odds, directly altering the edge.
- Deck count: the number of decks in play in card-based games affects the composition of the remaining deck at any point, which changes the true probabilities underlying each decision.
- Optional rule sets: rules that return part of a losing bet under specific outcomes, such as when the ball lands on zero in certain roulette variants, reduce the effective loss on those bets and lower the overall edge.
- Player strategy: whether the player uses the mathematically optimal decision at each choice point determines how closely the actual edge tracks the theoretical minimum for that game.
- Bet type within a game: which specific wager the player places among a game’s available bets determines the edge, since different bets within the same game carry different payout structures relative to their true probabilities.
House Edge by Game Category
The house edge varies widely across casino game categories, from fractions of a percent to well above twenty percent. Both the game type and the specific bet placed within that game shape the long-run expectation, meaning two players at the same table can face very different edges depending on which wager they choose. The game name alone doesn’t tell you the expected cost; the bet selection matters just as much.
Table games carry the lowest house edges in the casino, and those edges are more sensitive to player decisions than in any other category. Within a single game, choosing a different bet type can shift the edge by more than tenfold, so any published figure for a table game applies to a specific bet, not the game as a whole.
| Game | Bet Type | House Edge |
|---|---|---|
| Blackjack | Liberal Vegas rules, optimal strategy | 0.28% |
| Blackjack | Typical rule range | 0.5%–1% |
| Baccarat | Banker | 1.06% |
| Baccarat | Player | 1.24% |
| Baccarat | Tie | 14.36% |
| Craps | Pass / Come | 1.41% |
| Craps | Odds bet | 0.00% |
| Video Poker | Full-pay Jacks or Better | 0.46% |
Roulette is the clearest example of how a single structural change moves the house edge substantially across variants of the same game. The number of zero pockets on the wheel and whether the game returns part of a losing stake when the ball lands on zero are the two variables that account for the entire spread across variants. Checking the variant label on a roulette game is the first step in figuring out the long-run expectation, not just reading the game name.
| Variant | Structural Feature | House Edge |
|---|---|---|
| French Roulette | Single-zero wheel with La Partage rule (half stake returned on zero for even-money bets) | 1.35% on even-money bets |
| European Roulette | Single-zero wheel | 2.70% |
| American Roulette | Single-zero and double-zero wheel | 5.26% |
| Triple-Zero Roulette | Single-zero, double-zero, and triple-zero wheel | 7.89% |
Slots and lottery-style games cover the widest edge range in the casino and are generally the highest-edge segment. The exact house edge for an individual slot machine is typically not published on the game itself, so players have to rely on category-level ranges rather than a specific figure. The prize structure displayed on a slot or keno game doesn’t indicate an edge comparable to table games, and the figures below reflect that gap.
| Game Category | House Edge Range |
|---|---|
| Slot Machines | 2%–15% |
| Keno | 25%–29% |
Why Short-Term Results Diverge from the Long-Run Expectation
The house edge is a long-run statistical expectation, not a prediction for any single session, hour, or hand. Any individual period of play can produce an outcome that looks nothing like the published figure, and that’s a normal feature of probability, not an error in the stated edge. This section explains why that gap exists and describes how digital games preserve the same underlying mathematical structure.
The house edge describes the average outcome across a very large number of trials. It’s calculated from the full probability distribution of a game’s outcomes, weighted by their respective payouts, and it only converges toward its stated value as the number of trials grows large.
Any finite session produces a result that falls somewhere within a distribution around that average. That distribution isn’t a flaw in the game’s math; it’s a direct consequence of how probability works when outcomes are drawn from a fixed set of possibilities.
The fewer the trials in a session, the wider the range of results that can plausibly occur. A player who completes twenty hands of blackjack is sampling from a much broader spread of possible outcomes than a player who completes two thousand hands under identical rules.
A winning session is a valid statistical outcome that doesn’t contradict the published house edge. The edge doesn’t guarantee a loss on any given occasion; it describes the direction and size of the expected result across a large enough sample that variance becomes negligible relative to the mean.
A losing session carries the same interpretation in reverse. A result that exceeds the expected loss for a session is equally consistent with the stated edge, because the edge makes no claim about what will happen within any bounded window of play.
Online casino games use random number generators to reproduce the same mathematical structure that produces the house edge in physical games. The RNG determines each outcome by drawing from the same probability space that defines the game’s edge, so the stated return figure applies to digital play in the same way it applies to a physical table or machine.
In regulated markets, RNG-based games are subject to third-party fairness audits. These audits verify that a game’s recorded outcomes match its stated probability model, confirming that the distribution of results lines up with what the published house edge or return figure implies. They don’t assess whether any individual player’s results were favourable or unfavourable.
When independently published audit information exists for a digital game, it means the stated return figure has been externally verified against actual outcome data. When that information isn’t available, the stated figure hasn’t been confirmed by an independent party. That’s a difference in verifiability, not necessarily a statement about the game’s actual performance.
Reading the House Edge on Any Game You Encounter
The house edge is most useful as a framework for interpreting what any published return figure actually means. Once you understand that the edge is encoded in the payout structure and shifts with rule configurations rather than just game names, you can tell the difference between a long-run mathematical expectation and a session result, and spot when two figures describing the same game are actually measuring different things.
