RTP and house edge explained: what the numbers actually mean for your money

A complete guide to Return to Player percentages, house edge, and why they're the same number described two different ways

Every casino game has a mathematically defined long-run return - a number that, over enough spins, hands, or bets, describes exactly how much of your money the game is designed to return to you, and by extension, exactly how much the operator retains. This guide covers RTP and house edge - the two ways this same underlying number gets presented - how they're calculated, what "long-run" genuinely means for your individual session, and why understanding this properly changes how you should evaluate any casino game.

RTP and house edge are the same number, described two ways

RTP (Return to Player) is expressed as the percentage of all money wagered on a game that is returned to players over time. A slot with 96% RTP returns, on average, £96 for every £100 wagered across its full theoretical lifetime of play.

House edge is simply the inverse - the percentage the operator retains. A game with 96% RTP has a 4% house edge. These two figures always sum to 100%, because between them they account for every pound wagered: what comes back to players (RTP) and what the house structurally retains (house edge).

Why both terms exist: RTP is the player-facing framing - the number the UK Gambling Commission increasingly requires operators to display, and the number that makes intuitive sense when comparing games ("this one returns more than that one"). House edge is the mathematics-facing framing, more commonly used in table games and by analytically-minded players, because it directly describes the operator's structural profit margin on every wager.

How RTP is actually calculated

RTP is not measured from real player sessions - it's a mathematical property of the game's design, calculated by the game's developer based on the paytable, the probability of each outcome, and the game's specific rules, then independently verified by testing laboratories (organisations like GLI - Gaming Laboratories International - or eCOGRA) before the game receives regulatory certification.

The theoretical nature of the figure: A 96% RTP slot doesn't return exactly 96p per £1 in any single session, or even any single day of play across all players. It's a theoretical figure calculated across the game's full mathematical cycle - which for a modern video slot can run into the tens or hundreds of millions of spins before the actual returned percentage converges meaningfully close to the stated theoretical RTP.

Simulation-verified, not empirically measured in real-time: Testing labs verify RTP by running millions of simulated spins against the game's underlying random number generator and paytable, confirming the actual simulated return matches the developer's stated theoretical figure within an acceptable margin - this is what regulatory certification actually confirms, not real-world tracking of actual player outcomes.

What "long-run" actually means for a real session

This is the single most misunderstood aspect of RTP, and the one that matters most practically.

A 96% RTP slot does not mean you'll get back 96% of your stake in any given session. Over a 200-spin session, your actual result could reasonably range from losing your entire stake to winning several multiples of it - the stated RTP describes the average outcome across an enormous number of spins, not a guarantee for any specific, finite session you'll actually play.

Volatility determines how that average is distributed. Two slots can both carry 96% RTP while feeling completely different to play - a low-volatility slot delivers that 96% through frequent small wins, rarely producing a long losing streak but also rarely producing a large single win. A high-volatility slot delivers the same 96% average through long stretches without a win, punctuated by the occasional large multiplier hit. Both games are mathematically "fair" at the same RTP; they simply distribute that fairness completely differently across a real session.

The practical takeaway: RTP tells you what to expect on average over an enormous sample size. It does not tell you what a specific 30-minute session will look like - that's determined by volatility, and by simple variance, which is the mathematical term for the genuine randomness around any average outcome in a finite sample.

The precise numbers behind "long-run": how many spins convergence actually takes

Earlier sections of this guide explain that RTP describes an average across an enormous number of spins, not a guarantee for any specific session. This section puts precise, sourced numbers on exactly how many spins that convergence genuinely requires - because "an enormous number" is vague, and the actual figures are more concrete and more useful than that vagueness suggests.

The accuracy-by-sample-size relationship, with real figures. For even a rough directional estimate of a game's actual RTP, independent analysis suggests you'd need at least 10,000 spins of tracked data. To narrow that estimate to within roughly ±1% of the true theoretical figure, you'd need approximately 40,000 spins. To tighten it further to within ±0.5%, you're looking at roughly 160,000 spins. Professional testing labs, by contrast, verify RTP using millions of simulated spins specifically to eliminate this uncertainty entirely for certification purposes.

A concrete illustration of just how wide the swings are at low spin counts. After just 10 spins, your personally tracked return could reasonably show anywhere from 50% to 200% of your stake - a swing so wide it says almost nothing about the game's true underlying RTP. After 1,000,000 spins, your tracked figure would typically sit within about 0.1% of the game's published theoretical RTP. The core statistical principle: it's not that a slot "adjusts" or "self-corrects" toward its stated RTP over time - nothing about the game's mechanics changes based on your play history. It's simply that random variance, which dominates results heavily at low sample sizes, mathematically shrinks in relative significance as the sample size grows, a property of statistics generally rather than anything specific to how a particular slot behaves.

A genuinely interesting edge case worth knowing: your own session RTP can legitimately exceed 100%. No licensed slot lists a theoretical RTP above 100% - that would mean the operator loses money by design, which no commercially viable game does. But your own personal, actual RTP for a specific short session absolutely can exceed 100%: if you play a handful of rounds and hit a meaningful win before stopping, the ratio of what you won to what you wagered in that specific session can genuinely be well over 100%, even though the game's own long-run theoretical figure remains fixed below 100%. This isn't a contradiction - it's simply the difference between a single short sample (your session) and the enormous long-run sample the theoretical RTP describes.

Why this matters practically for interpreting your own play. If you track your own results across a typical recreational session - a few hundred to a few thousand spins - you should expect your actual, personally-observed return to differ from the game's stated RTP by a potentially significant margin, in either direction, purely as a function of normal variance at that sample size. A session running notably better or worse than the stated RTP isn't evidence the game is somehow miscalibrated or "not working as advertised" - it's the statistically expected, normal behaviour of a properly random game observed over a sample size far too small for convergence to have meaningfully occurred yet.

RTP across different game types

Slots: RTP varies enormously by title - from the low-90s for some higher-margin games up to 99%+ for a small number of specifically high-RTP titles (Mega Joker by NetEnt, for example, at 99.0% RTP, is among the highest-returning slots available at any licensed operator).

Blackjack: Using correct basic strategy, standard blackjack rules produce a house edge of approximately 0.5%, equivalent to 99.5% RTP - among the best mathematical odds available in any casino game, though this figure assumes perfect strategy application on every single hand.

European Roulette: A fixed 2.70% house edge (97.30% RTP) on every bet type except the specific La Partage or En Prison rules some operators apply to even-money bets, which reduce the effective house edge on those specific bets to approximately 1.35%.

American Roulette: A meaningfully worse 5.26% house edge (94.74% RTP), due to the additional double-zero pocket on the wheel - this is why European roulette, where available, is always the mathematically better choice over the American variant.

Live dealer games: Typically mirror their RNG equivalents' house edge closely, since the underlying game rules and payouts are identical - the "live" element changes the presentation and pacing, not the fundamental mathematics.

The RTP detail almost no guide mentions: the same slot can have multiple RTP versions

This is genuinely important information that most RTP explainers skip entirely, and it directly affects how much you should trust a headline RTP figure you find online.

Many popular slot titles are built with several selectable RTP configurations. Rather than a single fixed mathematical model, some game developers build their titles with multiple possible RTP settings - commonly ranging across several percentage points - and the operator deploying the game chooses which specific configuration to activate on their platform. This means the same slot title, by the same developer, can genuinely carry a different RTP at one licensed operator compared to another.

Why regulated markets handle this differently than unregulated ones. In tightly regulated markets, operators are required to declare which specific RTP configuration they've selected for each game, and that declared figure must remain fixed and displayed accurately - this is the standard both UK Gambling Commission and licensed US state regulators expect, and it's part of why sticking to UKGC-licensed and state-licensed operators (the only kind covered on this site) matters beyond the basic question of legality. At unregulated or offshore operators, a player may have no reliable way of knowing which specific RTP configuration is actually running, since there's no regulatory requirement forcing accurate disclosure.

The practical takeaway: When you look up a slot's RTP online - through a review site, a forum, or even the game's own information screen - that figure is accurate only for the specific configuration deployed at the specific operator you're checking. Always verify RTP through the actual in-game information screen at the specific operator you're playing at, rather than relying purely on a general figure quoted elsewhere, since the same title could legitimately be running a different configuration there.

Why buying a bonus feature can carry a different RTP than the base game

Many modern slots offer a "bonus buy" or "feature buy" option - paying a multiple of your normal stake to skip directly into the bonus round rather than triggering it through regular play. This creates a genuinely important RTP nuance worth understanding before using the feature regularly.

The mechanism: The bonus buy feature and the base game are, mathematically, separate betting propositions with their own independently calculated RTP - even though they exist within the same slot title. It's entirely possible, and common, for a slot's bonus buy RTP to differ from its base game RTP, sometimes meaningfully. Similarly, activating an optional "ante bet" (a side wager, common on certain video slot formats, that increases your odds of triggering bonus features in exchange for a higher base stake) creates its own distinct RTP calculation separate from standard base-game play.

Why this matters practically: If you specifically favour bonus-buy play - skipping base game spins to go straight for bonus rounds - the RTP figure you should be checking is the bonus buy RTP specifically, not the game's headline base-game RTP, since these can be different figures within the same title. Reputable game information screens typically disclose both figures separately where a bonus buy or ante bet option exists; if a specific game's information screen doesn't clearly separate these, that's worth treating as a transparency gap rather than assuming the base RTP applies universally across every way the game can be played.

Why RTP doesn't capture everything: the bonus contribution trap

This is a genuinely important practical extension of the RTP concept that most explainer guides skip: RTP describes the game itself, but bonus wagering requirements can effectively override the practical value of a high-RTP game if the bonus terms weight that specific game type down.

The mechanism: Casino bonuses commonly assign different "contribution rates" to different game categories when calculating progress toward a wagering requirement. Standard slots often contribute 100% (every £1 wagered counts fully toward clearing the bonus), while table games frequently contribute at a reduced rate - sometimes as low as 10-20% - and some high-RTP slots specifically are weighted down even further at certain operators, precisely because their high RTP makes them a less profitable vehicle for the operator to have players clear bonus wagering through.

Why this matters: A slot's 99% RTP is a genuine, real mathematical property of that game. But if you're using that slot specifically to clear a bonus wagering requirement, and the bonus terms only credit 30% of your wagers on that title toward the requirement, your effective wagering burden for that specific game is roughly three times higher than the headline wagering figure suggests - a detail entirely separate from the game's own RTP, but one that meaningfully affects the real value of combining a high-RTP game with an active bonus.

How to use RTP information practically

Compare RTP within the same game category. Comparing a slot's RTP against a table game's RTP isn't a like-for-like comparison, since the volatility profiles and bet-sizing conventions differ enormously between categories. Compare slots against slots, and table games against table games, for a meaningful sense of relative value.

Higher RTP doesn't guarantee a better session - it improves your long-run expected outcome. If you're playing purely for entertainment across a single evening, a lower-RTP but higher-volatility game might genuinely produce a more exciting, unpredictable session even though its long-run mathematics are objectively worse than a higher-RTP alternative. Neither choice is "wrong" - they serve different purposes.

Check whether an operator displays per-game RTP at all. Some UK operators publish RTP directly in the game lobby, letting you filter or sort by it; others require you to check individual game information pages, or don't display it prominently at all. The UK Gambling Commission has been pushing the industry toward greater transparency on this specific point, and operators that already lead on this are generally a positive signal about the operator's broader transparency practices.

Variance and the volatility index: the actual numbers behind "high" and "low" volatility

Earlier sections of this guide, and several category pages elsewhere on this site, describe slot volatility in the general terms of "low," "medium," and "high." This section provides the actual mathematics behind those labels - because regulators require every slot to carry a precise, calculable figure, not just a marketing label.

Variance and standard deviation, defined precisely: Variance measures how dispersed a game's payouts are from the statistical average - a low-variance game keeps outcomes clustered close to the mean, while a high-variance game produces outcomes that swing widely above and below it. Standard deviation is simply the square root of variance, which converts the (often very large and unwieldy) variance figure into a more intuitive number expressed directly in stake multiples.

A concrete illustration: Two hypothetical slots both average a 50x return on total wagers over the long run - identical RTP. The low-volatility slot has a standard deviation of roughly 10x, meaning outcomes typically swing between about 40x and 60x. The high-volatility slot has a standard deviation of roughly 100x, meaning outcomes can swing anywhere from 0x to 250x or beyond. Both games are mathematically "fair" at the same average - but the actual experience of playing them, session to session, is completely different.

The Volatility Index (VI): When slot developers submit games for regulatory approval, they must specify a formal Volatility Index - calculated as the standard deviation multiplied by a statistical confidence-interval weighting (commonly 1.645 for 90% confidence, or 1.96 for 95% confidence). This produces a standardised number, typically falling somewhere in the 0-20 range for most slots.

For comparison, beyond slots: Most table games carry a Volatility Index below 5 - table games are inherently lower-variance than slots because their payout structures are simpler and their outcomes cluster more tightly around the mean. At the extreme opposite end, a lottery draw like a major jackpot game can carry a Volatility Index in the tens of thousands, reflecting the enormous gap between the near-certainty of losing your stake and the vanishingly rare but enormous jackpot payout.

Why this number matters more than most players realise: Two slots with genuinely identical RTP can produce completely different bankroll experiences depending on their Volatility Index - a low-VI slot will feel like it's paying consistently even during a losing session, while a high-VI slot can produce dozens or hundreds of consecutive losing spins even when playing entirely correctly, purely as a function of its design. Checking a game's stated volatility (even in the simplified low/medium/high form most operators display, if not the precise numerical index) before choosing what to play with a fixed session bankroll is genuinely useful practical information, not just a preference detail.

Video poker: the one casino game where strategy genuinely changes your RTP

Every other game covered in this guide has a fixed house edge that basic strategy can approach but never improve beyond the game's built-in mathematics. Video poker is structurally different - it's the clearest example in the entire casino category of a game where your actual decisions directly determine your realised RTP, not just your variance.

Why video poker works differently: Unlike a slot (where outcomes are entirely determined by the random number generator with no player input beyond pressing spin) or blackjack (where basic strategy converges everyone toward the same near-optimal outcome), video poker deals you a hand and then asks you to choose which cards to hold and which to discard - and that choice has a directly calculable, precise effect on your expected return for that specific hand.

The numbers, using Jacks or Better as the reference game: A full-pay 9/6 Jacks or Better machine (paying 9x your bet for a full house, 6x for a flush) returns 99.54% RTP with mathematically perfect strategy on every single hand - among the best odds available anywhere in a casino. Play the same machine with a simplified, "good but not perfect" strategy, and the RTP drops to approximately 99.46% - a small-looking gap that compounds meaningfully over thousands of hands. Play with genuinely casual, error-prone decision-making, and realised RTP commonly falls to 96-98%, even though the machine's theoretical maximum never changed.

Why the pay table matters as much as your strategy. Not every Jacks or Better machine is a 9/6 table. An 8/5 pay table (paying 8x for a full house, 5x for a flush) drops the maximum achievable RTP to approximately 97.3%; further reduced pay tables (7/5, 6/5, and lower) can push the achievable maximum down to 96% or below - meaningfully worse than the full-pay version, regardless of how well you play it. Checking the specific pay table before playing any video poker machine is arguably more important than memorising the strategy chart itself, since no strategy can overcome a genuinely worse pay table.

Why max-coin betting is a structural requirement, not just a suggestion. On a full-pay Jacks or Better machine, the royal flush pays 800-to-1 at maximum coin bet, but only 250-to-1 at anything less than maximum - a difference so large that it alone accounts for roughly 2 percentage points of the game's total 99.54% RTP. Betting fewer than the maximum coins on a video poker machine quietly removes a meaningful chunk of the game's headline return, regardless of how well you play every individual hand.

The honest caveat about "beatable" games. Some very specific, rare pay table variants (occasionally advertised as 9/7 or 10/6 tables) can theoretically return slightly above 100% RTP with perfect strategy - genuinely player-favourable in pure mathematical terms. These are uncommon, require flawless strategy execution on every hand to realise the advantage, and are far more relevant to dedicated advantage players than casual bettors. For the overwhelming majority of video poker play, the realistic goal is finding the best available pay table and applying strategy consistently - not hunting for a mathematically beatable machine.

RTP reference: real figures across common games

The table below collects verified RTP figures for widely available games across categories, giving you a genuine reference point rather than abstract percentage ranges.

Figures reflect optimal or theoretical play as certified by independent testing labs; actual RTP achieved by any individual player depends on strategy accuracy (where applicable) and, for all games, remains subject to genuine short-term variance regardless of the theoretical figure.

Frequently asked questions

How many spins does it actually take for RTP to converge?

For a rough estimate, at least 10,000 spins; for accuracy within ±1% of the theoretical figure, approximately 40,000 spins; for ±0.5% accuracy, roughly 160,000 spins. Professional testing labs use millions of simulated spins for certification purposes. At just 10 spins, your tracked return could reasonably range from 50% to 200% - variance dominates heavily at low sample sizes.

Does a 96% RTP slot mean i'll get 96% of my money back?

Not in any single session - RTP is a theoretical figure calculated across the game's entire mathematical cycle (potentially tens of millions of spins), not a guarantee for any specific finite session you'll actually play. Your individual results will vary significantly from the stated average due to variance.

What's a "good" RTP for a slot?

Slots at 96% RTP or above are generally considered favourable within the industry; the highest-RTP titles available (Mega Joker, Blood Suckers, and similar) reach 98-99%. Slots below 94% RTP are on the less favourable end of what's typically available at licensed UK and US operators.

Is house edge the same as RTP?

Yes - they're the same underlying figure expressed two different ways, and always sum to 100%. A game with 4% house edge has 96% RTP; a game with 96% RTP has a 4% house edge.

Why do bonus terms sometimes make a high-RTP slot's bonus contribution lower?

Because a slot's high RTP means less of each wager is retained by the operator - using that game to clear bonus wagering is less profitable for the operator than a lower-RTP game, so some operators specifically weight high-RTP titles down in their bonus contribution tables to offset this.

Is European roulette really better than American roulette?

Yes, mathematically and unambiguously. European roulette's single-zero wheel produces a 2.70% house edge; American roulette's additional double-zero pocket nearly doubles that to 5.26%. Where both are available, European roulette is always the better choice.

This guide is for educational purposes. All casino games carry a built-in house edge regardless of RTP percentage, and no strategy guarantees winning outcomes. Gambling problem? In the UK: BeGambleAware.org | 0808 8020 133. In the US: 1-800-GAMBLER. Information correct as of September 2026.