Slot RTP and volatility: what the percentage does and doesn't tell you
On this page
Every slot sold to players in Great Britain has to tell you, before you play, how much it pays back. That figure, the return to player (RTP), is usually a number in the mid-90s: 94%, 96%, 97%. It's easy to read it as a promise about your next session. It isn't. It describes the game over an enormous number of spins, and it says nothing about how bumpy the ride is on the way.
That second part, the bumpiness, is volatility. Two slots with exactly the same RTP can feel completely different to play: one drips out small wins, the other goes quiet for hundreds of spins and then, very occasionally, pays out a lot. This guide explains what each number measures, what the rules require casinos to show you, and, with three hypothetical games, what a session of play actually looks like.
How to read this article. The three games below are hypothetical: we made up their pay tables to illustrate the maths. They aren't real slots and don't describe any real game. The rules are quoted from the Gambling Commission's own pages. Every figure can be reproduced with the script listed under Sources.
What RTP actually measures
RTP is the share of all the money staked on a game that the game is designed to pay back, averaged over a very long run of play. A slot with a 96% RTP is built so that, across many millions of spins by everyone playing it, it returns £96 for every £100 staked. The other £4 is the house edge: the game's built-in margin.
Three things follow from that definition:
- It's an average across all players and all spins, not a figure for you, today. Nothing in the rules or the maths says your own session will land anywhere near it.
- It sets your expected cost, not your result. If you stake £1 a spin for 200 spins on a 96% game, you stake £200 and the expected loss is 4% of that: £8. That's the average over a huge number of sessions like yours, not what will happen.
- A higher RTP is better for you on average, but only on average. At 97% the expected cost of the same 200 spins is £6; at 94%, £12. Our casino wagering requirements guide shows how the same percentage sets the cost of clearing a bonus.
What the rules say casinos must show you
The Gambling Commission's remote technical standards (the RTS) set what online casinos licensed in Great Britain must tell you about a game. RTS 3 is explicit that it's about your chances: its aim is "to enable customers to make informed decisions about whether to gamble based on their chances of winning". Before you commit to play, the standard requires "a description of the way the game works and the way in which winners are determined and prizes allocated", along with "house edge (or margin)" and "the return to player (RTP) percentage or the probability (likelihood) of winning events occurring". The Commission's guidance describes this as "the average theoretical return to player percentage".
Other standards protect how the game behaves:
- Outcomes must be random. RTS 7 requires that game results are "acceptably random" and that "adaptive behaviour (that is, a compensated game) is not permitted". A game can't tighten up after a big win or loosen after a losing run: it isn't "due" anything.
- No misleading near misses. RTS 7 also bans "substituting losing events with near-miss losing events".
- No celebrating a loss. RTS 14 says the game "must not celebrate a return which is less than or equal to the total stake gambled". A spin that pays back 80p of your £1 can't be dressed up with win sounds and graphics.
- Speed and stakes. RTS 14 requires "a minimum of 2.5 seconds" between the start of one game cycle and the next, and bans features such as "turbo, quick spin and slam stop". Separately, the Commission's stake limits cap an online slot game cycle (a spin) at £5 for players aged 25 and over, and £2 for those aged 18 to 24.
So the RTP is a real, required piece of information. The point of this guide is what it can and can't tell you.
Volatility: same RTP, very different games
RTP tells you how much a game pays back. Volatility describes how it pays it back: lots of small wins, or a few rare big ones. It's measured by how spread out a spin's result is, the standard deviation: the bigger the number, the wilder the swings.
Here are three hypothetical games. All three have exactly the same 96% RTP. We built them that way so that the only difference between them is how the 96% is spread out.
| Hypothetical game | RTP | Spins that return something | Spread of one £1 spin (standard deviation) | Biggest win |
|---|---|---|---|---|
| Game A (low volatility) | 96.0% | 28.8% | £2.25 | 20× stake |
| Game B (medium volatility) | 96.0% | 22.2% | £5.07 | 100× stake |
| Game C (high volatility) | 96.0% | 12.2% | £31.96 | 5,000× stake |
Game A returns something on almost three spins in ten, but never more than 20 times the stake. Game C returns something on fewer than one spin in eight, and most of its 96% is locked up in prizes so rare that most players will never see them. On paper, they cost exactly the same to play.
"Spins that return something" isn't the same as winning, either: many of those returns are smaller than the stake, which is exactly why RTS 14 stops them being celebrated as wins.
What a session looks like
We simulated 20,000 sessions of 200 spins at £1 a spin on each game: £200 staked per session. The expected loss is the same £8 on all three.
| 200 spins at £1 | Sessions that end ahead | Median result | Average result | 1 session in 10 ends at or below | 1 session in 10 ends at or above |
|---|---|---|---|---|---|
| Game A (low) | 36.6% | −£10 | −£8.20 | −£47 | +£34 |
| Game B (medium) | 38.2% | −£24 | −£8.10 | −£85 | +£88 |
| Game C (high) | 14.3% | −£104 | −£8.20 | −£142 | +£52 |
The average across the sessions sits close to the expected −£8 on every game. That's the RTP doing its job. But almost nobody gets the average:
- On the low-volatility game, results cluster: most sessions end between −£47 and +£34.
- On the high-volatility game, the typical session (the median) loses £104, more than half the stake, and only about one in seven ends ahead. The average still comes out near −£8 only because a small number of sessions hit very big prizes: the luckiest 1% of sessions finished £1,026 or more ahead.
That's the core of it: the same RTP can hide very different experiences. A high-volatility game spends most sessions taking money and pays it back to very few players in large amounts.
Over 1,000 spins (£1,000 staked, expected loss £40) the pattern holds. On Game C the median session loses £366, about one in five ends ahead, and the average across our 20,000 sessions came out at −£31.80, not the expected −£40. That gap isn't a flaw in the game or the maths: it's chance. A simulated average is itself a random result, and its likely spread is measured by its standard error. For Game C over 1,000 spins, that is about £7.15, so about 95% of simulations like ours would produce an average somewhere between roughly −£54 and −£26; −£31.80 is well inside that range. For Game A the same range is only about ±£1. The high-volatility game is so much harder to pin down because so much of its return comes from a 5,000× prize that lands on about one spin in 31,000: in 20 million simulated spins, a few more or fewer of those prizes moves the average by pounds.
Why one session tells you almost nothing about RTP
People sometimes judge a slot's RTP from their own play: "it can't really be 96%, I lost half my money". Whether or not that's true, a session can't show it. Here's a rough sense of scale: how many spins it would take before your average return is likely to be within one percentage point of the RTP.
These are statistical approximations, not thresholds. They assume every spin is independent and use the normal approximation for an average of many spins, with 95% confidence: after that many spins, about 95% of players' average returns would be within ±1 point of the RTP, and about 5% would still be further away. Nothing makes a game "reach" its RTP after a set number of spins; the average only tends to get closer as the number of spins grows. For a high-volatility game the approximation is rougher still, because its results depend so heavily on rare prizes.
| Hypothetical game | Approximate spins for a 95% chance of an average return within ±1 point of the RTP |
|---|---|
| Game A (low) | about 195,000 |
| Game B (medium) | about 990,000 |
| Game C (high) | about 39 million |
Calculation: spins ≈ (1.96 × the standard deviation of one spin ÷ 0.01)², using the exact standard deviations in the table above.
Even at the fastest pace the rules allow, one spin every 2.5 seconds, 195,000 spins is more than five days of non-stop play, and 39 million spins would take about three years. That's why RTP is a property of a game's design rather than something anyone can measure from their own play, and why a winning or losing session is not evidence that a game is "hot", "cold" or broken.
If you think a game didn't behave as its rules say, that's a matter for a complaint, not a hunch: our guide on how to complain about a UK gambling company explains the steps.
How to use RTP and volatility
- Use RTP to compare cost, not to predict results. Between two otherwise similar games, the higher RTP costs less on average. It doesn't make a winning session more likely in any useful sense.
- Find the figures before you play. RTS 3 requires the information to be available before you commit; it's usually in the game's help or information pages. Check the figure in the game you're actually playing, at the casino you're playing it at.
- Treat volatility as a warning about swings. A high-volatility game will usually lose faster in a short session, with a small chance of a large win. If you set a budget for a session, expect a high-volatility game to use it up more often.
- Don't chase a game. Outcomes are random and games can't be "compensated", so a long losing run doesn't make a win more likely. Chasing losses is one of the warning signs of harm.
Editorial view: RTP is worth knowing because it tells you what a game costs on average. Volatility is worth knowing because it tells you how unlikely the average is to describe your own evening.
Questions and answers
Does a 96% RTP mean I'll get £96 back from £100?
No. It means the game is designed to return 96% of all the money staked on it over a very long run of play by everyone. In a single session you could finish well above or well below that; in our hypothetical high-volatility game, the typical 200-spin session lost more than half its stakes.
Are high-RTP slots better?
They cost less on average, which is a real advantage. But RTP says nothing about how a game's wins are spread out, and the typical session on a high-volatility game can be much worse than on a low-volatility game with the same RTP.
Can a slot be "due" a win?
No. The Commission requires outcomes to be acceptably random and doesn't permit compensated games, so past spins don't change the chances of the next one.
Where do I find a game's RTP?
Online casinos licensed in Great Britain must make the RTP, or the likelihood of winning, available before you play. It's usually in the game's information or help pages.
Safer gambling
Gambling should be entertainment, not a way to make money: every slot has a house edge, and over time it costs money to play. If you're gambling more than you planned or chasing losses, free help is available 24 hours a day from the National Gambling Helpline, run by GamCare, on 0808 8020 133 in England and Scotland; in Wales, the Gambling Commission lists 0808 2819 265. You can set deposit limits with your casino, take a break, or self-exclude from online gambling with businesses licensed in Great Britain through GAMSTOP. Our safer gambling page explains the tools.
Related on DannyBet
- Casino wagering requirements after the 10x cap
- UK online casinos
- How we rate casinos
- How to complain about a UK gambling company
Sources and method
- Gambling Commission, RTS 3: Rules, game descriptions and the likelihood of winning (remote technical standards, last updated 7 May 2024; read 15 November 2026): the information that must be available before play, including RTP or the likelihood of winning.
- Gambling Commission, RTS 7: Generation of random outcomes (last updated 7 May 2024; read 15 November 2026): random outcomes, no compensated games, no misleading near misses.
- Gambling Commission, RTS 14: Responsible product design (last updated 12 January 2026; read 15 November 2026): the 2.5-second game cycle, no turbo or quick spin, no celebrating a return at or below the stake.
- Gambling Commission, Online slots stake limit guidance (updated 16 June 2025): £5 per game cycle for players aged 25 and over, £2 for 18 to 24.
- Calculations:
editorial/calculations/slot_rtp_volatility.pyin the DannyBet repository (Python, standard library). The three games are hypothetical pay tables, each with an RTP of exactly 96%. RTP, hit rate and standard deviation are exact; the spins for ±1 point are a statistical approximation (independent spins, normal approximation, 95% confidence), not a guaranteed threshold; the session results are from a seeded simulation of 20,000 sessions per game (seed 20261115), and the script also prints the standard error that explains how far a simulated average can sit from the expected result. - Help and self-exclusion: GamCare, GAMSTOP.
Sources last checked: 15 November 2026.