Why 96% RTP does not mean you get £96 back from £100
RTP is an average share of all stakes returned over a vast number of plays by everyone. It isn't a refund on your £100, and any one session can land far from it.
Add as a preferred source on Google
Because 96% describes the average share of all money staked on a slot that comes back as prizes. That average is worked out across an enormous number of plays by everyone using the game. It is not a refund rate on your £100, and it says nothing reliable about what any one session will return. Results are random, and over time the game is designed to keep the other 4% of everything staked.
What does the 96% actually describe?
For online games, the Gambling Commission requires operators to make information about your chances easily available before you commit to gamble. That can be the return to player (RTP) percentage, the house edge or the probability of winning outcomes, alongside a description of how the game works and how prizes are decided. Where an RTP is shown, it is the game's theoretical RTP: the figure that comes from the game's design and its maths.
That figure is checked before release. The Commission's testing requirements for online games include verifying the game's maths and theoretical RTP, partly by simulation. The game is set to play automatically over a high number of games to confirm that the actual return lands within an acceptable range of the expected one. Notice the scale involved. The 96% is a property of the game across a huge volume of play, not a promise about a few hundred spins.
Why is your £100 the wrong yardstick?
RTP is measured against turnover, which the Commission's guidance defines as the total of all stakes made on a game, including winnings that are staked again. That detail matters more than most people expect. Start with £100, keep staking whatever comes back, and the total you put through the game can quickly pass £100.
Here's a hypothetical example. Say a game has a stated RTP of 96% and every spin costs £1. On average, the game is designed to return 96p of every £1 staked and keep 4p.
| Total staked (hypothetical) | Average returned at 96% | Average kept by the game |
|---|---|---|
| £100 | £96 | £4 |
| £500 | £480 | £20 |
| £1,000 | £960 | £40 |
In the second and third rows, the same £100 could have been the starting balance, recycled through the game several times. The average cost is 4% of everything staked, not 4% of the money you started with. These figures are long-run averages, not forecasts. An actual session could finish well above or well below them, and losing the full £100 is entirely possible.
Why can one session land so far from the average?
Each result is random. The Commission requires online game outcomes to be acceptably random: distributed in line with their theoretical probabilities and unpredictable. It also prohibits compensated, or adaptive, behaviour in online games, so a game can't change its probabilities during play to steer itself towards its RTP. A run of losses doesn't make the next spin more generous, and a big win doesn't make it meaner.
So the average only emerges through sheer volume. In guidance about gaming machines in venues, the Commission makes the same point directly: a machine displaying an 85% RTP does not mean you should expect to win back an average of 85p for every £1 you stake during a playing session, because the figure is an average over a significant number of plays. Those machines operate under different rules from online games, but the arithmetic of averages works the same way.
Volatility widens the gap. A highly volatile game draws much of its return from prizes that are very large but rare, and many short stretches of play will contain none of them. Those rare prizes are still counted in the 96%, which is part of why a typical session can look nothing like the headline figure.
Is 96% exact even across everyone's play?
Not precisely, and the Commission's own guidance shows why. Online operators must monitor how their games perform live, comparing the actual RTP (total wins divided by total turnover over a period) with the theoretical figure. The acceptable gap depends on the game's volatility and narrows as more play builds up.
In one worked example in that guidance, a game designed to return 91.68%, with its particular volatility, could return anywhere from 89.93% to 93.43% after 400,000 games and still be treated as performing as expected. If a game measured across hundreds of thousands of plays can sit nearly two percentage points either side of its design, a single person's session is no test of the number at all.
What does the 4% tell you?
The gap between 96% and 100% is the share of stakes the game is designed to keep over time. The Commission's monitoring guidance calculates a game's yield as turnover minus wins, which is that margin expressed in money. A higher RTP shrinks the average margin, but it doesn't remove it, and it doesn't make any session profitable. Gambling costs money, and the more that goes through a game in total, the larger the amount it is designed to keep on average.
The point to keep
Read 96% as a description of the game, not a forecast for your money. It tells you the share of all stakes, from all players, over a very long run, that the game is designed to pay back as prizes. Your own result will be random and can fall a long way from it, and the cost grows with the total you stake, not the amount you started with.
Sources: UK Gambling Commission, Remote gambling and software technical standards: RTS 3, rules, game descriptions and the likelihood of winning; UK Gambling Commission, Remote gambling and software technical standards: RTS 7, generation of random outcomes; UK Gambling Commission, Testing strategy for compliance with the remote technical standards: procedure for testing; UK Gambling Commission, Live return to player performance monitoring of games of chance; UK Gambling Commission, How to calculate return to player (RTP); UK Gambling Commission, Return to player: how much gaming machines pay out.
How this was written. Drafted with AI assistance from the sources listed above in Harvey Watkins's style and published after automatic checks for accuracy and compliance, without a line-by-line read by an editor. If you spot a mistake, tell us. Our editorial standards.


