What does online slot RTP mean? Two games, one average

What 95% describes

Return to player is a ratio between expected prizes and stakes. In a simple game with a £1 stake and a 95% theoretical RTP, the expected gross prize is 95p per play. Gross matters: that prize figure is measured before subtracting the £1 paid to take part.

The expected net result is therefore a loss of 5p per play under the model's assumptions.

That statement does not predict a particular spin. It also does not mean a £20 deposit will become exactly £19. A deposit and total money staked are different quantities, especially if returned money is used again.

This workshop makes the distinction concrete with two invented games. They are mathematical examples with complete outcome tables, not available products, game recommendations or reconstructions of a commercial slot's hidden probability model.

Game A: frequent small returns

Imagine a random draw from 100 equally likely tickets, with the ticket replaced before the next draw. Fifty tickets return £0, 45 return £1, and five return £10. Every draw costs £1.

The possible outcomes and their probabilities are now completely specified; there is no undisclosed feature, progressive contribution or changing state.

Multiply each prize by its probability and add the results: 0.50 × £0 + 0.45 × £1 + 0.05 × £10 = £0.95. Dividing by the £1 stake gives 95% RTP. Half the tickets return something, but 45 out of 100 merely return the stake.

Calling all of those outcomes profitable wins would hide an important distinction. Under this particular table, only the £10 outcome produces a positive net result.

Game B: fewer returns, a larger top outcome

Keep the £1 stake and 100-ticket model, but change the contents. Ninety tickets return £0, five return £1, and five return £18. Its expected gross prize is 0.90 × £0 + 0.05 × £1 + 0.05 × £18 = £0.95.

The theoretical RTP is still 95%.

Only ten tickets in every hundred return anything in Game B's probability model. The top outcome is larger, while the average remains unchanged. Our comparison therefore disproves a simple inference: two games with the same RTP do not have to distribute prizes in the same way.

The chart shows probability weights, not a recorded hundred-spin session. Sampling 100 independent draws does not force every ticket type to appear exactly in its stated proportion.

Name the quantity before comparing it

For these toy games, define hit frequency as the probability of receiving any prize greater than zero. That produces 50% for A and 10% for B. Define profitable-result frequency as the probability of receiving more than the £1 stake.

That produces 5% for both. These definitions are explicit so the figures can be checked; a commercial description may use different terminology.

The spread of outcomes also differs. A compact way to quantify it is the standard deviation of the gross prize for one £1 draw. The specified tables give approximately £2.13 for A and £3.92 for B.

This is a calculation from our tables, not an operator's volatility label. It describes dispersion around the mean, not the maximum possible loss, a safe stake or a recommended playing style.

The average is not a settling-up promise

If the ticket is replaced and the model remains unchanged, a new draw uses the same probabilities after a loss as after a prize. There is no balancing mechanism that owes a particular participant money. A streak does not create an obligation for the next outcome to restore the advertised average.

That distinction also has a regulatory context. The Gambling Commission's RTS 7 requires acceptably random outcomes and prohibits adaptive, compensated behaviour within its stated scope. It addresses the implementation of regulated remote games; our ticket examples provide a separate transparent way to understand the arithmetic.

Neither the standard nor the example allows us to infer a commercial game's full probability table from its artwork or a short recording. The published rules and the actual game version remain important evidence.

Sources: RTS 7: random outcomes.

Use the calculator to check the model

The worksheet lets you choose A or B, a hypothetical stake and a fixed number of draws. It calculates total stakes, expected gross return and expected net result using the disclosed 95% ratio. Change the game and those expectations stay the same, while the hit frequency and single-draw spread change.

That is the comparison the exercise is designed to reveal.

At £1 for 100 predetermined draws, total stakes are £100, expected gross prizes are £95 and the expected net result is minus £5. These are model expectations across possible outcomes. They are not a likely-looking cash balance, a promise about an individual session or a calculation of what can be afforded.

The tool does not simulate play, accept money, contact a casino or save entries.

Why a deposit is not the denominator

Suppose a hypothetical account starts with £10. A person could stake some of that money, receive a return, and stake some returned money again. Total stakes would then count each amount committed, not just the original deposit.

A claim that 95% RTP means losing only 5% of the deposit confuses these two accounting measures.

Our worksheet avoids that confusion by fixing the number of draws in advance and displaying total stakes explicitly. It does not model stopping when funds run out, changing stakes, deposits, withdrawals or bonus conditions. Those processes can change the distribution of a final balance and the amount ultimately staked.

They cannot be reconstructed accurately by treating a deposit as though it were a single wager. Keep the denominator attached to the number whenever discussing return.

Take three questions back to the rules

When reading a real online slot description, ask which game version the information belongs to, which mode or feature it covers, and what the percentage actually describes. The Commission's RTS 3 sets requirements for understandable game rules and information about the likelihood of winning before commitment within its scope.

RTP is one form that likelihood information can take; the standard does not turn it into an individual outcome guarantee.

Then separate what is known from what is absent. A percentage alone does not supply a complete prize distribution, the chance of a particular feature, or a forecast for tonight.

The useful result of this workshop is an improved reading habit: identify the average, inspect the outcome structure where it is disclosed, and keep hypothetical mathematics clearly apart from evidence about a named commercial game.

Sources: RTS 3: rules and likelihood of winning.