CURRENT EDITION
UK guide · Maths & worked examples
RTP and house edge without session promises
Return to player describes a theoretical long-run relationship between stakes and prizes for a game. It is not a promise about one person, one deposit or one evening of play.
- Use the right denominator
- Recognise the model boundary
- Do not turn averages into instructions
Use the right denominator
A hypothetical 96% RTP refers to returns relative to total stakes in the model, not a guaranteed refund of 96% of deposits. Re-staked winnings create additional turnover. Mixing deposits and stakes can therefore produce an incorrect estimate even before randomness is considered.
Recognise the model boundary
House edge is often expressed as one minus RTP for a fully specified game model. A 96% theoretical return corresponds to a 4% theoretical margin. Actual results over a short sequence can differ greatly. A negative expectation does not prevent a win, and a win does not refute the negative expectation.
Do not turn averages into instructions
Multiplying turnover by a theoretical margin can illustrate an average-model cost. It cannot determine your eventual balance, chance of clearing a bonus or affordable budget. Game configuration, payout distribution and promotional restrictions add further questions beyond the headline RTP.
An average needs a distribution
Expected return is a weighted average of possible outcomes. To calculate it from first principles, each outcome needs both a value and a probability. Two games can share the same average while having very different chances of returning nothing in a short sequence. That is why multiplying a promotion target by a house-edge percentage cannot establish the likelihood of completing the target. A proper probability model would need the payout distribution, stake rules, stopping conditions and fund treatment, among other assumptions. This library does not turn those missing inputs into a profit forecast. The educational value of RTP is to explain the long-run cost built into a game model and the limitations of claims that a bonus makes the outcome predictable.
- Read theoretical return from the actual game information rather than assuming every version of a familiar title is identical.
- Keep the expected model loss distinct from the maximum loss, realised loss and chance of exhausting the available balance.
- Do not use a past losing sequence as evidence that a future random outcome must compensate for it.
Worked example · hypothetical
A model calculation
With a hypothetical 96% RTP, £100 of aggregate stakes corresponds to £96 expected aggregate returns and £4 expected loss in the model. A real short session could lose the whole available balance or finish ahead. £4 is not a spending cap or a forecast.
Explore the arithmetic with the wagering calculator →
Common questions
Will a game repay earlier losses to reach its RTP?
Random independent outcomes do not create a personal repayment schedule. Do not continue because you believe the game owes you a return.
Does this guide recommend participating in a promotion?
No. The purpose is to explain terms and consumer questions. Declining an offer, stopping play or seeking support are valid outcomes of reading it.
Sources & further reading
Read the source alongside the explanation. Rules and published terms can change; check the current official position where you live.
How we research and maintain guides →EU explanation →