What a Slot's RTP Number Means

Every licensed online slot in the UK carries an RTP figure on its information page. The percentage sits somewhere between 94% and 97%, next to the variance rating and the maximum win. The Gambling Commission requires operators to publish it, and operators make sure you see it when the number is high. The result is the most quoted and least understood figure in online gambling. Most players read 96% as a refund rate: put £100 through the game, get £96 back. The industry prefers that reading to stand.

RTP stands for return to player. It describes a mathematical expectation calculated across millions of spins, and it says nothing about any single session. The figure shows what a game would pay back if you wagered the same stake, spin after spin, and averaged the results. A slot with a 96% RTP keeps 4p of every pound wagered over that near-infinite run. The flip side is the house edge, which for that game is 4%. That 4% is the casino's margin and the price of the entertainment.

The definition matters because the marketing around slots encourages you to treat the figure as a personal outcome. It is not. Game developers calculate RTP from the full paytable and the probability of every symbol combination, then independent testing labs verify the result. When you see 96.2% on a slot's information page, you are looking at the average outcome of every spin the game can produce, weighted by how often each spin occurs. Nothing about that average has happened to you yet.

A coin toss makes the logic visible. Offer someone a flip where a £1 stake returns £1.90 on heads and nothing on tails. Each toss has an expected value of -5p: a £0.90 profit half the time against a £1 loss the rest, which averages to a 5p loss per toss. After a thousand tosses the average sits close to that, but no single toss pays 95p. A slot resembles that machine, with more outcomes and a wider spread, and you spin it hundreds of times per hour.

Why your session never matches the number

Take a slot with a 96% RTP and a stake of 50p. Spin it 500 times and you will wager £250. The expected loss is 4% of that, £10. The range of actual outcomes from 500 spins is far wider than that. You could end the session £80 down or £60 up. Both results are common, and neither proves anything about whether the published RTP is accurate.

The reason is variance, which casinos market under the name volatility. RTP is an average across a near-infinite sequence of spins, and the shorter the sequence you play, the wider the gap between the average and your result. No slot has a memory. A spin is an independent event, so a full-screen win lands with the same probability on spin one as on spin 400. A dead stretch of forty empty spins and a bonus round that pays 300x can sit inside the same 500-spin block. The order they arrive in determines your balance; the RTP calculation takes no account of order.

Public review pages let you watch the gap in action. Browse the Trustpilot reviews for slotastic casino and you will see the standard mix: rapid losses, occasional big wins, and nothing that looks like the published RTP arriving on schedule. The mix matches the paytable's probabilities. The average appears only over hundreds of thousands of spins.

Doing the maths on a paytable

A paytable turns the abstraction into numbers you can check. The example below is a simplified version of a modern six-reel slot, with every outcome squeezed into eleven rows. The payout column shows what the game returns for each winning combination, in multiples of your stake. The probability column shows how often that combination occurs. The contribution column is the payout multiplied by the probability. Adding up that column produces the RTP.

Read the table top to bottom and the big-payout rows catch the eye first. A 1,000x win from a single spin looks like the reason people play. Look at the probability column and the pattern flips. The 1,000x outcome contributes 0.5% of the 96% RTP. The modest 2x win contributes 16%, even though it pays 500 times less, because it occurs 16,000 times more often. The rare rows create the excitement; the common rows carry the percentage.

This is why the RTP on the info page never arrives in a single session. The percentage is the sum of hundreds of these contributions, and the sum settles down only after millions of spins. Play 2,000 spins at £1, a solid evening on most slots, and you have wagered £2,000. The expected loss at 96% RTP is £80. The actual result will sit somewhere in a wide band around that, and a £400 swing is well within normal range. The casino does not owe you £80 back, and it is not cheating you when the result differs. The paytable describes a million-spin game; you are playing a two-hour one.

What the house edge costs per hour

The house edge converts the RTP into a running cost. Spin speed on online slots sits around 600 spins per hour, sometimes faster. At £1 a spin, that is £600 wagered per hour, and a 4% house edge prices the hour at about £24. At £5 a spin, the same hour costs about £120. Those are the numbers to compare against any other entertainment you buy.

The cost scales with time and stake. If you play 100 spins a week at £2, you wager £200 and the expected loss is £8. Over a year that becomes 5,200 spins, £10,400 wagered and about £416 lost. You might lose it in one night or spread it across fifty; the price is the same.

What the published RTP does not show

The RTP number on an information page is not always the whole story. The same game can exist in several RTP versions, with the operator choosing which one to run. The Gambling Commission requires the displayed figure to match the version in use, and suppliers can change the maths when they update a game. A slot that showed 96.5% in an old review might show 95.2% after a refresh. Check the information tab inside the game, not a screenshot from last year.

Progressive jackpot slots add a second layer. A game funding a six-figure jackpot takes the jackpot contribution out of each spin, and the published RTP may include that contribution even though no player receives it in their own session. Strip out the jackpot and the base game returns less. The same logic applies to bonus rounds: some games pay better inside the feature than outside, and the blended RTP hides a split you cannot choose.

Bonus funds complicate the arithmetic further. A £20 welcome bonus with a 35x wagering requirement demands £700 of wagers before the winnings become withdrawable, and the house edge applies to every pound of that £700. At 4%, the wagering requirement alone carries an expected cost of £28 before the first spin. A generous-looking bonus on a low-RTP slot can end up worse than no bonus on a high-RTP one.

What the RTP number is good for

RTP works as a comparison tool, not a prediction tool. When you choose between two games, the one with the higher figure is the more generous one over the long run. Pair that with the variance rating above it. A 97% RTP high-volatility slot can drain a £50 deposit in twenty spins. A 94% RTP low-volatility game can stretch the same £50 across several hours. The higher RTP does not make it the better choice for your evening.

The second use is expectation setting. Once you know the house edge, you know the average price of an hour at that game, so you can decide what the entertainment is worth before the first spin. The same arithmetic justifies the responsible gambling tools built into licensed operators' sites: deposit limits, time-outs, reality checks, and self-exclusion through GamStop. The maths is not on your side over time, and those tools exist because of that fact. GamCare and BeGambleAware offer the same advice in plainer terms: set the limit before the session, not during it.

The next time a slot advertises 96% RTP, read it as a description of a million spins by a hypothetical average player, not a receipt for your next hour. Compare it across games, factor in the variance, count the cost of any wagering requirements, and set a budget that matches the price. The number will not tell you whether you win tonight. It tells you what the game costs, and that is the figure you can use.