Slot RTP and volatility explained: the house edge and hit frequency

Slot RTP and volatility are best explained with money on the table. Set a £1 stake on a game published at 96% and the arithmetic is unforgiving. Every £100 that passes through the machine returns £96 on average, so the house keeps £4. That average is calculated across millions of spins, which is a span no player ever completes. What you get instead is a slice of the distribution.

What slot RTP and volatility actually describe

Two figures do most of the work, explained properly below, and slot machine RTP is only the first of them. Return to player is the share of all stakes a game pays back over its full cycle. Volatility describes how roughly it does so. A high-volatility title reaches the same long-run figure through rare, large wins, while a low one drips small returns back steadily. Both can publish 96%.

That is the part worth holding on to. Slot RTP and volatility are not two ways of saying the same thing. Reading one without the other tells you very little about what an evening will feel like. A high-variance game at 96.5% can empty a small balance faster than a steadier one at 95.5%.

The house edge, in pounds

The house edge is easier understood in pounds than in percentages, so convert it. A 4% house edge on a £1 spin is four pence of expected cost per spin. Play six hundred spins in an hour, a routine pace on autoplay, and the expected cost is £24. Nothing about that figure is hidden or unfair. It is the published price of the entertainment, and it is the same arithmetic a roulette wheel runs at 2.7%.

The catch is that expected cost and actual cost rarely match over one session.

For scale, the table games sit an order of magnitude lower. A live blackjack table played to basic strategy runs a house edge of 0.72%, against 4% here. Our page on live dealer blackjack works through why that figure assumes flawless play.

Hit frequency is the number nobody quotes

Hit frequency is the share of spins that return anything at all. A game at 25% pays on one spin in four, which sounds generous until you notice what counts as a hit. A spin that stakes £1 and returns 30p is recorded as a win. Most paying spins on a modern video slot do exactly that.

So the machine can feel active while the balance falls. That is most of why you lose playing slots over a long session. That gap between how often something happens and how much it returns is the single biggest reason a session ends early. It is also why you lose playing slots even on a title with a healthy percentage. Providers publish hit frequency far less consistently than slot machine RTP, which makes it harder to compare than it should be.

Variance, explained with 40,000 simulated sessions

We modelled a high-volatility game at exactly 96% and ran forty thousand sessions of ten thousand spins at £1. The model is a simplified one, and its assumptions are stated below rather than buried. A model whose workings are hidden is only a tidier guess.

Outcome after 10,000 spins	Result	What it tells you
Median session	−£766	The middle outcome. Half of the forty thousand sessions finished worse than this, which is exactly what a headline average hides.
Sessions finishing ahead	41.3%	Four in ten still ended up, on a game with a negative expected return. A short run beats the edge often; a long one does not.
10th to 90th percentile	−£3,401 to £3,003	Eight sessions in ten finished somewhere inside this range. The width of it is the point — one game produced both ends.
Best single session	£12,982	One session out of forty thousand. Outliers this size are what drag the mean above the median, and why the mean flatters the game.
Mean session	−£409	Close to the theoretical −£400, which is the check that the model behaves — not a result any single evening would feel like.

Method: 40,000 independent sessions, 24.5% hit frequency, six prize tiers scaled to 96.00% expected return, fixed seed 20260803. Mean landed at −£409 against a theoretical −£400, confirming the model behaves.

The mean is the honest headline and the median is the useful one. Expected loss across every session was £409, close to theory. Yet the typical player did worse than that. A handful of very large wins drag the average up while most sessions sit below it. Four in ten still finished in profit.

Which points at something the percentage cannot tell you: the figure describes the game, not the session, and those are different objects.

Where the published RTP figures come from, and where they do not

In Britain the operator carries the licence, not the game. The Gambling Commission requires that the information needed to judge your chances of winning be easily available before you commit to gamble. It specifies that this must include the return to player percentage. The duty falls on the site serving the game rather than the studio that built it, which turns out to matter. Testing houses such as eCOGRA and GLI verify that a game’s output matches its stated return before it reaches the market.

There is still an awkward gap between that duty and what you can look up in advance. No major provider publishes RTP on a public web page. The figure lives in the in-game information panel. So it is genuinely disclosed and simultaneously out of reach, until you have an account at the site. Compare a game across two casinos and you have to open it twice.

Which is the honest position for this catalogue too. Our figures are cross-checked across independent slot databases, not lifted from a provider sheet, and we say which sources agree on each entry. Where they disagree the field stays blank. One title here has three irreconcilable published returns, so it carries no percentage at all. An estimate dressed as a measurement is worse than an empty cell.

Questions about RTP and volatility

Does a 96% RTP mean I get £96 back from £100?

No. It describes the long-run average across millions of spins, not your session. In our simulation of 40,000 sessions at that figure, the median player finished £766 down after ten thousand spins. Fewer than half finished ahead at all. The figure describes the game over its full cycle, not the slice of it you play.

Do slot RTP and volatility always move together?

Usually, but not always at the same stake. A 96.5% game with brutal variance can empty a small balance faster than a 95.5% game that pays often. Read the volatility band alongside the percentage rather than sorting on one column.

Do some titles ship with more than one RTP?

Yes, and this catches people out. Providers often release the same game at several settings, and the operator chooses which to run. The figure on a provider sheet may not be the figure at the site where you play it.

Is there a reason why you lose playing slots more often than you win?

Two, and neither is bad luck. The return is set below your stake by design, and most paying spins return less than you put in. Hit frequency counts any win, including a spin that hands back a third of your bet.

Does a higher hit frequency mean a better game?

Not on its own. Hit frequency counts any paying spin, including one that hands back thirty pence on a pound staked. A game can pay on one spin in four and still drain a balance steadily, which is why the figure only means something read next to the return.

Can a session actually beat the house edge?

Over a short run, easily. The top tenth of our simulated sessions finished more than £3,000 ahead, and the best single run cleared £12,982. The same distribution puts the bottom tenth past £3,400 down. Four in ten of those sessions still finished in profit despite the edge.