What RTP actually describes
A game at 96% RTP returns 96% of everything wagered on it back to players, measured across millions of rounds. It does not mean you get 96 pence back from every pound tonight, and it does not mean a losing session is unusual.
The figure is a property of the game’s mathematical model, calculated by the studio and verified during certification. It assumes optimal play where decisions exist, and infinite rounds where they do not.
Your actual session sits somewhere in a distribution around that average, and for any realistic number of spins the distribution is wide.
The distribution nobody publishes
Here is what the market looks like across the slot catalogue we track:
| RTP band | Games | Share of sample |
|---|---|---|
| below 94% | 50 | 1.9% |
| 94-95% | 17 | 0.6% |
| 95-96% | 47 | 1.8% |
| 96-97% | 2472 | 92.5% |
| 97% and above | 87 | 3.3% |
Across 2673 game entries tracked in our database.
Ninety-two percent of games sit inside a single percentage point. That is not a coincidence and it is not one operator’s choice: studios build to the band operators will accept, because a game paying noticeably more is harder to place.
The practical consequence is that choosing a slot on RTP alone is mostly noise. The difference between a 96.2% title and a 96.6% one is swamped by volatility long before it appears in your balance.
Where RTP does matter
Three places, all of them at the edges rather than the middle.
The tails. A small number of games clear 97% and a smaller number sit below 94%. Those are worth identifying, and our RTP rankings list them.
Between builds of the same game. Studios licence several RTP configurations of the same title and operators choose which to run. Gaps above two percentage points exist under identical names and artwork, which is covered in same game, different RTP.
Between game types. The gap between a blackjack table and a keno draw is enormous, far larger than anything within a single category.
Why you should not trust a quoted RTP
The figure a review quotes usually comes from a studio press release. The figure that applies to you is the one running at the casino where you are playing, and those can differ.
Open the game’s information panel before you play. It takes a few seconds and it is the only source that describes the build actually in front of you.
Be sceptical of any site publishing a “real” or “measured” RTP derived from live bet feeds. Those feeds over-represent wins, and the arithmetic produces returns above 100%, which is impossible. We explain why we refuse to use that data on our methodology page.