Return to player is the most flattering phrase in the lobby, since it names the money coming back and stays silent about the money that does not. A 96% slot returns 96 units of every 100 wagered and keeps four. Most industries call that a margin, and this one prints it on the tile as a selling point.
The awkward part is whose lifetime we are discussing, because the game gets millions of spins and you get an evening. A slot can hold 96% for its entire commercial career and still empty your balance before the kettle boils. The game keeps its promise, because the average belongs to the machine rather than to you.
A slot arrives with a menu of RTPs rather than a single one. Play’n GO, in example, ships Book of Dead in five certified builds, running from 96.21% down to 84.18%. Same Rich Wilde, same expanding symbol, same 5,000x ceiling, with twelve points between the top version and the bottom. The casino picks, the game says nothing, and the answer sits two taps away in a menu nobody opens.
Reviewed byNadia WinchesterThe maths model produces the RTP long before a game reaches a lobby. Studios total every possible outcome, weight each by how often it occurs, and divide what comes back by what goes in. Billions of simulated spins confirm it behaves, and a testing lab checks the sums before launch. The figure arrives by arithmetic, with no real player involved anywhere in it.
A slot needs an enormous number of rounds before real results settle anywhere near the published figure. Across a few hundred spins the actual return can land almost anywhere, and players finish sessions at 300% or at nothing without either outcome saying a word about the game. The percentage describes a curve that only straightens across a scale no individual will ever play. Your session is a sample of one, drawn from a distribution built to be read in millions.
A 96% RTP does not mean a €100 deposit leaves with €96, because the percentage applies to everything staked. Win €20, stake it again, and that €20 arrives as fresh turnover with the same four percent coming off it. The same money passes through the margin repeatedly, which is how €100 generates around €2,500 of turnover at a euro a spin before it runs out. That buys 2,500 spins, and it explains why the house edge collects far more than four euros.
Three bands cover the whole market, and the gaps between them are smaller than the arguments about them suggest. A couple of percentage points separates a generous slot from a stingy one, which sounds trivial until you multiply it by every spin of a session.
| RTP band | How common | What lives there |
|---|---|---|
| 97% and above | Rare | Classic formats, thin feature sets, low ceilings |
| 95% to 96.99% | The default | Almost every modern video slot, at every volatility level |
| Below 95% | More common than the lobby admits | Trimmed builds, licensed brands, progressive jackpot titles |
Games above 97% exist, and nearly all of them are paying for the privilege somewhere else on the spec sheet. Relax Gaming publishes Book of 99 at a flat 99%, then honours it with a stripped-back book slot carrying one feature and a modest ceiling. Thunderkick’s 1429 Uncharted Seas holds 98.6% across twenty-five lines with nothing a modern player would recognise as a bonus round. The maths has to balance, so generosity in the return arrives as smaller peaks rather than bigger ones.
The famous figures also come with small print worth reading. Mega Joker reaches 99% only in Supermeter mode at maximum coins, which describes a way of playing rather than a property of the game. Open it, spin at base level, and the number on every top-ten list stops applying to you entirely.
Rated 92.4 of 100%
Rated 93.6 of 100%
Rated 96 of 100%
Most of what any lobby stocks sits in this band, and the clustering answers to commerce rather than coincidence. Studios build toward a target that leaves the operator a workable margin while staying respectable enough to print on the tile. Push higher and the revenue share stops working, drop lower and the players who check start noticing.
Calling this range average obscures how much it contains. Two slots both certified at 96.2% can behave like unrelated products, since the percentage describes the total and stays silent on the delivery. One drips small wins across an hour and the other holds everything back for a bonus round you may never reach, which is a question of volatility rather than return.
Rated 78 of 100%
Rated 81 of 100%
Rated 91 of 100%
Rated 85 of 100%
Rated 94 of 100%
Three separate things push a game under 95%, and only one involves the studio being stingy. Trimmed builds account for most of it, since a slot certified at 96.21% and deployed at 91.25% lands here without a single pixel changing. Licensed titles carry the cost of the film, band or franchise in the maths. Progressive jackpots divert a slice of every stake into a pool almost nobody collects, which drags the base game down while the advertised total stays flattering.
None of that makes these games unplayable, and plenty of people knowingly pay the extra margin for a theme they love or a jackpot worth chasing. The bill still arrives though, because moving from 96% to 92% doubles the house edge rather than nudging it. That turns four cents a spin into eight, and across a few thousand spins it is the difference between a long session and a short one.
Rated 81 of 100%
Rated 87 of 100%
Independent labs test every game before it reaches a lobby, and the same names keep appearing on the paperwork: GLI, eCOGRA, iTech Labs, BMM Testlabs. They put the random number generator through statistical batteries, run the maths model across simulated volumes no human could reach, and confirm the game returns what its documentation claims. Each RTP build goes through this separately, so the 84% version of a slot carries its own certificate sitting right beside the 96% one.
That detail deserves a moment. Certification confirms a game behaves as declared, which is a different job from deciding the declaration is reasonable. A stamped 88% build is exactly as certified as a stamped 96% one, because the lab audits arithmetic and holds no view on whether the arithmetic likes you.
A certificate confirms four things and stops there:
Regulators sit above the labs and agree on remarkably little. Malta dropped its floor to 85% in 2021, which clears the way for builds no player would knowingly choose. Italy holds fixed-odds games including slots to 90%. Elsewhere the rules cover disclosure rather than generosity, and a few markets ask for neither, which is how the same title ends up wearing four different percentages across four different licences.

A studio does not build one game and attach one percentage to it. It builds the maths model, certifies several versions as separate approved builds, and the aggregator serving that game to casinos presents those builds as settings. An operator selects one per brand or per market, and the reels, the bonus rounds and the max win stay precisely where they were.
Play’n GO got there first and the industry followed, so Pragmatic Play, NetEnt, Red Tiger, Nolimit City, Quickspin, Relax Gaming, Thunderkick and Hacksaw all ship games this way now. The tiers land in roughly the same places across the market:
| Tier | Typical RTP | Where it turns up |
|---|---|---|
| Default | 96% | The figure quoted in reviews, marketing and provider sheets |
| Second | 94% | Sites funding heavy bonus programmes |
| Third | 92% | High-tax markets and margin-led brands |
| Floor | 88% and below | Licences with no disclosure requirement |
Some studios stop at three settings and keep the whole range inside two points, while others run five and open a canyon reaching down past 84%. The risk of meeting a poor build therefore depends as much on which studio made the game as on which casino serves it. Your version was chosen against a margin target by somebody who will never meet you.
Take the same turnover through two builds of one game and watch what happens to the operator’s side. At 96% the house keeps four units in every hundred wagered, and at 94% it keeps six. Nothing about the player experience changed, and the site just increased its revenue on that game by half.
Drop to the 88% tier and the hold reaches twelve units per hundred, three times the default. That is why the settings exist and why the lowest ones keep finding homes, particularly in markets where nobody has to publish the figure. Two percentage points feel like rounding to a player and read as a strategy to a portfolio manager.
Every game in the casino carries a return figure, and slots sit closer to the bottom of that list than the marketing suggests. Line them up against the table games and the reels start looking like an expensive way to spend an hour.
| Game | RTP at its best | What it asks of you |
|---|---|---|
| Video poker, 9/6 Jacks or Better | 99.54% | The correct hold on every single hand |
| Blackjack | 99.5% | Memorised basic strategy and a friendly rule set |
| Baccarat, banker bet | 98.94% | Ignoring the tie bet forever |
| Craps, pass line | 98.59% | Staying on the simplest bet at the table |
| European roulette | 97.30% | Nothing at all |
| Slots | 92% to 97% | Nothing, and the build was picked before you arrived |
| American roulette | 94.74% | Nothing, at twice the cost of the European wheel |
Those table game numbers come with homework attached. Video poker hits 99.54% only for someone making the mathematically correct decision on every hand, and a few sloppy holds an hour drag it down by more than a point. Blackjack pays best to players who learned basic strategy and follow it when it feels wrong, which is most of the interesting moments.
Slots ask nothing, and that is both the appeal and the invoice. No decision at the reels moves the return, so the figure stays wherever the operator left it before you opened the game. European roulette demands equally little and still beats a great many slot builds, which is worth knowing before dismissing the wheel as an old person’s game.
The cheapest games in the building are frequently the simplest ones. Crash and dice titles often run between 97% and 99%, so a crash game with two buttons and no artwork can hand back more than a slot with a full orchestral score.
A published RTP covers the base game under normal conditions, and modern slots spend a lot of time outside normal conditions. Every button that changes how you play can change which figure applies to you.
RTP describes a total and stays silent on the shape of it, which is why two identical percentages produce sessions with nothing in common. That shape has a name and a number of its own, and volatility is where the difference between a gentle evening and a brutal one actually lives.
The figure your casino is running sits inside the game itself, and checking it costs less time than reading the bonus terms. Do it once per title per site, because a version that was 96% last year may not be today.
Free play loads the same build the site serves for real money, so the number you find there is the one you will be playing against. Nothing needs depositing and no account has to exist yet, which makes this the cheapest research available anywhere in the lobby.
Every licensed slot carries one, usually bottom left or bottom right of the control bar. Tapping it opens the rules, the paytable and the information screens, all of which the studio built and the regulator approved.
It appears in the rules section, often on the final screen after the paytable and the feature descriptions. Some studios put it up front, others bury it behind three swipes, and the wording stays close to identical across the industry.
Provider websites list the figure they built the game around, and a gap between that number and the one on your screen means the operator selected a reduced build. A game showing 92% when the studio publishes 96% is telling you exactly what kind of site you are on.
The buy carries its own return figure and the rules screen states it alongside the base game. Two seconds of reading settles what a hundred times your stake actually purchases.
Reputable sites display it because their licence expects them to, and a game hiding its return has told you something useful about the operator. Plenty of casinos publish the figure without being asked, and those are the ones worth your deposit.

Percentages stay abstract until money gets attached, so put €100 into a game at €1 a spin and keep re-wagering whatever comes back. Nothing on the screen changes between these three builds, and the sessions end nowhere near each other.
All three builds lose. Choosing the top one converts a bad expectation into a slightly less bad one rather than into any kind of edge, and anyone selling high RTP as a strategy has confused a smaller loss with a profit.
What the extra points genuinely buy is time on the reels, which means more bonus rounds reached and more room for a good run to turn up before the money runs out. Treating the percentage as the price of an evening puts it in the right frame, and the gap between paying €4 and €8 per hundred wagered is worth the thirty seconds it takes to check.
RTP is the most useful number on a slot and the most frequently oversold one. It tells you what a game returns across a lifetime of spins, which makes it excellent for comparing two titles and useless for predicting the next hour. Reading it as a promise about your session is the mistake almost everyone makes on the way in.
The number also moves. A slot sitting at 96% on one site can run at 92% on another with nothing visible to separate them, and the version you get was decided by an operator weighing margin rather than by anything you did. Checking takes less time than reading a bonus term, and it stays the single most valuable thing on this subject that a player can actually control.
Picking a better build buys time on the reels rather than an edge, and time is worth having. More spins means more chances for the game to do something interesting before the balance runs out. Read the figure, know what it does and does not cover, then choose the games you actually want to play.
No, and the two measure completely different things. RTP describes how much a game returns across its lifetime, while hit frequency describes how often a spin pays anything at all. A 97% slot can leave you empty-handed for fifty spins and then hand back everything in one, and a 94% game can pay small amounts constantly while draining the balance steadily. The percentage covers the total rather than the pacing.
Not mid-session, since the build loads with the game and stays put until you close it. Operators can switch a title to a different certified version between sessions though, so a slot you checked six months ago deserves a fresh look. The change happens at deployment rather than on the fly, and nothing obliges the site to announce it.
The game’s return stays exactly where it was, because a bonus changes how you funded the spin rather than how the maths works. Wagering requirements are the thing worth reading instead, since they decide how much you have to stake before the balance becomes yours. A generous bonus attached to a low RTP build is a slower way of arriving at the same place.
They are running different certified builds of one game. Studios ship several approved versions, the operator picks one against its margin targets, and the artwork stays identical across all of them. The figure in the game’s own info screen is the only one that applies to you, regardless of what any review page says.
That depends entirely on the licence. Malta sets a floor of 85%, Italy holds fixed-odds games to 90%, and several jurisdictions set no numeric minimum at all while requiring operators to make return information available. A game licensed somewhere with neither rule can run as low as the studio built it.
Anything at 96% or above puts you in decent company, since that is where most studios set their defaults. Below 95% is worth a second look at what you are paying for, and below 92% means the operator chose a build with you in mind rather than the studio. The number matters most as a comparison between two versions of the same game.

The Author

The Author
Slots and News Reviewer