Reading a return figure without being misled by it
What the percentage is a statement about
Treat a return figure as an engineering specification rather than a promise. When Pragmatic Play prints 96.50% against Starlight Princess 1000, it is describing where the machine settles as the number of spins grows without limit. It is a property of the design, in the same way a stated fuel consumption is a property of an engine - true under measurement conditions nobody encounters on a Tuesday.
Two consequences follow immediately, and both get lost in casual explanations. The first is that the percentage applies to turnover rather than to your deposit. Every dollar passing through the machine is measured, including dollars you already won and are staking again, so a session that cycles a balance repeatedly is generating far more turnover than the sum you put in. The second is that the figure becomes trustworthy only at volumes no individual reaches. An evening is not a small version of the long run; it is a different thing entirely, governed by the distribution rather than by its average. That is the department of our volatility guide, and on a machine like this one it is the more useful of the two pages.
The arithmetic below shows what the specification implies as a pure average at various volumes. Read the right-hand column as the theoretical cost of the entertainment rather than as a forecast, and note how little the middle column resembles any actual session on a game whose payouts arrive in a handful of widely spaced events.
| Spins at C$0.40 | Turnover generated | Average returned at 96.50% | Average cost |
|---|---|---|---|
| 100 | C$40 | C$38.60 | C$1.40 |
| 500 | C$200 | C$193.00 | C$7.00 |
| 1,000 | C$400 | C$386.00 | C$14.00 |
| 5,000 | C$2,000 | C$1,930.00 | C$70.00 |
Averages only. A real run of 500 spins on a very high volatility machine will land nowhere near that middle column, in either direction, and the reason is the subject of the next section but one.
The slice the machine keeps
Whatever a return figure does not give back is retained, and on a 96.50% build that retained slice is 3.50% of everything staked. It funds the studio, the operator and everyone in between, and no commercial slot exists without one. Anybody implying otherwise is selling something.
The retained slice is useful for exactly one purpose: ranking two machines you already want to play. It is a poor guide to anything else, because it says nothing about how the remainder is handed back. A game keeping 3.50% and returning it in a thousand small pieces and a game keeping 4.00% and returning it in three enormous ones will feel like different hobbies, and the second one can empty a bankroll while behaving exactly as designed.
Where the 96.50% is actually handed back
This is the section most return explainers skip, and on the machines this site covers it is the one that changes decisions. A single percentage describes a distribution, and the distribution is nothing like flat. On a pay-anywhere multiplier game the great majority of that 96.50% is delivered inside the free spins round and inside the rare chains where a multiplier lands on a sequence that pays. The base game is not built to hand back 96.50%. It is built to hand back considerably less, and to finance the round that hands back considerably more.
Two things follow. A run that never opens the door was not cursed; it simply never reached the part of the machine where the money is stored, which on this title requires four scatters rather than the three most of its siblings ask for. And a huge multiplier symbol is not a claim about return at all. A 1,000x symbol landing on a grid with no match pays nothing whatsoever, and the 96.50% already accounts for how routinely that happens - the marketing image of the symbol does not.
So when a promotion sets "96.50% return" beside "multipliers up to 1,000x", read two unrelated facts rather than one argument. Our Starlight Princess 1000 review logs the exact moment - a 100x symbol arriving on a chain that produced nothing - because it is the most representative event in the whole genre and nobody puts it in an advertisement.
| Game | Return | Slice kept | Where that return is concentrated |
|---|---|---|---|
| Starlight Princess 1000 | 96.50% | 3.50% | Overwhelmingly inside the free spins and the chains where a multiplier meets a win |
| Big Bass Bonanza | 96.71% | 3.29% | Mostly in the feature, but with visible small line wins keeping the base game alive |
| Sugar Rush 1000 | 96.00% | 4.00% | Almost entirely in the feature, where marked spots stop resetting between spins |
| Buffalo King Megaways | 96.52% | 3.48% | Partly on the trigger itself, which pays before a free spin has run |
| Elvis Frog TRUEWAYS | 96.7% | 3.30% | Split across two features, so it arrives in smaller pieces more often |
Return figures read from each manufacturer's own catalogue page; the retained slice is arithmetic on those figures. The right-hand column is our reading of the mechanic, not a studio statement.
Notice how narrow the spread in the second column is - under a point separates the top of that list from the bottom - and how little that narrowness predicts. Those five machines do not feel remotely alike. The gap between "nearly identical percentages" and "completely different experiences" is the whole reason a return figure works as a filter and fails as a ranking.
One title, several builds
A detail that reliably surprises people: the percentage is a property of a configuration, and manufacturers issue more than one configuration of the same game. Pragmatic Play and BGaming both do it - identical reels, identical bonus round, identical artwork, different payout settings - and the operator licensing the title decides which version its lobby runs. Nothing in the interface announces the choice, and no marketing copy will mention it.
This is disclosed industry practice rather than a conspiracy, and it is precisely why a review can only ever report the manufacturer's reference figure. When we write 96.50% against Starlight Princess 1000, that is Pragmatic's catalogue entry. Whether the site you are sitting on licensed that build is a separate question with a separate answer, and it is answerable in under a minute.
Getting the figure out of the machine itself
Every legitimate slot will tell you its own return if you ask it, and the asking is trivial. The game client carries a rules or information screen, reachable from the menu inside the game window, and the percentage is printed there alongside the paytable and the terms. Where a manufacturer ships several builds, that screen names the one currently live.
Two habits make the check worth more. Do it before funding an account rather than afterwards, because the point is to inform a decision and not to explain a disappointment. And while the screen is open, read the title the software gives itself - on this family that is how you separate the 1000 build from the earlier release, from Pachi and from Super Scatter, four machines sharing one costume and four different sets of maths. A game that will not disclose a return figure anywhere is telling you something about the operator hosting it rather than about the studio that built it.
What the headline figure does not cover
The percentage on a catalogue page describes base play on the reference build. Several things players assume are folded in are quoted separately or not at all, and knowing which is which stops a lot of bad comparisons.
- A purchased feature. Buying entry usually carries its own figure, often slightly different from base play. It removes the wait, not the retained slice, and it consumes capital far faster. Ontario-registered sites take the button away entirely.
- Optional side bets. The Chance x2 wager on Elvis Frog TRUEWAYS raises your stake to raise trigger frequency. It reshapes the distribution rather than improving the return, and it is quoted apart from the base figure.
- Promotional credit. Nothing an operator gives you reaches inside the machine. Extra credit changes how many spins sit in the account, which is a bankroll effect and a real one, but it is not a return effect.
- Your stake size. The percentage is identical at twenty cents and at a hundred dollars. Stake alters how much money flows through the same maths and nothing else.
- The operator's own margin. The figure belongs to the game, not to the venue. Whatever the site keeps, and whatever an affiliate such as this one is paid, sits entirely outside it.
Who verifies that the number is true
A return figure is not a marketing claim a studio may simply assert. Builds are submitted to independent testing laboratories, which confirm the random number generator behaves as specified and that simulated play converges on the stated percentage. Regulated markets require that certification before a title may be offered at all, so a game reaching an Ontario-registered operator has been through it.
An offshore site may carry exactly the same certified build - many do - or it may not, and there is no local body you can ask. That asymmetry is one more reason the licensing question on our rules by province page is not a technicality.
A ceiling is not a return figure
The two numbers usually appear together, which encourages people to read them as one claim. They are not related. Pragmatic Play states a return for Starlight Princess 1000 and states no maximum win for it; the same is true of Buffalo King Megaways. Plenty of sites print a confident ceiling anyway, having taken it from a site that took it from another site.
We leave the cell blank and explain the blank. A figure you cannot trace back to the people who built the machine is not a specification, it is a rumour with a decimal point in it, and our review method sets out how those gaps are handled.
The cost that outruns almost any return gap
Readers spend enormous energy comparing two machines a point and a half apart, applied to turnover, accumulating slowly. Meanwhile an account settled in US dollars trims a slice off the deposit going in and off the withdrawal coming back, applied to capital, charged immediately. That one decision can cost more than the comparison being agonised over, and unlike the return figure it is entirely yours to make.
The conclusion is not that return figures are unimportant. It is that they are a tiebreaker rather than a starting point. Settle the expensive, controllable things first - a ledger denominated in Canadian dollars and a payout route that matches the deposit route - and then let the percentage decide between titles you already want to play. Our real money slots page converts all of that into spins, and the previews are gathered on the free slots page if you would rather test a machine than read about one.