What RTP in slots means: a long-run average, not a promise

Return to player is a figure worked out over a huge number of spins, & this guide explains how it is set, tested, monitored & shown to players.

What RTP means

The UK Gambling Commission’s player guide to return to player says percentage return to player (%RTP) “can be thought of as part of the money paid to play the machine that is then given back to the player as prizes”.

GLI defines it as “the expected percentage of wagers that a specific game will return to the player in the long run”.

As a worked example, a slot with a 96% RTP would, on average over a very large number of spins, pay back £96 for every £100 staked. The other £4 is the share the game is designed to keep.

A long-run average, not a session promise

The Commission’s guide is explicit: “The %RTP is an average achieved over a significant number of game plays and not each time the gaming machine is played.”

It says the average is generally measured over 10,000 or 100,000 games for compensated machines, which vary the chance of a prize depending on earlier results. It is measured over more games than that for random machines, whose odds stay the same from one game to the next.

The guide’s own example: if a machine displays an 85% RTP, “you should not expect to win an average of 85 pence for every £1 you stake during a playing session.”

The Commission’s guidance on live RTP monitoring makes the same point from the operator’s side. “Measuring a game with only a small amount of play will be pointless as the tolerance will be too large for meaningful results,” it says, where tolerance means the range above or below the target that still counts as normal.

How a game’s RTP is set

A slot’s maths is written into the maker’s design documents. In a 2009 peer-reviewed paper, University of Waterloo researchers Kevin Harrigan & Mike Dixon say characteristics such as “hit frequency, payback percentage, and odds of winning” are held in probability accounting reports, or PAR Sheets, “sometimes called paytable and reel strips”.

Through Ontario’s Freedom of Information & Protection of Privacy Act, the authors obtained 23 PAR Sheets from the manufacturer International Game Technology. They included seven versions of Lucky Larry’s Lobstermania & eight versions of The Phantom of the Opera.

The paper discusses “how two identical looking slot machines can have very different payback percentages”.

On how lower-RTP versions are built, iGaming Business reports that Eyal Loz of the game developer RubyPlay explains developers typically adjust the frequency of key events, such as bonus features. “In the short term, players might not notice. But over time, they experience fewer rewards, and the cost accumulates,” he says.

How RTP is tested before release

GLI says RTP “can be calculated via either a theoretical or simulated approach”, depending on the game type. It works the figure out “by evaluating or simulating every possible combination”, using “the payout table data provided by the game developers”.

Its evaluations cover the contribution to RTP from the base game, from feature games & from jackpots.

For online games regulated by the UK Gambling Commission, its monitoring guidance says games “must be tested prior to release (and after updates which may affect fairness)”.

The game’s designers calculate the theoretical RTP & the game’s volatility, which the guidance expresses as a standard deviation: a measure of how widely results swing around the average. The Commission says these figures are “reviewed as part of the required external testing”.

How RTP is monitored after launch

Testing can miss problems. The Commission’s guidance says “Remote operators are required to monitor the performance of the games they offer”, because a design or operational issue could leave a game overpaying or underpaying.

Actual RTP comes from dividing total wins by turnover, meaning total stakes. In the Commission’s example, a game designed at 91.68% takes £1,200,000 of turnover & pays £1,085,000 in wins in a month, giving an actual RTP of 90.42%.

That gap is not automatically a fault. With a volatility of 5.6, the guidance says that after 400,000 games the allowable tolerance is 1.75 above or below 91.68, so the game “could return between 89.93% and 93.43% and still be considered to be performing as expected”.

The tolerance narrows as play builds up. It is set at a 95% confidence level, so a game with no fault might still fall outside the range in roughly 1 in 20 tests, & operators “should raise alerts where a game appears to be falling outside the acceptable performance range”.

GLI offers a similar check called Field Data Analysis. It compares theoretical values with “the actual values gathered from the live environment” & reports whether each game payout is “within statistically acceptable bounds”.

Minimum RTPs & the 100% ceiling

Rules differ by place. The UK Gambling Commission’s player guide, last updated on 16 June 2021, states: “There is no minimum RTP on a gaming machine.”

In the US state of Pennsylvania, 58 Pa. Code § 461a.7 says a slot machine’s theoretical payout percentage may not be less than 85%, calculated at the lowest possible wager. It also may not equal or exceed 100%, calculated at the highest eligible wager.

The Pennsylvania code calculates that percentage over “the cycle of the game”, the finite set of all possible combinations. It also states: “The odds of any winning combination may not exceed 50 million to 1.”

Where to find a game’s RTP

Under the Commission’s remote technical standard RTS 3, last updated on 7 May 2024, information on the chances of winning “must be easily available before the customer commits to gamble”.

That information must include the house edge (or margin), the RTP percentage or the probability of winning events occurring. The guidance says it may sit in game artwork, in ‘help’ or ‘how to play’ pages, or behind links on game selection pages, menus or within individual games.

For gaming machines, the Commission’s player guide says each machine must show its category, its %RTP & whether it is compensated or random, on a label or notice, or on a game help screen.

The same game does not always carry the same figure. According to iGaming Business, “Operators can deploy different RTP versions of the same game”, typically 96%, 94% or 92%.

The publication reports that, faced with higher taxes, many UK operators appear to be quietly reducing RTP. G Games co-founder Helen Walton told it: “Standard RTPs have moved from around 96% to 94%, and 92% is now the requested maximum.” The figure that applies is therefore the one shown inside the game at the site offering it, though the article notes players “may not read paytables”.

RTP vs house edge

The Commission describes the house edge as “the percentage that a casino would expect to keep, on average, from each hand or spin, given normal patterns of play”. It is the other side of RTP: at a 96% RTP, the share kept is 4%.

RubyPlay’s Eyal Loz, speaking to iGaming Business, put the difference between 96% & 92% this way: “If a player wagers €1, and the operator takes four cents, that’s the cost. If that becomes eight cents, you’ve doubled the cost of entertainment.”

As a worked example, moving a game from 96% to 94% raises the average amount kept from £100 of stakes from £4 to £6, an increase of half.

Setting deposit limits

When a game’s RTP is below 100%, the long-run average is that players get back less than they stake. The Commission’s remote technical standard RTS 12 sets out the limit tools online operators it licenses must provide.

Customers “must be prompted to set a limit as part of the registration process or at the point at which the customer makes the first deposit or payment”. As a minimum, systems must offer deposit limits over periods including 24 hours, 7 days & 1 month.

A limit can be raised only at the customer’s request, after a cooling-off period of at least 24 hours & a further confirmation. Reductions “must be implemented immediately” unless systems or technical failures prevent it.

Where daily & weekly limits are both set, the most restrictive applies. The standard’s example is that a £10 daily & £100 weekly limit allow at most £10 a day & £70 a week.

Questions

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

Only as an average over a very large number of spins. The UK Gambling Commission says RTP is “an average achieved over a significant number of game plays and not each time the gaming machine is played”.

Can the same slot have different RTPs at different sites?

Yes. iGaming Business reports that operators can deploy different RTP versions of the same game, typically 96%, 94% or 92%, & Harrigan & Dixon found identical-looking machines with very different payback percentages.

Is there a minimum RTP for slots?

The UK Gambling Commission’s player guide says there is no minimum RTP on a gaming machine. Pennsylvania’s code sets a theoretical floor of 85% at the lowest wager & requires the figure to stay below 100% at the highest.

How do regulators know a game pays out as designed?

Under UK Gambling Commission guidance, games are externally tested before release & online operators must monitor actual RTP against the theoretical figure within a tolerance set by the game’s volatility. GLI offers similar field data checks.

What is the difference between RTP & house edge?

RTP is the share of stakes returned as prizes over the long run, & the house edge is the share the operator expects to keep, according to the UK Gambling Commission. At 96% RTP, the share kept is 4%.

Sources

• gamblingcommission.gov.uk
• gamblingcommission.gov.uk
• gamblingcommission.gov.uk
• gaminglabs.com
• cdspress.ca
• law.cornell.edu
• igamingbusiness.com
• gamblingcommission.gov.uk

Sources checked 30 October 2026. Information only. 18+.