What we found across 13 games

The same name did not always mean the same displayed percentage. Our strongest documented example is Legacy of Dead: its clients showed 91.51%, 94.51% and 96.58%, a 5.07-percentage-point spread. All three figures appear in Play’n GO’s game settings documentation, making this the largest observed spread in the study with both endpoints supported by documented configurations of the same game.

Some larger numerical spreads appear elsewhere, but they need more care. Eye of Horus, for example, includes different game builds, while Starburst contains a 98.05% reading that we could verify on screen but could not match to a documented NetEnt mathematical configuration.

The table below summarises what the screenshots showed. Observed range means the lowest and highest recorded display across our sample, not necessarily a developer-certified range of otherwise identical game versions.

Differences are expressed in percentage points. Fluffy Favourites retains its three-decimal readings in the full matrix, with the spread rounded to 5.20 pp. Sweet Bonanza has no single spread because its rules give RTP ranges, not one identified active value.

Where the source evidence stops: For Big Bass Splash, Bonanza Megaways, Eye of Horus, Fishin’ Frenzy, Fluffy Favourites, Rainbow Riches and Sweet Bonanza, we verified the screenshots of the rules or help screens but have not independently reconciled each RTP tier against a complete developer or certification configuration set in this study. The table therefore describes recorded displays, not certified RTP ranges.

Why the same slot can show different RTPs

Legacy of Dead and Book of Dead: documented configurations

Legacy of Dead supplies the clearest answer to the question. Four clients displayed 91.51%, six showed 94.51% and Slots Temple displayed 96.58%. Play’n GO’s Legacy of Dead Rules and Settings (GRS1611) lists all three settings. That establishes a plausible documented explanation for the observed differences, though the screenshots do not expose the precise build identifiers of every casino deployment.

Book of Dead tells a similar story, with 91.25%, 94.25% and 96.21% displayed across the sample. These percentages are also in Play’n GO’s Rich Wilde and the Book of Dead Rules and Settings (GRS0595). The figures show why quoting a single generic RTP for a popular slot can be misleading.

To put those percentages into perspective, a model with 91.51% RTP corresponds to a theoretical cost of £84.90 per £1,000 staked, compared with £34.20 per £1,000 at 96.58%. That is about two and a half times as much at the lower setting, in theoretical terms. It is a long-run mathematical comparison, not a prediction of anyone’s losses or results in a session.

See our Legacy of Dead and Book of Dead guides for the games themselves.

Big Bass Bonanza, Gates of Olympus and Sugar Rush: UK configuration records

Big Bass Bonanza appeared at 92.05%, 94.02%, 94.50% and 96.71%. Pragmatic Play’s structured supplier data lists 96.71%, 95.67% and 94.50%. The UK certification material also documents 94.02% and 92.05%, along with a 90.63% configuration. Our casino sample showed four values: 92.05%, 94.02%, 94.50% and 96.71%. The 95.67% and 90.63% options did not appear in these checks. The 96.71% Slots Temple reading was confirmed on page three of that game’s rules.

Gates of Olympus showed 92.10%, 94.50% and 96.50%. A 2021 UK certification documents 94.50%, 95.50% and 96.50%; a later 2024 UK Remote report documents 92.10%. The 95.50% option was not observed in our sample. Sugar Rush showed 92.03%, 94.50% and 96.50%. Its supplier material and original 2022 UK certification document 94.50%, 95.50% and 96.50%, while a March 2024 UK Remote report documents 92.03%. Again, 95.50% did not appear in the observed clients. These records explain why the displayed values can differ, although they do not give us every casino client’s exact build identifier.

Big Bass Splash and Fluffy Favourites: additional variations

Big Bass Splash showed three percentages in our sample, from 92.05% to 96.71%. The 96.71% figure at Betfred, unusual beside some other Betfred readings, is supported by its own game-rules screenshot. We have not established from that screenshot alone that all the clients were running otherwise identical mathematics.

Fluffy Favourites showed 90.187%, 93.29% and 95.388% across nine casinos. We did not find a matching game in the recorded lobby checks at Betfred or Sky Vegas. The William Hill screenshot came through its Bingo interface. Those are important conditions of this comparison, not reasons to invent a value for the two missing cells.

Where a direct comparison needs more care

Eye of Horus and Fishin’ Frenzy: game builds matter

The Eye of Horus figures span 90.25% to 96.31%, an arithmetic spread of 6.06 percentage points. But the captured games are not all the same identified build. William Hill’s 90.25% client shows version 1.7.4.1, and MrQ’s 94.00% client shows 1.7.4.2_HOTFIX_2. Both are identified in the captured material as newer 1,500x maximum-win versions, yet they display different RTPs. Meanwhile, bet365’s version 1.6.2.6 and Slots Temple’s version 1.7.0.2_hotfix_4 both display 96.31%. That is a useful distinction: build numbers and displayed RTP are separate things. Several other builds remain unidentified. A 6.06-point headline therefore describes the recorded displays, not a proven spread between settings of one identical build.

For Fishin’ Frenzy, MrQ showed 94.00% in a newer version, while bet365’s 96.12% screen identifies 1.6.1.3_Hotfix_1, the version reported at Slots Temple. The overall recorded figures run from 92.00% to 96.12%, but build identity has not been established for every client.

Rainbow Riches and Bonanza Megaways: different source contexts

Rainbow Riches appeared at 92.02%, 93.99% and 95.00%. The screenshots include different delivery contexts: Paddy Power used a Playtech/Virtue Fusion client, while bet365 and Slots Temple displayed a classic Barcrest/Scientific Games game version. That establishes a platform distinction; it does not by itself explain why every RTP figure differs.

Bonanza Megaways ranged from 91.21% to 96.00% in the recorded material. One figure needs a specific qualification: Betway’s 95.98% came from a separate developer-hosted game-help page opened alongside the casino client, not the game’s embedded rules screen. That reading is neither the minimum nor the maximum, so it does not change the observed spread. It is still useful as an example of why the source of an RTP percentage matters.

Sweet Bonanza: the rules give a range, not one fixed figure

The Sweet Bonanza screenshots show one of three published minimum-to-maximum ranges: 92.04–92.10%, 95.45–95.50% or 96.49–96.60%. The full ranges appear in the casino matrix below. We have not established which value within each range applies to a particular betting mode or game state, so it would be misleading to call the upper endpoint the sole active RTP.

Starburst: a verified display, an unresolved mathematical version

Slots Temple displayed 98.05% on its Starburst RTP help screen on 25 September, with the same figure captured again on 4 October. The on-screen observation is supported; its underlying developer-certified configuration has not been reconciled with the NetEnt material available to our research. There is a smaller discrepancy elsewhere: bet365 and Paddy Power displayed 96.09%, while the current NetEnt game page gives 96.08% and an older OpenBet document gives approximately 96.10%. The available records do not establish whether this is rounding or a separate configuration. Nor do they document the full set of lower figures displayed in our sample. So 98.05% remains a verified client display, not a developer-certified maximum for Starburst.

All 13 games: the casino-by-casino RTP matrix

This is the complete comparison, with 13 game rows and 11 casino columns. Every numeric entry has a matching source screenshot from the September 2026 research archive. Some captures show different builds, and the Betway Bonanza Megaways figure comes from an external help page. A screenshot confirms what a source displayed, not that two deployments used exactly the same mathematical build.

Swipe sideways within the table to compare casinos. The game-name column stays visible where your browser supports it.

What RTP means, and what it does not

Return to player is the theoretical proportion of cumulative stakes a game is designed to return over a very large amount of play. A 94.50% RTP corresponds to a theoretical £94.50 returned per £100 staked over that long-run model. It is not a promise that depositing £100 will result in £94.50 back, and it cannot tell you what will happen in a short session.

Game providers may document more than one mathematical configuration under the same title. The configuration that appears in a specific client can also depend on the build and market. The comparison only becomes like-for-like when those details are properly established.

Different published RTP versions do not mean a game can adjust its probabilities in response to whether someone has been winning. The UK Gambling Commission’s RTS 7A and 7D prohibit adaptive behaviour and restrict changes to game probabilities and payouts while a game is available for gambling, subject to stated exceptions and notification requirements.

How to check the RTP shown by a slot

Look in the game’s rules, help, information or How To Play menu. The figure may be labelled theoretical RTP, average RTP or return to player. Record the actual game title, the figure displayed and, where the screen provides it, the game version or mode. A general developer listing may not identify the version in that particular casino client.

For UK-regulated remote games, the Gambling Commission’s RTS 3 requires accessible rules and information about the likelihood of winning before gambling. Some games express that through a theoretical RTP, house edge or probability information. A range on the rules screen should be read as a range until the applicable configuration or mode has been established.

How Slotathon carried out the research

Slotathon checked game information at 11 selected UK-facing casinos: bet365, Betfred, Betway, LeoVegas, MrQ, Paddy Power, PlayOJO, Sky Vegas, Slots Temple, Virgin Games and William Hill. The same 13 slot titles formed the research matrix. The original source captures are dated 24–25 September 2026; Starburst at Slots Temple was checked again on 4 October.

The complete matrix contains 143 game-casino combinations: 141 numerical RTP readings and two Fluffy Favourites entries where the game was not found. We checked every numerical reading against its dated source screenshot and repeated the Slots Temple Starburst check on 4 October. One Bonanza Megaways percentage came from a separate developer-hosted help page captured alongside the Betway game; Sweet Bonanza displayed minimum-to-maximum ranges rather than one identified active RTP.

The casinos were a selected research sample, not a random selection or a survey of every UK gambling site. Slots Temple was included as a specialist comparator alongside broader-market sites, not as a recommended operator. Checks for the six original benchmark games were recorded in real-money clients. Across the wider set, the source images do not consistently establish whether every session was logged in, logged out or in demo mode. Some games were opened through bingo or other help interfaces, and some screens do not identify their build. No real-money outcomes or actual returns were measured.

Limits of the findings: The figures show what the captured screens said at the time. They do not establish a casino-wide average RTP, current game availability, what an individual will receive, or the mathematical certification of an unidentified game build. Differences between builds, ranges and unresolved sources remain visible instead of being forced into one explanation. For the broader evidence process, see How We Research.

Commercial disclosure: This page contains no casino affiliate links or recommendations. Casinos are displayed alphabetically, not ranked.

Frequently asked questions

Can the same slot really display different RTPs at two UK casinos?

Yes. Several games in this study show different on-screen percentages, and developer settings documents explicitly support multiple configurations for titles such as Legacy of Dead and Book of Dead.

Can a casino change the RTP while I’m playing?

Different versions of a slot may have different documented RTPs, but UK Gambling Commission RTS 7A prohibits adaptive behaviour that changes outcome probabilities in response to play. RTS 7D restricts changes to rules, payouts and probabilities while a game is available for gambling, except where the game rules provide for them, with relevant changes brought to customers’ attention. Two casinos displaying different RTPs is not evidence of an in-session change.

Does a different displayed RTP always mean the exact same build was changed?

No. Some clients have different game builds or interfaces. The title and RTP alone are not enough to prove identical mathematics.

Does a higher displayed RTP guarantee a better outcome?

No. RTP is a theoretical long-run property, not a prediction for a deposit, a session or a series of spins.

Why do some Sweet Bonanza cells contain two percentages?

The rules screens published minimum-to-maximum RTP ranges. We show both endpoints, rather than choosing the higher one and presenting it as a single active value.

Why is Starburst 98.05% marked as unresolved?

The display was captured twice at Slots Temple, but Slotathon has not reconciled that number to independent documentation of a corresponding NetEnt mathematical version.

Why are some Fluffy Favourites cells marked n/a?

The game was not found in the recorded lobby checks at Betfred and Sky Vegas. That is an availability finding, not a 0% RTP.

The takeaway

A familiar slot name does not necessarily identify one RTP. Across the 11 casinos we checked, all 13 titles produced more than one numerical RTP figure or published RTP range in the screenshots. The strongest straightforward example remains Legacy of Dead, whose three displayed percentages match settings documented by its developer.

The exceptions matter too. Different builds, range-based rules and one unresolved Starburst display show why a comparison needs more than a list of percentages. The useful question isn’t which casino heads a ranking. It’s which game version and which source the number actually describes.