What the chart does and what it cannot do

The top 20 slots UK chart on this page is a model, and the model is the honest part. Rather than naming twenty live titles and quoting figures at them, the table sorts the market into twenty generic title archetypes, from the Book of Dead style high-volatility adventure to the Starburst style low-volatility arcade, and gives each archetype a modelled RTP band and a volatility class. Every figure in the table is labelled modelled, the assumptions sit under the table, and nothing in it claims to describe one named game.

The archetype route exists because a return figure tied to one named title is only as good as the version behind it. A lobby can print one number, a different variant can run another, and a reader has no way to check which version loads on their screen. A class-level band claims less and holds still: it says that the fishing-collector shape of game, wherever a reader meets it, sits inside a modelled range, and it leaves the exact figure to the game's own published terms.

What the chart cannot do matters just as much. It cannot tell anyone what an evening will return, it cannot rank one archetype above another as a buy, and it cannot turn a percentage into a withdrawal. The percentage is an expectation spread across an enormous volume of play, the volatility class describes how unevenly that expectation arrives, and the terms wrapped around any offer decide what a player can actually take out. The rest of the page works through those three layers in order.

Use the page in the order it is built. The table gives the classes, the chart prices the spread, the arithmetic section runs the numbers, and the calculator lets a reader swap in their own stakes. The baseline table and the licence section then price what the wrapper around any claim adds or removes. Nothing here needs a named title to work, which is the point of a chart that holds still.

How to read an RTP figure and what it does not promise

RTP stands for return to player. It is the share of everything staked that a game is built to hand back across a very long run of play. A title modelled at 95 per cent is built so that, over a long enough volume, £95 of every £100 staked returns to players in total and £5 does not. The figure is a property of the game's maths, measured over millions of spins, not a promise held for any one of them.

What it does not promise is the part readers most often miss. No session inherits the average. A hundred spins at 20p stakes £20, the long-run expectation at 95 per cent is £19.00 returned, and a single batch can return nothing at all or several times the stake. The distance between the average and the short run is not bad luck waiting to be corrected; it is the design, and the volatility class in the next section describes its size.

The bands on this page are modelled. We take the published-class range for each archetype, treat the midpoint as the class's expected return, and print both ends so the spread stays visible. The twenty modelled bands run from 93.0 per cent to 96.7 per cent, which means the modelled cost of play runs from £3.30 to £7.00 of every £100 staked. Where a figure is needed for arithmetic later, the assumption is stated in the sentence that uses it.

One reading habit is worth building: read the figure as a rate, not as a reward. A rate lets you price volume. A thousand spins at 10p stakes £100, so a one-point difference in the band is £1 of expected cost across that thousand. A reward framing, the sort that whispers a high figure is owed back, is exactly what the arithmetic does not support.

Top 20 slots UK in one table

The table below is the page's core: twenty archetypes, each with a modelled RTP band and a volatility class, every row labelled modelled. The band is the range the class is modelled to sit inside, the class describes how unevenly the return arrives, and the final column translates both into what play tends to feel like. The note under the table carries the standing assumption: these are category-typical, modelled figures, and the terms that govern any real game are the ones printed at the operator's end.

Figures are category-typical and modelled; a payment desk's own terms govern.

Two rows reward a second look. The progressive jackpot band sits lowest because a slice of every stake feeds the meter, so the base game returns less by design. The bonus-buy carrier prints a narrow, respectable band and pairs it with the sharpest swings on the page, because the whole variance of the feature can arrive at once. Neither fact makes one archetype better than another; both change what a session feels like.

Mapping a real title onto an archetype takes three questions. Does the base game pay in lines, ways or clusters? Does most of the return sit behind a feature, and can that feature be bought? Does a meter, a drop or a collector mechanic skim part of the stake before the reels pay? The answers place a title in a row, and the row's band does the rest. A title that sells its own bonus round behaves like the bonus-buy carrier whether or not it carries a famous name.

Reading the payback chart

The chart takes the midpoints of eight of the modelled bands and lines them up. The bars start at 90 per cent rather than at zero, because the honest differences between classes are small in percentage terms and the truncation lets them be seen. The caption carries the model wording, as every figure on this site does.

Read the gaps as money, not as decoration. A point of RTP is £1 on every £100 staked. The gap between the top bar and the bottom bar, 96.3 to 94.0, is £2.30 per £100, which is £23 across £1,000 of turnover and £230 across £10,000. The copper bar marks the fishing-collector archetype, and the worked example in the arithmetic section uses its 95.6 midpoint, so the chart and the sums stay consistent.

Volatility classes and the shape of a session

Volatility describes the shape of the ride, and it is a separate dial from RTP. Two archetypes can share a band and behave nothing alike: the low-volatility arcade and a high class can both model 95 per cent, while one hands the return back in small, frequent amounts and the other piles most of it behind a feature that may not land for hundreds of spins.

The classes used in the table run from low through medium-low and medium to medium-high and high. A low class returns often and swings gently, which suits short sessions and small bankrolls. A medium class trades some frequency for larger features. A high class concentrates the return hard: long quiet stretches, then a round that carries most of the session's return inside it. The class tells a reader what bankroll and what patience an archetype asks for, which is more useful at the stake-selection screen than any headline percentage.

The practical use is matching, not ranking. A reader who wants an hour of gentle play prices a low class differently from one chasing a single large round, and the same £10 bankroll survives very differently across the classes. What volatility never changes is the expected cost: the band governs how much the model expects the game to keep, and the class only governs the route the money takes.

Stake sizing is where the class earns its keep. The same £20 bankroll is a long, gentle evening at a low class and a five-minute prelude at a high one. The class does not change the expected cost per £100 staked; it changes how many decisions a bankroll buys and how wide the swings run around the expectation.

Spins, stake and expected return

Three worked examples, each with its assumption in the open. First, a plain session: a hundred spins at 20p on the fishing-collector archetype stakes £20.00. At the modelled midpoint of 95.6 per cent, the expected long-run return is £19.12, so the modelled cost of the session is 88p. Volatility decides whether an actual batch lands near that, above it or at zero; the expectation itself does not move.

Second, a wagering ladder. We take 40x, because that is what this part of the market asks most often on spin winnings. A £10 spin bonus at 40x on winnings asks for £400 of turnover before a withdrawal clears. At the same 95.6 midpoint, that turnover carries an expected return of £382.40 and an expected cost of £17.60, before any cap, expiry or game weighting narrows it further. The ladder, not the spin value, is where most of the cost of a claim sits.

Third, volume: a thousand spins at 10p stakes £100.00. At the top of the modelled range, 96.7 per cent, the expected return is £96.70 and the cost £3.30. At the bottom, 93.0 per cent, the return is £93.00 and the cost £7.00. Same stakes, same thousand spins, and the band alone moves the expected cost by £3.70. That is the whole argument for reading bands rather than headlines.

The British baseline, table by table

The second table sets the British baseline beside the offshore bracket. Rows carrying a UK licence print "UK licence" and show the category-typical shape of a British-licensed welcome: match offers or free bets, wagering in the thirties where it applies, a £10 minimum deposit in most cases, and payment rails that never include credit cards, because the British ban has bound licensees since April 2020. Rows at the offshore end print "No UK licence" and show the shape of that bracket, where crypto rails are common and credit cards still appear, because the ban binds licensees only.

Every figure in the table is a category-typical range assembled from terms operators published at the time of writing, not a quote of any single promotion, and the note under the table says so in the site's standing wording. An offshore licence claim is made in an operator's own footer, and this site treats any such claim as unverified until a register lookup records it. The licence column reports the class, and the class is the point.

Figures are category-typical ranges compiled from operator-published terms at the time of writing, not quotes of any single promotion. Rows carrying a UK licence are the British baseline shown for comparison; none of them are claimable at the offshore offers this page discusses.

What a UK licence would have bought you

Framed as what it buys the player, a UK licence is a bundle of protections with the price already paid by the operator. It buys age verification before play is allowed and identity checks under anti-money-laundering rules. It buys a place on Gamstop, the national self-exclusion scheme that every British-licensed online operator must join, with minimum terms of six months, one year or five years that cannot be cut short once they run.

It buys a dispute route. A British licensee must offer an approved alternative dispute resolution service, and IBAS adjudicates disputes between customers and licensed operators independently and at no cost. An operator outside the register sits outside that route, outside Commission enforcement and outside the scheme, so a withheld withdrawal has nowhere independent to go. A licence also buys the payment protections, among them the credit card ban, and verified games with a regulator holding enforcement powers behind them.

The tax position is part of the same settled baseline: player winnings are not subject to income tax in the UK, and the duty falls on operators instead, at 21 per cent of gaming yield and rising to 22 per cent from April 2027. None of this is an accusation aimed at any offer; it is the price list of the protections an offshore offer does not carry, printed plainly so a reader can cost the gap before claiming anything.

The terms that move the return

The band describes the game; the terms describe the wrapper, and the wrapper moves the money. Wagering on bonuses and on free-spin winnings at the offshore end typically runs 25x to 50x, most commonly 35x to 45x. Spin values typically run 1p to 25p, most commonly around 10p, and no-wager spin offers exist with lower values or tighter caps in exchange for the missing ladder.

Cashout caps on no-deposit spin offers commonly sit between £50 and £100, with expiry of 3 to 30 days, most often 7. Minimum deposits sit at £10 to £20, with low-deposit offers at £5 and, on specific rails, £1 to £3. Payout windows run from same day to 48 hours on e-wallets, 1 to 5 working days on cards and bank transfer, and often under 24 hours on crypto. Rails commonly include MuchBetter, Visa, Mastercard, Trustly, Paysafecard, bank transfer and crypto, with credit cards appearing offshore because the British ban binds licensees only.

Game weighting is the quiet term worth hunting for. Ladders are usually priced as if every pound staked counted the same, and where weighting applies, certain games count at a fraction, so the real turnover needed grows beyond the headline multiple and the expected cost grows with it. The weighting table sits in the terms, not in the lobby.

Read those figures as the second half of the chart. The archetype band prices the game's expectation; the wagering multiple, the cap, the expiry and the rails price the claim around it. A reader who costs both halves before opting in holds the full arithmetic, and the terms at the operator's end always govern over any model.