Our Rating Method: How an Offer Is Priced and a Room Is Simulated

What a bonus is worth once its wagering is paid

A deposit bonus is worth whatever survives its wagering, and one line settles that: B × (1 − W × (1 − RTP)), where B is the bonus, W the times it must be turned over and RTP the long-run return of the game clearing it. At 96% the house keeps 4p in the pound, so every extra 1x of wagering costs 4% of the bonus.

£75 × (1 − 38 × 0.04) = −£39

Take £75 at 38x, the middle of the range an offshore welcome package asks for. Clearing it pushes £2,850 through the games at an expected cost of £114, so the bonus falls £39 short of covering it. At the 10x ceiling British licensees have been held to since 19 January 2026 that same £75 is worth £45. The multiple, not the headline, is the story.

That value assumes the turnover gets finished, and in simulation most sessions do not, because the balance runs out first. Packages are therefore ordered by which is cheapest to clear, not by which pays you.

Why no bingo room shows you a return

A slot has to show its return before you spin. A room does not, and the reason sits in the Commission’s technical standards: where the chance of winning depends on how many people turn up, “a description of the way in which prizes are allocated will be sufficient” in place of a return-to-player percentage. The carve-out excuses the percentage, not the explanation, so a lobby offering neither is the one to walk past.

With no room publishing a payback, we use the sector’s. Across remote bingo in Britain, 83.25% of ticket money came back as prizes in the year to March 2026, leaving operators 16.75%; a decade earlier the players’ share was 91.67%. Both are from the Gambling Commission’s industry statistics, and they are the only sourced answer to what British bingo pays. Where a simulated room needs a payback setting, that setting is printed and labelled an assumption.

How a room gets simulated, ball by ball

Bingo figures here come from dealing tickets and drawing balls rather than a tidy closed-form guess. A 90-ball ticket is fifteen numbers on a nine-by-three grid, five to a row, one to three from a column’s own decade band, dealt off strips of six as a British room sells them. Ball order is random, one line, two lines and the full house are all live on the same ticket, and a prize splits equally between every ticket completing it on the same call.

Put 220 players in a room, give each three tickets at 5p, and set the pool at the sector’s 83.25% as the stated assumption: 660 tickets sell, your own stake is 15p a game and the pool comes to £27.47. Across 200,000 simulated games this player wins something in 1.46% of them, about one game in 68. The average return is 12.49p, exactly 83.25% of the stake, because a symmetric room hands back its payback share. The median is nothing, 98.54% of games pay nothing, and while the room’s full house typically falls on call 60, this player’s best ticket waits until call 82.

The gap between that average and that median is why the draw gets run at all. One figure describes a room, the other describes your evening.

The five model slots behind every spin figure

Spin and bonus results come off five model slots of our own, with pay tables built so each hits its target return exactly. Where a page discusses a published game, the model is built to that game’s published return and maximum win, and says the real pay table is not public.

Every result quotes a spread as well as an average, because an average alone is the most misleading number a gambling page can print. Thirty spins at 10p on the Wild model, run 250,000 times, average £2.85; the middle run returns £1.42, and only 10.4% of runs reach £5.

Four things these pages will not say

• Trials that never happened. Nobody here opened an account to see what landed in it. A modelled figure says it is modelled; a published one is cited.
• Brand claims the register will not carry. It lists domain names against licence holders and nothing more, so we never say who built a site or whose skin it is unless the register shows it.
• Bonus codes printed as current. Codes are dated, often limited to selected players and pulled without notice, and a code alters no term anyway. Read the operator’s rules page and its revision date instead.
• A figure that contradicts the terms. One brand’s blog called its welcome £50 in bingo tickets while its dated rules page gave thirty tickets worth up to £30. The rules page wins, every time.

Anything that would help somebody get round Gamstop Online, a blocking app or a bank gambling block is out as well.

Where a brand’s facts are checked

Two sources settle a brand fact. The Gambling Commission’s public register gives the licence holder, its account number and the domains hanging off it, which is how a brand’s identity is checked, not assumed. On 10 October 2026 it put Gamesys Operations Limited under account 38905, with 13 domains listed and 3 active, and Jumpman Gaming Limited under account 39175, one Alderney office with 234 domains on the register and 19 active. Counts like those move, so the date travels with the figure.

The second source is the operator’s own dated rules page, which governs anything to do with a promotion. Rule dates come straight from the Commission: the 10x wagering ceiling and the condition stopping one incentive from spanning more than one product both took effect on 19 January 2026.

If a figure here looks wrong, send the page and the number through get in touch. It gets held against the published source, and where we had it wrong the page is amended and redated.