£250,000
Staked
Pick three teams. Each matchweek is a fresh table: first takes 50% of the pot, fourth takes 20%, last takes 30%. Invite your Sweepmates into group chat and private leagues, or talk with all players and see in-event polls in real time. Around it — a market for buying and selling teams, guaranteed jackpots and rollovers that carry forward until they're won, one question a day. Pari-mutuel throughout: the operator takes a fixed fee, never a position.
The league table predicts none of this — R² ≈ 0.14. And the mechanic runs on anything with a finishing order: from the Euros to the US Open, via Eurovision.
Once picks lock, the market opens on the live week: teams bought, sold and asked for while the table moves in real time, prices capped at what the team could still win. Teams are locked in fifteen minutes before that week's first kick-off — no new teams are minted after that; the market trades only what exists. Ten per cent of every trade to the house.
We recorded the first 32 weeks of the 2025–26 Premier League season. Twenty-five turned on a goal after the 80th minute. Nineteen of twenty clubs placed. Crystal Palace never did.
10% of stakes off the top, 2% to the jackpot, the rest paid back out. Then 10% of trades, 10% of transfers, £5 a switch. A fixed fee, never a book — no conflict, so we can hand players every piece of data a bookmaker never would.
Every table is built from one week's results alone, on Sportradar's live feed. The week ranks on goal difference, then goals scored, then shots on target: win 5–1 and you finish top of the week; lose 0–4 and you finish bottom — and bottom pays. Here is a single week, as it happened — but pick first.
One weekend, drawn live: the four clubs the money touched, through the real table, with the other sixteen behind. Watch what the last ten minutes do. Then pick your own three and play the whole season — it starts at Gameweek 1, in Play It.
Ten matches, Friday night to Sunday teatime — the table below ranks on the official rules: goal difference, then goals scored, then shots on target. Press play.
Eight goals landed at the 89th minute or later that weekend alone. Across the thirty-two recorded weeks, twenty-five turned on a late goal — one scored in the 80th minute or later — and one club never appeared in any of these three rows: Crystal Palace.
Every one follows the player's own three teams — nothing generic ever fires. A selection:
Teams are locked in fifteen minutes before that week's first kick-off — no new teams are minted after that; supply is fixed, and the market only redistributes it. Ten per cent of each trade to the house. The market and the live table sit side by side: you watch the league move in real time with a price on everything you hold.
A room holds up to twenty-five players — a pub, an office, a group chat moved in. Everyone's names, picks and standing for the week are on the table, and you chat with your Sweepmates all season: before the picks, through the matches, after the money. On matchdays a bigger chat opens across the whole game.
The retention isn't a notification. It's that your absence is visible to people you know — an empty seat gets dealt with by the room, not by a win-back email.
A preview before the round, a review after it. And because the operator holds no position, the Show can say everything — form, volatility, near misses, who the numbers like this week. A bookmaker's tipster is paid against you. Ours has nothing to hide and no reason to hide it.
Under eight words, yes or no, answered before the tally shows — the world's split, then your room's. “Would you have stolen the Jules Rimet?” “Would you handball on the line in the last minute?”
Questions are tuned to divide — forty to seventy per cent yes, or the reveal doesn't work. Each tally splits by region and the outlier is a headline: a dozen local stories a day, generated rather than bought.
In matches, the same machine fires on penalties and VAR calls. Thousands of verdicts on the referee, one split, screenshotted out of the app with the brand attached.
This is the PR strategy. The Jules Rimet question is a press release with a tally attached; the regional split is the local angle; the daily cadence is the editorial calendar. Players submit questions of their own, and every answer is data nobody else holds.
We recorded the first 32 weeks of the 2025–26 Premier League season — a deliberate sample: 96 paying positions, decided one week at a time. The Premier League crowned one champion; this game crowned fifteen different weekly winners, handed fourth to fifteen different clubs and last to fourteen. If league position decided it, the sharps would strip the pool and the casuals would leave. R² ≈ 0.14 — it doesn't.
The design rule is maximum drift from the Premier League's actual ranking. The further a week's payees sit from the league table, the better the game is working. Fourteen different clubs took last place across the sample — five of them from the league's top eight.
From 32 recorded weeks of the 2025–26 Premier League season. The full week-by-week record, with volatility and form indices, comes to the meeting.
The Near Miss Index — 32 weeks, every club. 2nd is one place from 50%, 3rd and 5th one place from 20%, 19th one place from 30%. This is the money that didn't pay.
| # | Club | 2nds | 3rds / 5ths | 19ths | Total | PL |
|---|---|---|---|---|---|---|
| 1 | Forest | 3 | 4 | 4 | 11 | 16 |
| 2 | Villa | 2 | 4 | 3 | 9 | 4 |
| 3 | Chelsea | 5 | 2 | 2 | 9 | 6 |
| 4 | Brentford | 2 | 4 | 3 | 9 | 7 |
| 5 | Palace | 2 | 3 | 4 | 9 | 14 |
| 6 | Liverpool | 1 | 6 | — | 7 | 5 |
| 7 | West Ham | 2 | 5 | — | 7 | 17 |
| 8 | Arsenal | 3 | 3 | — | 6 | 1 |
| 9 | Man City | 2 | 4 | — | 6 | 2 |
| 10 | Man Utd | 1 | 4 | 1 | 6 | 3 |
| 11 | Everton | 1 | 5 | — | 6 | 8 |
| 12 | Sunderland | 1 | 3 | 2 | 6 | 11 |
| 13 | Fulham | 3 | 2 | — | 5 | 9 |
| 14 | Newcastle | 3 | 2 | — | 5 | 12 |
| 15 | Burnley | 1 | 2 | 2 | 5 | 19 |
| 16 | Wolves | 1 | 1 | 3 | 5 | 20 |
| 17 | Brighton | 4 | — | — | 4 | 9 |
| 18 | Bournemouth | 4 | — | — | 4 | 13 |
| 19 | Spurs | 2 | 2 | — | 4 | 17 |
| 20 | Leeds | 1 | — | — | 1 | 15 |
Forest came one place short eleven times — more than any club. And Crystal Palace: nine near misses, never a pay-day. Every one of them is a returning player. PL = league position over the same 32 weeks.
The asymmetry. A near miss costs the pool nothing — the pot still pays exactly three. But the player one place short leaves with evidence their read was right, and evidence enters again next week. At a bookmaker's, a near miss is a dead loss. Here it's the cheapest asset in the game: zero cost to the operator, and it returns the one thing marketing budgets can't reliably buy — the same player, back next week.
The Clean Sweep Pot. Hold first, fourth and last in one week and the rolling jackpot is yours. One player has ever done it. Two pence in every pound staked feeds it — one rule, printed beside the number.
Plus 10% of every trade · 10% of every transfer · £5 a switch, available up to fifteen minutes before first kick-off. Pari-mutuel throughout — winners share one pool, every line above is a fee for playing, and none of it depends on a result. That is what a pool licence means.
Friday 15 August 2025 — the opening fixtures, exactly as they were. Pick three teams and play the recorded season one matchweek at a time. First takes 50% of the pot, fourth 20%, last 30%, every week. Change your three whenever you like.
Pool betting is its own category at the Gambling Commission — the pools and the Tote, not bookmaking. Licence 065457-R-340334-001 is held, its obligations met and running. Following means clearing all of it, and the Commission does not hurry.
And the Treasury just made the moat wider. From April 2026, remote gaming duty rose from 21% to 40%. From April 2027, a new 25% duty lands on remote betting. Pool betting was explicitly excluded from the overhaul — its duty stands where it stood. The Budget left two regimes alone: horse racing and pool betting. This licence occupies one of them. For an operator holding liquidity and a 40% tax bill, that is not a hedge. It is the open door.
The engine that ran the league season is running a World Cup game now — free, three teams a week, a place earning £50 credit towards a £1,000 pot at the quarter-finals. And the mechanic prices any finishing order, not just football's: from the Euros to the US Open, via Eurovision.
The gap: Super 6 is once a weekend, no money, no conversation. The betting apps have money and no company — need the customer to lose, and now pay 40% on gaming with 25% coming on betting. In between stands a licensed pool with rooms, a market and a daily ritual, in the one duty regime the Budget didn't touch. Nothing else is standing there.
A sponsorship is a logo on someone else's programme. This is a routine that belongs to the partner: picks Sunday, the market midweek, the question each morning, the table Saturday.
Each pub, branch or show runs a room of twenty-five, filled by invitation, keeping its own attendance.
Regional splits, near misses, a week decided at the 89th minute, a club that never placed — stories on a schedule.
No stake touched, no licence held, no book carried. A share of fixed fees, generated by the partner's own rooms.
A partner risks no capital, holds no licence, carries no book and never needs a player to lose — the worst case is a season of content that didn't grow. The best case compounds: rooms in every pub, branch or show, fifty-two weeks a year, each new competition running on the same engine, and a share of every fixed fee those rooms generate. For the player the same shape holds — the game is free until the Last 8, stakes start at £10, and the Clean Sweep Pot grows every week until somebody takes it. Small, known downside. Open upside. That is the whole trade.