We sell the real number.
Every fixture, priced by a calibrated engine across fourteen markets — then handed to your inbox as a one-page match card. No promised winners. No hidden losses. Probabilities that mean what they say.
Cancel anytime · 18+ · Information, not betting advice
3 of 21 lines from this card · from our live ledger, unsettled
Tier labels are defined model thresholds (ANCHOR = model ≥ 85%), not promises. Probabilities are the model’s, shown exactly as published.
One page per match. Everything priced.
A match card is not a tip. It is the whole fixture, worked through properly, on a single page.
Fourteen markets, priced
Result, both teams to score, goals, corners, cards, shots — each line carries the model’s probability next to the market’s implied one, so you can see where they disagree.
Accas, done properly
Combination maths the way it should be: probabilities multiplied with correlation haircuts, and the dead leg — the line most likely to sink the whole thing — named on the card.
Context that matters
Form, head-to-head, referee tendencies, injuries — the why behind every number, in plain English, on the same page.
Straight to your inbox
Cards arrive by email on matchdays as one-page PDFs. Sign up with your email at checkout — nothing else to install. Read the numbers, make your own call — including the call to sit a match out.
When we say 85%, it happens 85% of the time.
That is the whole product: a probability you can take at face value. It is called calibration, and unlike a quoted win rate it is falsifiable — you can check it against every pick we have ever published.
| We said | It happened | Off by | Sample |
|---|---|---|---|
| 70–80% | 73.7% | −1.3 pp | 1,200+ picks |
| 80–90% | 85.5% | +0.5 pp | 1,200+ picks |
| 90–100% | 94.0% | −1.0 pp | 1,200+ picks |
What “backtest” means: we ran the engine over the season and graded every published pick against what actually happened — the full set, not the good weeks. Our live public record opened August 2026 and settles pick by pick; until it matures, this backtest is the only performance evidence we will show, and we will always label it as one. Calibration is not a profit claim — see how we’re different.
Built to be checked.
We publish everything
Every pick we publish goes into a public ledger, settled against real results — wins, losses and voids alike. Nothing is deleted and nothing is cherry-picked. If we could not settle a pick, it stays in the log and counts against our coverage.
We never sell profit
Here is the sentence no other service will print: our engine does not beat the bookmakers’ closing prices — we have measured it, and it does not. What it does is quote honest probabilities. Anyone promising you profit should be asked for their full ledger; ours is one click away.
We show the working
Model probability next to implied probability. Sample sizes on every claim. Coverage next to every win rate — because “84% winners” means nothing without knowing how many picks could actually be settled. A percentage without a denominator is marketing; ours come with both.
Read this before you subscribe
- A high strike rate is not profit. Over 0.5 goals lands roughly 94% of the time and pays pennies. Winning often and making money are different things, and we will never blur them.
- Our live record is young. The public ledger opened August 2026. Until it matures, the only performance evidence on this site is the backtest above, labelled as a backtest.
- Nothing here is a recommendation. Some lines on a card are shown precisely because they are poor value. We price the match; what you do next — including nothing — is your decision.
One price. No tiers, no upsells.
- A one-page match card for every covered fixture, on matchdays
- Fourteen markets priced per match, plus accas and bet-builders with the combination maths done properly
- Form, head-to-head, referee and injury context on every card
- Delivered by email — no apps, no channels
- Full public record — misses included — always free to check
- Cancel anytime, in two clicks
18+ only. This is an information service — we never tell you what to back.
Want to see one first?
Buy a single match card for £2.99 — the full one-page PDF for one fixture, exactly as subscribers receive it. No subscription, no account.
Honest answers.
Is this a tips service?
No. A tipster tells you what to back. We tell you how likely things are and show the working. Some lines on a card are there precisely because they are poor value — knowing what to avoid is half the point.
What you do with the numbers is entirely your decision, and “do nothing” is always one of the options.
Why don’t you promise winners?
Because nobody honestly can. Bookmaker closing prices are the sharpest public forecast in football, and our engine — measured properly, over four seasons — does not beat them. We print that on our own website because it is true.
Services quoting “78% winners” are quoting a number with no denominator: no sample size, no coverage, no log of the losses. Our promise is different and testable: the probability on the card means what it says.
What is calibration?
Take every pick where we said “80–90%”. If the engine is honest, those things should happen 80–90% of the time — and in our backtest they happened 85.5% of the time, across more than 1,200 picks. Same story in the other bands.
A calibrated probability is one you can actually use: it means the same thing on every card, every week. An uncalibrated “confidence score” is decoration.
Why should I trust you?
Don’t — check. Every pick we publish lands in a public ledger, settled against real results, with misses and unsettleable picks counted. Our headline claim is falsifiable on purpose: if the numbers stop matching reality, you will see it there before we could hide it.
How do I cancel?
Anytime, in two clicks, from the billing link in any email we send you. No phone calls, no retention scripts. If it is not useful, stop paying for it.