A research engine for betting exchange markets.
Find an edge. Prove it holds.
Watch it live.
A hundred and forty-two million recorded price movements, eight seasons deep, across twenty-three leagues and eighteen markets. We look for the moments a market moves before it should — then test whether it survives on seasons the search never touched. Built on exchange data, where the same patterns run through every book.
What the market does in the hour before kick-off
every match-odds runner we hold · averaged · by starting price
130,750 runners, each indexed to its own price an hour before kick-off, then averaged minute by minute. Short favourites come in. The longer the price, the harder it drifts.
Measured to a T-Statistic of 21 across every runner we hold — and still not a strategy. That is what the tools are for.
Accurate as of 26 August 2026. Price movements from July 2019, when the data plan begins; results run a season further back, to August 2018.
What you get
A subscription to a working research engine and to what it produces. Not tips and not opinion — measured findings, the statistics behind them, and the tools that found them. We take no wagers, hold no stakes and place no bets.
What the engine found, before the off
The market, the runner, the price on record and the minute it was measured to — published before kick-off, so nothing can be written up with hindsight. What you do with it is your own business entirely.
The live record
Every finding's full history from the day it was published — winning runs, losing runs, and the closing-line figure that says whether the edge is still there.
The statistics behind every one
Sample size, t-statistic and p-value. Which seasons were held back from the search, and how far the figure fell when it met them. Published alongside, never on request.
The research tools
The same price grids, timing maps and bank models the findings were built with. Search eight seasons yourself, hold your own data back, test your own ideas — rather than only taking ours.
Statometrics is a data and research service. We do not accept wagers, hold stakes, place bets or operate any market. Nothing published here is advice or a recommendation.
The tools you get access to
Everything here runs against the full database. These are the tools the signals were found with, not a cut-down version of them.
Timing map
Every combination of when to open a trade and when to close it, scored across every fixture at once. Spot the region where a move actually happens instead of guessing minutes.
Price windows
Any fixture, price by price. The real opening and closing prices, and any window you care to set between them.
Held-back testing
Split the seasons yourself. Search on one half, and the other stays covered until you ask for it — so a result can't quietly be fitted to the data that judges it.
Bank and exposure
Run a starting fund through any strategy at your own stake. Worst drawdown, most at risk in a day, and how the fixtures actually land across a week.
Most services show you the winners
We show you when an edge dies. Any strategy has losing runs; the hard part is telling a bad month from a dead signal. Get that wrong and you either quit something that was fine, or keep feeding something that stopped working months ago.
Tested on data it has never seen
Found on early seasons, confirmed on later ones held back. Anything that only works on the years it was found in gets thrown away.
Judged on the closing price
Profit is noisy and slow to tell you anything. Beating the price the market settled at is the earliest honest sign a signal knows something.
Costs counted before the claim
Commission and the gap between the two prices come off first. Plenty of strategies look good until you take those off, and then don't.
Seventy-five seats, and the real reason why
The edges we publish are smaller than one price increment. That sounds like a technicality; it's the whole constraint.
A market moves in steps. On a price around 2.40 the smallest possible step is roughly 0.8% — and the edges worth publishing sit under that. It means the price only has to move a single increment against you for the advantage to be gone several times over.
So if seventy-five people take the same selection at the same minute on a thin market, they move it themselves. The first few get the price we published. Everyone after gets something worse, and past a certain number, worse than no signal at all.
That number is what the cap is. Not a marketing device, not urgency — the point above which the product stops working for the people who bought it. We'd rather turn away members than quietly sell the same seat twice.
Founding twenty
Open nowLocked at that price for as long as you stay. These twenty are backing a record that is still being built, and they should be paid for taking that risk.
JoinThe remaining fifty-five
Not yetPriced once there's a full year of published results to judge it against. Charging properly for something unproven would be the same trick this whole site argues against.
Why the exchange, and nothing else
Traditional bookmakers close or limit accounts that win. On an exchange you bet against other people rather than the house — so winning isn't a reason to be shown the door. Betfair has the liquidity, but the same patterns run through every exchange, and the research travels.
Twenty seats at £40
Locked at that price for as long as you stay. The remaining fifty-five are priced once there is a published year of results. After seventy-five there are no more.
JoinCancel whenever you like. Your price does not move while you are a member.