Every fixture on this page carries one prediction, published before kick-off, together with the market odds available at the moment it was published. The odds are recorded rather than live, so a pick can be judged later against the price it was actually offered at instead of one that moved afterwards.
Each fixture gets the market it fits, not the same market as every other fixture. A tight match between two similar sides and a mismatch between a title contender and a struggling side are different problems, and forcing both into a match-result call would mean publishing a weak opinion on one of them.
Fixtures, line-ups, injuries, live scores and bookmaker prices are licensed from commercial sports-data providers. On top of that feed we run our own selection procedure, which profiles each match and matches it to a market. The methodology page sets out exactly how that works, including the inputs used and when our own profile overrides the provider.
If the feed has no data for a fixture, the row stays empty and reads NO MARKET. We do not fill gaps with estimates, and where a fixture has no bookmaker price we show no percentage at all rather than a number with nothing behind it.
There are no hit rates, return-on-investment figures or profit claims anywhere on this site. Those numbers are easy to present favourably and nearly impossible for a reader to verify, and they are the most common way prediction sites mislead people. The picks and their published prices are here; anyone who wants to keep score can.
We do not accept bets, process payments or handle money in any form. This is an analysis site. 18+. Betting involves financial risk and nothing here is a guarantee of any outcome.
Every market we model has its own page, listed on the markets overview, and a fixture appears on exactly one of them — the one its shape fits. The all leagues page shows the same day grouped by competition instead, which is the better view if you follow one league closely.
Value bets collects the fixtures where our probability is materially higher than the price implies. Popular bets shows where the day's money is concentrating, which is useful mainly as a map of the crowd rather than a shortlist. Accumulator suggestions and the slip generator combine published picks into priced slips, with the arithmetic of why fewer legs is almost always the better trade.
For a specific fixture, the match insight breaks the pick down into what drove it — form, head-to-head, goals scored and conceded, and the current team news. Live scores run continuously while matches are on.
The percentage beside a pick is the model's probability for that outcome in that fixture. It is not a claim about how often our selections win, and the colour band beside it — high, medium, low — is a readability aid over the same number rather than extra information.
The odds shown are the market price at publication, before kick-off, and they are deliberately frozen. Prices move as money comes in, so recording the number lets a pick be judged later against what was genuinely available rather than against a price that drifted afterwards.
A high probability does not make a good bet on its own. A 90% shot at a price implying 95% loses money; a 45% shot at a price implying 35% makes it. That distinction is the whole of what the value bets page is for, and the methodology page explains how we arrive at the numbers behind it.