Football predictions today, with the probability on every pick

How our football predictions are made

Each team gets an attack rating and a defence rating, updated every day from recent results, with newer matches counting more. The model adds home advantage (none at neutral venues) and corrects for low-scoring results such as 0-0 and 1-1, which simple models get wrong. It then calculates the probability of every scoreline from 0-0 to 10-10. Adding those up gives the home, draw and away percentages and the Over/Under goals chances you see on each match. The same scoreline grid gives our both-teams-to-score and correct score probabilities. We publish a prediction only when both teams have at least three past results, so thin data never produces a guess dressed up as a number.

Read the full methodology

Are our football predictions accurate?

We grade every prediction after full time and never remove the misses. Over the last 30 days, 72% of picks we rated 60% or higher were correct, from 734 graded matches. We also track calibration (do our 70% picks win about 70% of the time?) and compare our numbers with bookmaker closing prices on the same games.

See the full accuracy record

How to use free football predictions: a practical guide

By Mike Jeremy, Senior Sports Betting Editor

Free football tips are everywhere, and most of them are a team name with no reasoning and no record. Ours are different in one way that matters: every prediction is a probability from a statistical model, published before kick-off and graded in public afterwards. This guide explains how to read those numbers and how to use them without fooling yourself.

1. Read the percentage, not just the pick

The Pick column shows the single most likely outcome for each match, but the percentage beside it is the real information. A 55% pick and an 85% pick are both "the pick", yet the first fails almost as often as it lands. Our model is built to be calibrated: across many matches, things we rate at 70% should happen about 70% of the time. That also means a 70% pick fails three times in ten, and that is the model working, not failing.

To turn a percentage into odds, divide 100 by it. 50% is fair odds of 2.00, 70% is 1.43 and 85% is 1.18. We call these fair odds because they contain no bookmaker margin.

2. Pick the market that fits the match

Each market answers a different question, and the same match can be uncertain in one and clear in another. A derby between two even sides may be a coin toss for the result but a strong Over 1.5 goals call.

• Home win and away win: the 90-minute result. Clear favourites often sit at 60–80%.
• Win & BTTS, BTTS & no draw and Win & Over 2.5: combinations summed from our full scoreline grid, so they always agree with the single-market numbers.
• Draw predictions: the matches most likely to end level. Draws are the hardest result to call, so even the top of the list is usually around 30-35%.
• Double chance: two of the three results in one pick (home or draw, draw or away, either team). It lands far more often, so the fair odds are short.
• Over 1.5 goals, Over 2.5 and Over 3.5: total goals by both teams. Over 1.5 is the most common high-probability market.
• Under 2.5 and Under 3.5 goals: low-scoring matches, often tight games or strong defences.
• Both teams to score: each side scores at least once, whatever the result.
• Correct score: the exact result. Even the most likely score usually has only a 10–14% chance, so treat it as a guide to plausible results, not a strong pick.

3. Bankers and accumulators: the maths of adding legs

A banker is the pick you trust most. Our banker tips are simply the day's strongest picks at 78% or more, with the real hit rate of past picks at that level. The danger starts when you combine them. In an accumulator every leg must land, so the chances multiply: two legs at 85% land together about 72% of the time, five legs at 85% only about 44%, and ten legs at 80% about 11%. Our accumulator tips show that combined chance on every ticket so you can see the risk before you take it.

4. Compare our number with the bookmaker's price

A bookmaker's odds imply a probability too: 1.50 implies 67%, 2.00 implies 50%. If our model rates an outcome at 70% and the price implies 62%, the price is longer than our fair odds. If the price implies more than our number, it is shorter, and the bet is worse value than it looks. Bookmakers build in a margin, so most prices are a little short of fair. Our value bet calculator does this comparison for any match and price, and the odds converter handles fractional and American odds.

5. Plan ahead

Predictions are published every morning for the next week. See tomorrow's predictions, this weekend's predictions or a single league. Probabilities can move until kick-off as new results come in, so check again on the day.

6. Mistakes to avoid

• Treating a high percentage as a certainty. 85% still fails about one time in seven.
• Adding legs to chase bigger odds. Each leg cuts the combined chance.
• Ignoring team news. The model does not see late injuries, rotation or weather; you can.
• Judging any tipster, including us, on a few days. Look at hundreds of graded picks and at calibration, which is why our accuracy record keeps every miss.

New to betting terms? Our football betting guides explain each market, odds formats and bankroll basics in plain English.

Using football predictions responsibly

A probability is not a promise. Even a 75% chance fails one time in four, and no model sees late team news, injuries or weather. Treat these estimates as one input, set a budget, and never chase losses. 18+ only.

Responsible gambling · BeGambleAware.org · GamStop