Football predictions guide: how they are made and how to read them
A football prediction is an estimate of the probability of each result in a match (home win, draw or away win) and of other outcomes such as goals. It does not say what will happen: it says how often each outcome would happen if the same match were played many times. This guide explains how it is calculated, how to read it and how to check whether a source is reliable, using our own data.
By the PronosLab editorial team · Methodology ·
What is a football prediction?
A serious prediction is not a hunch or a "sure thing". It splits 100% across the possible results: for example, 52% home win, 26% draw and 22% away win. That split comes from data (form, goals, fixtures, absences) and from a model that combines them.
What separates it from an opinion is that it can be checked. If a source gives 60% to many outcomes, roughly 6 in 10 of them should happen. That is why PronosLab publishes every prediction before kick-off, with its seal, and then shows whether it was right or wrong.
How a football prediction is made, step by step
This is how we work on every match. The order matters: numbers first, then context, and a final check at the end.
- Collect the match dataResults from this season and previous ones, the table, goals scored and conceded home and away, and each team's fixture list.
- Run a statistical modelA Poisson model estimates how many goals each team will score and an Elo rating measures their strength. That gives the probability of every scoreline and of the 1X2.
- Add the contextConfirmed absences and doubts, rest days and travel distance, the referee, the weather and what the press reports, always with a source.
- Compare with the marketThe market's implied probabilities are a useful benchmark. If our number moves far away from it, there has to be a specific fact behind it.
- Check before publishingAn analyst writes the preview and a verifier checks every figure against the database. If something does not match, it is not published.
- Seal it and be accountableThe prediction is sealed with a public fingerprint before kick-off and, once the match ends, it goes into the track record as a hit or a miss. Nothing is deleted.
A real example
This is the latest match from a top league that we have settled, exactly as we published it before kick-off:
Copa del Rey · Tue 10 Nov Tavernes 0–6 Valencia We gave Tavernes 7 %, the draw 17 % and Valencia 76 %. We got the 1X2 right. See the full analysis →Which data matter most in a football prediction
Not all data are worth the same. These are the factors that move a match probability the most:
- Each team's strength
- Goals scored and conceded, adjusted for the quality of the opponent and for how recent they are. Everything else builds on this.
- Home or away
- Home advantage is real, but it varies a lot from team to team and from league to league.
- Recent form
- The last 5 to 10 matches say more than the table, especially early in the season.
- Absences and doubts
- Losing a first-choice goalkeeper or the top scorer changes a match. We use official absences and press reports with a source.
- Rest and travel
- Playing again after three days or after a long trip takes its toll, especially in weeks with European games.
- What is at stake
- A derby, a knockout tie or a relegation fight changes how teams approach a match.
- Referee and weather
- They affect cards, penalties and tempo. They are small factors, but they show up in the goals.
What a prediction covers: 1X2, goals and both teams to score
A full prediction goes beyond who wins. These are the ones we publish and what each of them means:
- 1X2
- 1 is a home win, X a draw and 2 an away win. The three probabilities add up to 100%.
- Over/under 2.5 goals
- The probability that the match ends with 3 goals or more. It depends mainly on both teams' expected goals.
- Both teams to score
- The probability that both teams score at least once, whoever wins.
- Who avoids defeat
- Who avoids defeat: 1X (home or draw), 12 (no draw) and X2 (draw or away). It comes straight from the 1X2.
- Most likely score
- The exact score with the highest probability. Even the most likely one is usually below 15%.
How to read the probability in a prediction
60% is not a promise: it means that in 10 matches like this one, that result would happen about 6 times and the other 4 it would not. A favourite can lose and the number can still be right.
The draw deserves its own attention: it happens in roughly one in four matches, even though it is almost never the single most likely result. We explain it in detail, with data, in this guide: How to read a probability →
Are football predictions reliable? How to check
Be wary of anyone who does not publish their misses. A reliable source shows its whole track record, lets you check that the prediction existed before the match and measures whether its probabilities hold up. These are our numbers today:
Brier score on the same 59 matches (lower is better): PronosLab 0.536 · market 0.506 · model alone 0.556.
A hit rate alone is not enough: picking clear favourites is easy. That is why we also measure calibration and the Brier score, which reward giving the right probability, not just picking well.
Common mistakes when reading football predictions
- Judging by a single matchOne result neither proves nor disproves a probability. You need dozens of matches.
- Treating a favourite as a certaintyA 65% call misses one time in three. There is no such thing as a "sure" prediction.
- Forgetting the drawAlmost a quarter of matches end level. When the draw is above 30%, the match is wide open.
- Reading it as a fanBias towards your own club or the big names is the most common mistake. Data do not wear a shirt.
- Trusting streaks and headlinesA short run or a story without a source weighs less than it seems. Look at how many matches are behind it.
Predictions by league
Each league has its own page with the round's matches, the table, the season simulation and our hit rate in that competition:
- LaLiga predictions
- Champions League predictions
- Premier League predictions
- Serie A predictions
- Bundesliga predictions
- Ligue 1 predictions
- LaLiga 2 predictions11 matches in the next 7 days
- Copa del Rey predictions4 matches in the next 7 days
- Eredivisie predictions
- Primeira Liga predictions
- Championship predictions
- Europa League predictions
- Conference League predictions
- 2. Bundesliga predictions
Football predictions: frequently asked questions
What is a football prediction?
It is an estimate of the probability of each result in a match (home win, draw or away win) and of other outcomes such as goals or whether both teams score, calculated from data rather than opinion.
How is a football prediction calculated?
With a statistical model that estimates each team's expected goals from their results, adjusted for the match context (absences, rest, referee, weather) and compared with the market. At PronosLab a verifier also checks every figure before publishing.
What does 1X2 mean in a prediction?
1 is a home win, X a draw and 2 an away win. A 1X2 prediction splits 100% across those three options.
Are football predictions reliable?
It depends on the source. A reliable one publishes all its hits and misses, proves the prediction existed before the match and measures its calibration. No prediction can guarantee a result.
What is a good hit rate for 1X2 predictions?
In the major leagues, getting the most likely result right in around half of the matches is already a good level, because draws and upsets are common. More than the percentage, what matters is that the probabilities hold up.
Are PronosLab predictions free?
Yes. Full predictions are open every day and, when each match ends, the prediction and whether it was right are visible to everyone. Premium adds every match and the in-depth analysis.
PronosLab is not a bookmaker and does not recommend betting. If you bet, do so responsibly and only if you are over 18. Responsible gambling
