Guides
Football probability, explained.
Short guides to the numbers on this site. Every worked example uses live Premier League data.
The Dixon–Coles model, explained
How the Dixon–Coles model turns attack and defence ratings into expected goals, scorelines and match probabilities, and how our version is set up.
Read the guide →Over/under 2.5 goals and both teams to score, explained
What over 2.5 goals and both teams to score mean, how often they happen in the Premier League, and how a score model estimates them for a match.
Read the guide →Expected goals from a score model, and how it differs from xG
What the expected goals on our match pages mean, how they come from team ratings rather than shots, and how to read them alongside shot-based xG.
Read the guide →How log loss and Brier score measure football predictions
Why accuracy alone misleads for football forecasts, how log loss and Brier score reward honest probabilities, and how our model compares with baselines.
Read the guide →Most common football scorelines and how models price them
Why 1–1, 1–0, 2–1 and 0–0 are the most frequent football results, why even a heavy favourite's top scoreline rarely exceeds 15%, and how score grids work.
Read the guide →How home advantage and scoring compare across six leagues
Compare home win rates, draws, goals per match and over 2.5 rates across the Premier League, Championship, La Liga, Bundesliga, Serie A and Ligue 1.
Read the guide →Does home advantage still matter?
Home, draw and away rates by Premier League season, from completed results.
Read the study →