Everything we write, newest first: previews of the matches ahead and reviews of the ones just played, each built from the model's own probabilities and expected goals, and longer pieces on what the numbers do and do not say. Kept current rather than written once: title races and relegation odds, and the model's full track record. How the probabilities are produced, including where the model is known to be wrong, is in the methodology.
Long reads written once from our own data: what the numbers show across whole leagues and seasons, and where our model gets it wrong.
Home sides outscored their visitors in every one of the 51 leagues we measured: they scored 23% more goals. How much more ranges from 45% in Brazil's Serie A to 3% in Qatar. But a single season is a poor guide to which league is which.
More →Our model has a prediction for 402 matches this weekend (31 October - 1 November), across 71 competitions. It is surest about PSV Eindhoven: 91% to beat Willem II, most likely 4-0. 42 matches are within three points either way, games where the draw is a real outcome.
More →The injury and suspension report covers 106 of the 402 matches on our board this weekend (31 October - 1 November). Across them, 645 players are ruled out and 218 are doubtful. Burnley have the longest list: 11 players out for their match against Portsmouth.
More →Fiorentina 2–2 Atalanta finished with a draw. The model got this one wrong. It leaned to Atalanta at 42%, and rated the actual result at 26%.
More →Atletico Nacional 2–0 Santa Fe finished with Atletico Nacional winning. The model had the right result: it gave that outcome 54%, though it expected 1-0 rather than 2–0.
More →Gremio 2–0 Atletico Paranaense finished with Gremio winning. The model had the right result: it gave that outcome 38%, though it expected 1-0 rather than 2–0.
More →Our model leans to Athletic Club here, but not by much: 42% (fair odds 2.38) against 25% for the draw. Published scoreline: 2-1, the likeliest exact score within that result.
More →Juventus host Napoli in a game our model cannot separate: 32% for the home side, 30% for the draw and 38% for the visitors.
More →Our model leans to Arsenal here, but not by much: 45% (fair odds 2.22) against 28% for the draw. Published scoreline: 0-1, the likeliest exact score within that result.
More →Monza 1–4 Napoli finished with Napoli winning. The model had the right result: it gave that outcome 59%, though it expected 1-2 rather than 1–4.
More →Bahia 3–2 Sao Paulo finished with Bahia winning. The model had the right result: it gave that outcome 49%, though it expected 1-0 rather than 3–2.
More →Frosinone 5–1 Lecce finished with Frosinone winning. The model had the right result: it gave that outcome 61%, though it expected 2-1 rather than 5–1.
More →Remo 2–1 Botafogo finished with Remo winning. The model had the right result: it gave that outcome 42%, though it expected 1-0 rather than 2–1.
More →America de Cali go into this one as a clear favourite by our numbers: 61%, which is fair odds of 1.64. Our published scoreline is 2-0 — the likeliest exact score within that result rather than across the whole match — and the model gives Fortaleza FC 15% to win outright.
More →Our model leans to Flamengo here, but not by much: 48% (fair odds 2.08) against 27% for the draw. Published scoreline: 1-2, the likeliest exact score within that result.
More →Our model leans to Inter here, but not by much: 48% (fair odds 2.08) against 25% for the draw. Published scoreline: 1-2, the likeliest exact score within that result.
More →Chelsea host Manchester United in a game our model cannot separate: 37% for the home side, 23% for the draw and 40% for the visitors.
More →Our model leans to Manchester City here, but not by much: 49% (fair odds 2.04) against 24% for the draw. Published scoreline: 2-1, the likeliest exact score within that result.
More →Torino 1–3 Como finished with Como winning. The model had the right result: it gave that outcome 52%, though it expected 0-1 rather than 1–3.
More →AC Milan 2–4 Bologna finished with Bologna winning. The model got this one wrong. It leaned to AC Milan at 46%, and rated the actual result at 26%.
More →