—
From xG per 90.
At least one.
At least one foul won.
At least one; top picks run a bit high.
Yellow or red, from his card rate; in testing no better than a flat guess — referees matter a lot.
From xA per 90 — in testing no better than a flat guess.
Expected total; in testing no better than the league average.
How it works: each player's season rate per 90 (shrunk toward his position's average), times his expected minutes (average over his team's last 3 league games, times FPL's chance of playing), adjusted for what the opponent allows and for home/away. Tested on gameweeks 4–7 using only the data available before each game. For interest, not betting advice.
Predictions are frozen at kickoff, then checked against what happened.
Each cell is the predicted chance, then ✓ if it happened. Cards: of the picks given 50% or more, how many came off, and how many events the model expected vs how many happened, across every scored game. Results marked provisional are refreshed once the source finalises player stats.
Filled dot = home game, hollow = away. Dashed line = 50%.
Average team distance per game in km (all players who played, added up). Follows the All / Home / Away switch.
Per-game averages. Fouls = committed by the team, Fouled = fouls won. Tap a header to sort, a row to make it Team A.
Every league game for Team A, newest first.
Season totals in the Premier League. Fouls = committed, Fouled = fouls won. Km/90 needs 90+ minutes. Tap a header to sort.