Top Outfield Players — Monte Carlo Simulation (correlated bonus)
Each curve shows the probability distribution of FPL points a player could score in the upcoming gameweek, estimated from 5,000 Monte Carlo simulations.
How the simulations work: Each trial samples a player's appearance and minutes, then shares one match score across goals, assists, clean sheets, and goals-conceded penalties. Goals and assists are allocated to players; saves, cards, defensive contributions, and BPS-based bonus are scored within the same trial. Expected points and threshold probabilities come from these same simulated totals, including non-appearances. The resulting 5,000 totals are smoothed into the KDE curves you see below.
How to read the plot:
Model inputs include rolling xG, xGA, fixture difficulty, home/away, and historical per-90 stats. All predictions are for a single gameweek.
Last week’s highest projected players and what they returned
MAE, all predicted players with 1+ min: GW9 1.81 · season avg 2.03 (GW1–9)
Graded on the projections published before each deadline, bonus included (301 players)
fplbench-style MAE, players owned by 1%+ incl. non-starters: GW9 1.86 · season avg 2.21
Held-out evaluation: fit on 65,556 player-match appearances (4,536 team-sides), 2020/2021 GW1 through 2025/2026 GW37. Scored on 3,083 players who appeared in 2025/2026 GW38 through 2026/2027 GW9 (200 team-sides). The holdout outcomes were never available to the evaluation fit.
Final 2026/2027 GW10 forecast: refit on all 68,639 appearances (4,736 team-sides), 2020/2021 GW1 through 2026/2027 GW9. Predictions cover 460 current players across 10 fixtures.
Same 3,083 played-player rows from 2025/2026 GW38 through 2026/2027 GW9
3,083 played-player rows from 2025/2026 GW38 through 2026/2027 GW9; the same sample used for overall FPL-points evaluation.
Expected score uses each team’s mean model goals. CS is the team probability before player-minute risk.
Monte Carlo Simulation · Tap card for details
Last week’s highest projected players and what they returned
MAE, all predicted players with 1+ min: GW9 1.81 · season avg 2.03 (GW1–9)
Graded on the projections published before each deadline, bonus included (301 players)
fplbench-style MAE, players owned by 1%+ incl. non-starters: GW9 1.86 · season avg 2.21
Held-out evaluation: fit on 65,556 player-match appearances (4,536 team-sides), 2020/2021 GW1 through 2025/2026 GW37. Scored on 3,083 players who appeared in 2025/2026 GW38 through 2026/2027 GW9 (200 team-sides). The holdout outcomes were never available to the evaluation fit.
Final 2026/2027 GW10 forecast: refit on all 68,639 appearances (4,736 team-sides), 2020/2021 GW1 through 2026/2027 GW9. Predictions cover 460 current players across 10 fixtures.
Same 3,083 played-player rows from 2025/2026 GW38 through 2026/2027 GW9
3,083 played-player rows from 2025/2026 GW38 through 2026/2027 GW9; the same sample used for overall FPL-points evaluation.
Expected score uses each team’s mean model goals. CS is the team probability before player-minute risk.