Graded in public. Every week.

Before every deadline, Volt publishes a projected score for about 175 regular starters. After the gameweek, we grade the numbers people actually saw against the points those players really scored, next to FPL's own forecast for the same players. Nothing is re-run with hindsight, nothing is edited, and bad weeks stay up.

How to read it

• Average miss: how far a projection was from the real score, in points per player (big misses count more). Lower is better. Around 3 points is normal in FPL: one game is noisy, and nobody predicts a hat-trick.
• Top-rated player: the single player each source rated highest that week, and what he scored. It is not your captain (that depends on your squad), but it shows whether the very top of the list delivered.
• Top 10: the average real score of each source's ten highest-rated players.
• FPL's forecast: FPL's own expected points for the same players, captured before the same deadline.

What changed, and when

GW6 to GW8: Volt's number was half our machine-learning model and half FPL's own forecast. From GW9 (24 October 2026) the second half is each player's season-long output with finishing luck taken out, because FPL's forecast had turned into little more than recent form. Every week is graded on the method it was published with.

Small print, honestly

A few gameweeks prove nothing either way. We'll keep publishing every week, good or bad, and the sample grows from here. The grading runs automatically each day; a gameweek appears once FPL marks it finished.

Questions: founders@fplreward.com · Not affiliated with the Premier League or Fantasy Premier League.