Every number on this site comes from a real calculation. This page explains exactly how — no black boxes.
The Fixture Ticker uses three independent difficulty models, not just FPL's own FDR. You can switch between them with the view controls.
The default view uses FPL's official difficulty values — the same numbers visible on the FPL website. We display these as-is because they're what managers are already familiar with.
For attackers and midfielders, this rating tells you how leaky the opponent's defence has been — combining their season-long xGC with their last 5 games, weighted toward recent form. Home advantage is factored in. A green rating means the opponent has been conceding chances freely.
For defenders and goalkeepers, this rating tells you how dangerous the opponent's attack has been — again blending season and recent xG, adjusted for home and away. A green rating means the opponent has struggled to create chances.
Buckets use percentile thresholds rather than equal intervals, so roughly the bottom 15% of fixture scores become a 1, next 20% a 2, and so on. This prevents ratings clustering in the middle.
Every projected points figure across PitchIQ — in the Transfer Planner, Player Scout, Differential Finder and Captain Picker — is produced by the same engine. It combines seven signals:
GW1 trusts FPL's EP directly — FPL's own expected points figure already accounts for the next fixture, recent form and team strength. It's the single most accurate short-term signal available.
GW2+ blends EP with points-per-game — ep_next is noisy for multi-week planning because it reflects a single fixture. Season points-per-game is more stable and corrects for one-off blanks or hauls. The 55/45 blend balances short-term signal with season reliability.
Fixture ratio scaling — rather than applying arbitrary difficulty multipliers, we scale the blended EP up or down based on how the upcoming fixture compares to the GW1 fixture in terms of difficulty and home/away advantage. Difficulty weights: 1=1.28 · 2=1.12 · 3=1.00 · 4=0.84 · 5=0.68. Home=1.06 · Away=0.94.
Availability from chance_of_playing — FPL reports a player's chance of playing next round as 0, 25, 50, 75 or 100%. We use this directly rather than a binary available/doubt distinction, which is more accurate for players with niggling injuries.
Bonus point expectation — players with consistently high BPS scores earn bonus points every week regardless of fixture difficulty. We add a flat bonus expectation (season bonus ÷ starts, capped at 1.5) to all projected figures.
The Captain Picker ranks players using a composite score across three signals — recent form, FPL's own predicted points for the next gameweek, and ICT index. Expected points is weighted most heavily because it already accounts for fixture difficulty and form. Form and ICT are added to capture recent momentum and goal threat that predicted points can sometimes lag on.
The result is a single ranked list. The higher the score, the stronger the captain recommendation for your specific gameweek.
We believe in transparency. Here is an honest account of how the signals behind PitchIQ's models performed in the 2024/25 FPL season — the last completed season before this tool was built.
The Captain Picker uses three signals: FPL Expected Points (EP), Form, and ICT Index. EP is weighted most heavily. In 2024/25 those three signals clearly identified the right captains across the season:
No model is perfect. FPL involves variance — players blank on good fixtures and haul on bad ones. PitchIQ's signals improve your probability of making the right call, but they cannot eliminate it. The captain score is a ranking tool, not a guarantee. Use it alongside your own judgement of team news and matchday context.
We do not publish fabricated backtest numbers or cherry-picked results. The 2024/25 context above reflects the season as it happened, not a simulated run of our exact algorithm against historical data. We think that's the honest approach.