xG and xGI

xG in FPL: over- and underperformers

Expected goals tell you what should have happened. The Butler is more interested in the players who spent a whole season disagreeing.

The largest single-season overperformance against expected goal involvement in Fantasy Premier League is 16.24 goals and assists above xGI, by Watkins in 2023/24, who returned 37 against an xGI of 20.76. The largest underperformance is 6.65 below xGI, by Yeremy in 2025/26. FPL publishes expected statistics from 2022/23 onwards, and the figures count only players with at least 900 minutes in the season.

Biggest overperformance
16.24
Watkins, 2023/24
Biggest underperformance
6.65
Yeremy, 2025/26
Latest season's top overperformer
9.93
B.Fernandes, 2025/26
Seasons with xG data
4
FPL publishes expected stats from 2022/23

Updated: Seasons: 2022/23 to 2025/26

How to read the difference

Expected goal involvement, xGI, adds a player's expected goals to his expected assists: the return his chances and his key passes were worth, before anyone actually finished them. The difference published here is his real goals plus assists minus that xGI, so it is measured in goals and assists, not in Fantasy Premier League points.

A positive difference is not automatically a warning. Penalty takers, set-piece specialists and genuinely elite finishers sit above their xGI year after year. It becomes a warning when a player with an ordinary shot profile is a long way above it, because that is the gap most likely to close.

  • Only players with at least 900 minutes in the season are counted, so a substitute with two lucky goals cannot top the list.
  • FPL has published expected statistics since 2022/23, which is why the series starts there.
  • The difference is in goals and assists. Converting it to points depends on the player's position, so the page deliberately does not.

The biggest overperformers, season by season

These are the players who finished a season well ahead of what their chances were worth. Watkins holds the record for the period, returning 37 goals and assists in 2023/24 against an xGI of 20.76. In 2025/26 the leader was B.Fernandes, on 33 against an xGI of 23.07.

Players furthest above their expected goal involvement, 2025/26
#PlayerGoalsAssistsGoals and assistsxGIAbove xGI
1B.Fernandes9243323.079.93
2Bowen9132212.439.57
3Semenyo1762314.218.79
4Wilson1091910.658.35
5João Pedro1592416.877.13
Players furthest above their expected goal involvement, 2024/25
#PlayerGoalsAssistsGoals and assistsxGIAbove xGI
1Salah29184733.6913.31
2J.Murphy8132112.038.97
3Bowen13112415.078.93
4Wood2032315.017.99
5Cunha1572214.027.98
Players furthest above their expected goal involvement, 2023/24
#PlayerGoalsAssistsGoals and assistsxGIAbove xGI
1Watkins19183720.7616.24
2Gordon11162715.4211.58
3Mateta1672312.1410.86
4Jota107176.9910.01
5Foden1982717.129.88
Players furthest above their expected goal involvement, 2022/23
#PlayerGoalsAssistsGoals and assistsxGIAbove xGI
1Kane3093924.9614.04
2Haaland3694531.6513.35
3Trossard8132112.398.61
4Saka14122617.468.54
5Martinelli1592416.137.87

The biggest underperformers, season by season

And these are the players whose chances never turned into returns. Yeremy is the extreme of the period, returning 4 in 2025/26 against an xGI of 10.65. In 2025/26 it was Yeremy, on 4 against an xGI of 10.65.

This is the more useful half of the page for a transfer. A player a long way below his xGI is getting into the right positions, which is the part that is hard to fix, and the finishing is the part that usually corrects itself.

Players furthest below their expected goal involvement, 2025/26
#PlayerGoalsAssistsGoals and assistsxGIBelow xGI
1Yeremy22410.656.65
2Brooks1349.215.21
3Gudmundsson0003.983.98
4Mateta1201215.743.74
5Enzo1051518.523.52
Players furthest below their expected goal involvement, 2024/25
#PlayerGoalsAssistsGoals and assistsxGIBelow xGI
1Archer2026.434.43
2Robertson0115.394.39
3Paquetá4158.723.72
4Lindstrøm0114.113.11
5Akanji0002.872.87
Players furthest below their expected goal involvement, 2023/24
#PlayerGoalsAssistsGoals and assistsxGIBelow xGI
1João Pedro931216.654.65
2Calvert-Lewin731014.434.43
3Beto3036.763.76
4Enzo3369.503.50
5Garner1236.373.37
Players furthest below their expected goal involvement, 2022/23
#PlayerGoalsAssistsGoals and assistsxGIBelow xGI
1Havertz71814.336.33
2Gray4159.994.99
3Mac Allister1021216.324.32
4Maupay1015.154.15
5Damsgaard0003.633.63

Questions and answers

What is xGI in FPL?

Expected goal involvement is a player's expected goals plus his expected assists: what his chances and his key passes were worth, whatever actually happened to them. Fantasy Premier League publishes it in the official game data from 2022/23 onwards.

Which FPL player overperformed his xG the most?

Watkins in 2023/24, with 37 goals and assists against an xGI of 20.76, a difference of 16.24. That is the largest single-season gap in the 4 seasons FPL has published expected statistics for.

Does overperforming xG mean a player will decline?

Not on its own. Penalty takers and elite finishers stay above their xGI season after season. The signal is a large gap from a player whose chances are ordinary, and that is the gap most likely to close in the following months.

Is the xGI difference measured in FPL points?

No. It is goals plus assists minus xGI, measured in goals and assists. What it is worth in points depends on the position, since a goal is 4 points for a forward, 5 for a midfielder and 6 for a defender, so the figure is deliberately left in its own units.

Read next

Now check your own league

Benchmarks are one thing. The Butler applies them to your own mini-league: a verdict after every gameweek, awards for the heroes and the blunders, and a season archive that never forgets. Enter your league code and see where your table really stands.

See your league

Source: official Fantasy Premier League expected statistics, which the game publishes from 2022/23 onwards, with earlier completed seasons in this dataset rebuilt from the community FPL archive maintained by vaastav, which is MIT licensed and itself derived from the same official API. Only players with at least 900 minutes in a season are included. Figures as of 3 September 2026.

Method

Biggest over- and underperformers against expected goal involvement (xGI). The official FPL API publishes expected stats from the 2022-23 season, and end of season totals from the official player data are used for each season. The difference shown is actual goals plus assists minus xGI, measured in goal plus assist units, not FPL points. Only players with at least 900 minutes in the season are included. Historical seasons are built from the community FPL archive maintained by vaastav (MIT licensed, itself derived from the official FPL API) combined with our own calculations.

These are official Fantasy Premier League figures from the game's own API, covering every registered manager rather than a sample of them, so there is no selection bias to correct for. Seasons before the current one are rebuilt from the community FPL archive maintained by vaastav, which is MIT licensed and itself derived from the same official API.

Derived from publicly available Fantasy Premier League data. FPL Butler is not affiliated with the Premier League or Fantasy Premier League. The figures may be reused with attribution to FPL Butler.