xG, short for expected goals, is the chance that a shot is scored, given as a number between 0 and 1. A shot worth 0.30 xG goes in about three times in every ten. Add up a player's or a team's shots and you get the goals an average finisher would have scored from the same chances.
That makes xG the clearest measure of who is creating chances, whatever the scoreline says. Over a run of games it is a steadier guide to what happens next than goals are, because hot and cold finishing streaks even out.
An xG model learns from a huge record of past shots. For each new shot it asks how often shots like it went in, judged on:
Each provider builds its own model from its own data, and some also count the goalkeeper's position or the defenders in the way. That is why the same match can show slightly different xG on different sites.
Most shots are poor chances. The average Premier League shot was worth 0.12 xG in 2025/2026, and about one shot in nine was scored. A one on one with the keeper is usually worth 0.3 to 0.4, and a tap in from a yard out can be worth more than 0.9.
The average Premier League side had 1.47 xG a game in 2025/2026. Manchester City created the most, 1.92 a game, and Arsenal conceded the least, 0.88 a game, on their way to the title. Above 2.0 in one match is a lot of chances; below 1.0 is very few.
For a striker, 0.5 xG per 90 minutes is elite. Only nine Premier League players with 1,500 minutes or more reached it in 2025/2026, led by Erling Haaland on 0.70.
Updated 10 October 2026, after the last game
| Rank | Player | Club | Appearances | Goals | xG | Goals minus xG |
|---|---|---|---|---|---|---|
| 1 | Bukayo Saka | Arsenal | 6 | 3 | 3.8 | -0.8 |
| 2 | Erling Haaland | Manchester City | 5 | 5 | 3.8 | +1.2 |
| 3 | Alexander Isak | Liverpool | 5 | 4 | 3.4 | +0.6 |
| 4 | Igor Thiago | Brentford | 6 | 1 | 3.1 | -2.1 |
| 5 | Brian Brobbey | Sunderland | 5 | 3 | 2.9 | +0.1 |
| 6 | João Pedro | Chelsea | 5 | 5 | 2.7 | +2.3 |
| 7 | Bryan Mbeumo | Manchester United | 6 | 3 | 2.7 | +0.3 |
| 8 | Kevin Schade | Brentford | 6 | 3 | 2.6 | +0.4 |
| 9 | Gonzalo GarcÃa | Fulham | 6 | 1 | 2.5 | -1.5 |
| 10 | Thierno Barry | Everton | 5 | 2 | 2.5 | -0.5 |
Premier League player xG and xA, all 430 players
Scoring more than your xG means finishing better than an average player would from the same chances. Over a handful of games that is mostly luck. Over several seasons the best finishers do stay above their xG, but only by a little.
Big gaps tend to close. Antoine Semenyo scored 17 Premier League goals from 11.89 xG in 2025/2026, the widest margin above xG in the league, while Liam Delap scored once from 5.93 xG for Chelsea.
A player with plenty of xG and few goals is still getting into the right places, and the goals usually follow. A player far above their xG is more likely to slow down than to keep it up.
xGA, or expected goals against, is the xG of the shots a team concedes: the goals an average side would expect to let in from the chances it gives away. Arsenal conceded 0.88 xGA a game in 2025/2026, the fewest in the Premier League.
xGD, or xG difference, is a team's xG minus its xGA, showing whether it makes more and better chances than it gives away. Arsenal had the best in the Premier League in 2025/2026, +37.6, and won the title.
npxG, or non-penalty xG, is xG with penalties taken out, at 0.79 for each penalty taken, so a regular penalty taker does not look busier in front of goal than they are. It shows the chances a player gets from open play and set pieces.
The xG of the shots a player sets up, credited to the passer whether or not the shot goes in. Expected assists reward the chance created, not the finish. How expected assists are calculated
The points a team would expect from each game, given both sides' xG. Arsenal took 85 points from 74.9 xPts on their way to the 2025/2026 title. In Fantasy Premier League, xPts means something else: a player's expected FPL points. How xPts is calculated
After a match, xG shows which side made the better chances, whatever the score. Over a season, a team that keeps winning games it loses on xG is riding its luck: Aston Villa took 65 points from 52.4 expected points in 2025/2026.
xG per 90 minutes is the quickest way to find the forwards and midfielders who are getting chances. Goals follow xG over time, so a player with a high xG per 90 and few points is usually worth keeping. xG, xA and the rest of the game's terms are in the FPL glossary.
Bettors use xG too, and the goals betting guide shows how.
Yes, standard xG counts every penalty, each worth about 0.79, so a regular penalty taker can look busier in front of goal than open play suggests. Non-penalty xG, or npxG, takes penalties out to show the chances a player gets from everything else.
xG per 90 is a player's xG for every 90 minutes played, so starters and substitutes can be compared fairly. Erling Haaland averaged 0.70 per 90 in the Premier League in 2025/2026, the highest of anyone who played 1,500 minutes or more.
Over a run of games xG predicts future goals better than past goals do, because finishing streaks even out. Over one match it shows who made the better chances, not who deserved to win, and a team can lose a game it dominated on xG.
Each provider builds its own model from its own shot data, and some add details such as the goalkeeper's position or the defenders in the way. The same match can come out a few tenths apart on two sites, while the picture stays the same.
xG values the chance before the shot is struck, while xGOT, expected goals on target, values the finished shot, counting only efforts on target and their placement. A well placed shot can score higher on xGOT than xG, and keepers are judged on it.
Expected goals for every club and player in the Premier League, La Liga, Serie A, Bundesliga and Ligue 1.