What is xG in Soccer? Explaining a Key Betting Stat

Expected goals, commonly shortened to xG, is a soccer statistic that measures the quality of scoring chances.

The stat takes into account a number of factors and exports an xG rating based on previous attempts from the same vicinity.

Other factors include the number of defenders in the way and where the ball is placed within the goalmouth.

Shots closer to the goal receive a higher expected goals rating. Screamers from outside the box that find the top corner comparatively receive a lower expected goals rating.

Penalty kicks all receive the same expected goals rating. Those attempts on goal are worth 0.79 expected goals from 12 yards out.

Expected goals goes a step further with post-shot expected goals. Expected goals measures all attempts, whereas post-shot xG only measures shots that hit the target.

These attempts receive a rating purely based on where players place the shots on target.

Shots hit directly at the goalkeeper in the middle receive a low post-shot xG rating, while attempts hitting the top corner receive a high rating.

How to Use Expected Goals for Soccer Betting

Bettors can apply these statistics to find teams either overperforming or underperforming their metrics.

For example, a team that scores three goals off 1.5 xG is referred to as a negative regression candidate. Teams that fail to score off two expected goals could see positive regression.

The trick? Regression often happens over a larger sample size. Plus, timing the regression can often prove a tricky proposition.

Take a team like Nottingham Forest in 2024-25. The Tricky Trees finished seventh in the Premier League with a +12 goal differential against a -3 expected differential.

In the 2025-26 Premier League season, Forest regressed to a bottom-half finish.

Additionally, it’s worth noting that some teams constantly run unlucky. In 2024, Chelsea scored 64 goals off 73 expected. In 2025, they scored 57 goals off 71 xG in its first 37 matches.

Generally speaking, it’s safer to assume teams that overperform will see their luck run out at some point.

Positive regression candidates still see improvement, but at a slower rate than the overperformers.