Expected Goals (xG): What the Stat Really Tells You
xG is everywhere in football coverage. What it measures, why a 1–0 win can be an xG defeat, and the mistakes people make when they quote it.
The idea
Expected goals (xG) gives every shot a value between 0 and 1: the probability that a shot like it — from that position, in that situation — ends up as a goal. A tap-in from two metres might be 0.7 xG; a speculative shot from 30 metres, 0.02.
What goes into it
Models are trained on hundreds of thousands of historical shots. The most important factors are distance and angle to goal, followed by the body part used, the type of assist, and whether the shot came from open play, a set piece or a counter-attack. A penalty is usually valued at roughly 0.76–0.79 xG, depending on the model.
How to read a match total
Add up a team's shots and you get its xG for the match. If a team wins 1–0 with 0.6 xG while conceding 2.3 xG, it won — but the chances suggest it was fortunate. Over one game, finishing and goalkeeping swing results; over a season, xG is a strong signal of underlying performance.
Common mistakes
- Treating xG as a score. A team "winning on xG" did not win the match.
- Over-reading small samples. A striker 3 goals above xG after five games may simply be on a hot run.
- Comparing models. Different data providers produce different numbers for the same shot.
Used well, xG answers one question: were the chances good or bad? That is often more useful than the scoreline.
