Expected Goals (xG)

A statistical model that assigns each scoring chance a probability of becoming a goal, summing to an estimate of how many goals a team should have scored.

Definition

Expected Goals is an analytics metric that rates every shot or chance by its historical probability of being scored, based on features like distance, angle, body part, and type of assist. Adding those probabilities across a match gives a team's xG, a measure of chance quality that is often a better guide to underlying performance and future results than the actual scoreline, which is noisy. Bettors use xG to spot teams over- or under-performing their chances (likely to regress), but xG values differ between model providers and say nothing about a specific finish.

Worked example

A team generates chances totalling 2.3 xG but scores only once. The single goal understates how many good chances they created, hinting they may score more in similar future games, useful context for a totals or match bet.

Why it matters

xG gives learners a way to look past scorelines to underlying performance, and analysts fold it into probability models to hunt value the raw results table hides.

Related

Note: There is no single standard xG model; providers such as Opta, Understat and StatsBomb can assign different values to the same chance.