Skip to content

Explainers / TermUpdated September 30, 2026

Expected goals (xG)

Expected goals, or xG, measures the danger of scoring chances. Each attempt gets an estimated probability of becoming a goal. Add those probabilities together and you have the expected-goal total.

A simple example

Imagine three attempts with estimated scoring chances of 5%, 15% and 30%. These are hypothetical probabilities.

0.05 + 0.15 + 0.30 = 0.50 expected goals.

The attempts might produce no goals, one goal or more. The 0.50 describes the combined opportunity; it does not mean half a goal went in or that a goal is now owed.

That makes xG useful when the shot count tells only part of the story. Ten attempts from poor positions can carry less danger than a few excellent chances.

What makes a chance dangerous?

Distance and angle matter, along with the type of shot, manpower situation and events leading into the attempt. A rebound, for example, can ask the goalie to make a second save before getting set.

wRink estimates these probabilities from the recorded play, without using whether that particular attempt scored. The record cannot show everything a viewer sees: a screen, a passing lane or a goalie’s exact position may be missing. Expected goals adds useful context to the play rather than replacing watching it.

Which attempts count?

Our modern skater calculation includes goals, saved shots and missed attempts during regular-season play with the net occupied. Blocked attempts, empty-net attempts, penalty shots and shootouts sit outside it. Missing the net is still an outcome of a scoring opportunity.

When comparing actual goals with xG, both must cover the same attempts. A player’s official goal total and an xG total from a smaller sample do not make a fair comparison.

What the number helps explain

For a team, chances created and allowed describe the balance of danger. Our five-on-five process estimate goes further, estimating individual contributions while accounting for teammates, opponents and game context.

Finishing asks a different question: did the shooter add goals beyond the chances he had? A player can generate lots of excellent chances and score plenty without finishing much above expectation.

The exact xG baseline also depends on the job. Goalie evaluation uses attempts that reached the net and accounts for the shooter. Those totals should not be mixed with the skater calculation as though every sample and adjustment were identical.

Full method: Expected goals