METHODOLOGY
How is xG calculated in football?
xG (expected goals) gives every shot a probability, between 0 and 1, of becoming a goal based on the circumstances of the shot; the xG of a team or a player in a match is the sum of the probabilities of their shots. We do not calculate it on this site: we receive it from an external data provider, match by match, and add it up.
Frequently asked questions
What are xG statistics?
They measure the quality of chances. A team that scores far fewer goals than its xG was wasteful or unlucky; one that scores far more was clinical or lucky. xG describes chances that have already happened: it is not a forecast.
What does "goals minus xG" mean?
It is the difference between the goals scored and the accumulated xG, and we show it for every player in the top scorers table. A positive value means he scored more than his chances suggested, a negative one that he scored fewer.
How do you get from xG to expected points (xPts)?
For each match, each team's goals are treated as a Poisson distribution whose mean is that team's xG in the match. From there the probabilities of a win, a draw and a defeat are obtained, and the expected points of the match are 3 times the probability of a win plus 1 times the probability of a draw. A club's expected points are the sum over the matches for which we have xG.
Does the match forecast use xG?
No. The forecasting model, a Dixon-Coles, is fitted to goals scored and does not use xG.
Limits and missing data
We show xG as we receive it and add it up: we neither recalculate nor correct it.