xG over the past 9 PL seasons has....

overestimated goals scored every single season.

That's a pretty big sample size.

As an example its claiming based on chances the whole of the PL should have scored 1142 goals this past season yet only scored 1038 (9.1% difference)
xG comers from an EGV (expected goal value)
Clearly if this is off as much as 10% overall (and much higher with poor teams) then clearly its not a good predictor of team strength or performance.

If you look at the PL bottom 4 teams and bottom 10 teams you will see an obvious pattern where it totally overvalues bad clubs. (every single season)

If we look at this seasons bottom 4 points.

17 Everton 36 40.71+4.71

18 Leicester 34 45.07+11.07

19 Leeds 31 40.58+9.58

20 Southampton 25 36.47+11.47

It always hugely overvalues the very worst teams
(Feel free to go back through the last 9 seasons)

Again, a points total is a much better indicator of how good a "team" is, how good the strikers are, how good the defenders and goalkeepers are.

Lionel Messi has “outperformed” his xG prediction throughout his whole career.
That's a hell of a sample size to be lucky!

Posted By: usacanary, May 28, 20:53:30