Quote:
Originally Posted by Elrazor
That's really the same as saying it's ridiculous to decide the PL title over 38 games as the sample size is too small. Casual observation is usually enough to tell us the best team wins the title, and the worst team finishes bottom.
I'd say there are enough data points over a season to have a high degree of confidence in the results of xG models - the main area of contention is whether the method is valid, not the accuracy of the result it produces.
The post I quoted referred to xG being used to predict, not assess.
They are two very different things.
When used as a tool to assess performance, xG is valid but as I said, the transfer market makes it ridiculous as a prediction tool.
You, of all people, should know nothing is black or white.