Quote:
Originally Posted by Bingo
It's a pretty data savvy writer compiling his own goals saved above average based on his own metrics and arriving at Markstrom ranked 5th.
Certainly that has to have more weight than someone just saying Markstrom sucks because I think he sucks.
|
I really don't want to write a long post about the perils of building a stat like that, because no one will read it anyways.
Building a composite stat like that requires all kinds of choices as to what inputs to use, at what weights, etc. What back testing has been done to determine the accuracy of the stat (None, because how could you even?). It's a compilation of other stats (most of which have issues of their own, because a lot of things are difficult to measure in hockey), chosen by the author's own opinion of import, and weighted in the same way.
To conclude that there were 3.66 potential goals in that game is ridiculous. Just say it out loud. It's blind arrogance. Then, to make it worse, it is presented with 2 decimal places! making it sound so acutely accurate! and causing readers to take it as gospel
It's pretty much garbage. And I know your reply is 'it's the best we have'. But that's the problem. It isn't - or at least we have no way of knowing whether it is or not. So the correct response is to ignore it.
As to your second point, it was literally posted along with other stats that contradict it. No one has to just go with 'I think he sucks'.