I’ll say what I said in r hockey where you took this from: this metric is not great because it uses Fenwick instead of Corsi. It makes our defense look worse than it is because we block shots at a high rate compared to the league. The metric also claims to measure shot quality but only does so using total xG. But here’s the thing about total xG – it’s correlated with fenwick because it goes up when shots go up!
So in the metric xG/FF, the numerator and denominator are related.
And don’t get me started on the dog shit that is FF/GP. This whole chart ONLY is a decent analysis if it’s restricted to 5v5 play. And if you restrict it to 5v5, then guess what? 1 GP is not equivalent to another GP because the number of PIM varies every game. So you’d have to use per 60 min instead of GP.
In summary, this chart would be more interesting if it were explicitly restricted to 5v5, “shot quality” was something like 70th percentile xGA, and the “shot volume” was Corsi Against per 60 minutes.
Sincerely- a grumpy data analyst
Big_Bluebird8040
so basically lots of bad shots for, lots of good shots against? great!
Samuel_Vader
I’ve said every game so far that shot volume doesn’t matter if the vast majority of your shots are directly into the goalies chest. We’re not challenging opposing goalies if they barely have to move to make saves.
3 Comments
I’ll say what I said in r hockey where you took this from: this metric is not great because it uses Fenwick instead of Corsi. It makes our defense look worse than it is because we block shots at a high rate compared to the league. The metric also claims to measure shot quality but only does so using total xG. But here’s the thing about total xG – it’s correlated with fenwick because it goes up when shots go up!
So in the metric xG/FF, the numerator and denominator are related.
And don’t get me started on the dog shit that is FF/GP. This whole chart ONLY is a decent analysis if it’s restricted to 5v5 play. And if you restrict it to 5v5, then guess what? 1 GP is not equivalent to another GP because the number of PIM varies every game. So you’d have to use per 60 min instead of GP.
In summary, this chart would be more interesting if it were explicitly restricted to 5v5, “shot quality” was something like 70th percentile xGA, and the “shot volume” was Corsi Against per 60 minutes.
Sincerely- a grumpy data analyst
so basically lots of bad shots for, lots of good shots against? great!
I’ve said every game so far that shot volume doesn’t matter if the vast majority of your shots are directly into the goalies chest. We’re not challenging opposing goalies if they barely have to move to make saves.