I've seen a lot of talk about how our goalies aren't really doing a good enough job while citing their respective Sv%, which aren't exactly in the top of the league. I just wanted to show everyone what happens when you use xG as the basis for goalie performance rather than just counted shots and goals.
Here are the top 30 goalies with 300+ minutes (total sample: 57 goalies) in terms of goals saved above expectation per game (all situations):
*When Natural Stat Trick lists GSAA/60 they do so only in terms of regular save %, which isn't the way Corsica did it with xG. I have here recalculated it with xG instead.
As
@Mac n Gs pointed out in some other thread, we are giving our goalies an insanely difficult task and they are doing a very good job with it.
This is the goaltending stat that has had Hank as the best goalie in the league pretty much every year between 2007-08 and 2015-16 and on aggregate has him so far ahead of his peers they might as well play different sports.
Another note is that we can clearly see how the increase in scoring isn't coming from a higher degree of dangerous shots, but what looks like a general increase in shooting percentage. The xG algorithm is trained on historical data, where scoring was lower. If the increase was merely teams generating more dangerous chances, the goalies would perform "average" according to GSAA but only 16/57 goalies have managed to do that so far this year (and we it last year as well). So Georgiev's +0.04 is actually +0.35 compared to the average GSAA/60 among these 57 goalies (-0.31) and Hank's +0.2 is more akin to a +0.55.