Quote:
Originally Posted by rickroll
can you explain why you lean so much into this stat
it seems to be the best public stat, that i know about, that captures a player's value, in specific game states anyway (e.g. 5v5). i know dom's game score model thing takes other factors into consideration, like special teams, taking/drawing penalties, etc. there are certainly other war and game score models out there, if you guys know of any lmk. but i'm mostly interested in 5v5 play anyway. afaict it tries to take the shooting and goaltending luck out of things by focusing on what is
expected based on the play.
Quote:
from my understanding, these things are highly situational, linemate dependent, and not very strong indicators - ie mcdavid will be used very differently from game to game and at away games is going to always have the nightmare matchup due to last change rules
yeah ppl say this kind of thing a lot, but if that's true, then you can see how it invariably evens out because he gets the opposite matchups at home. in any case i think ppl have found that quality of competition isn't much of a factor, because it does indeed mostly even out over large samples. sure certain guys like mcdavid probably get tougher matchups on average, but it's not that big of a deal.
from what i understand quality of teammates is much more important.
e.g. playing on mcdavid's line vs playing on any other oiler line
but over a big enough sample i think this stuff washed out. so when you see that marner has 57% xg over the last 3 seasons vs nylander's 55%, it might be fair to say that the former is better. but there's other factors to consider: how they are on special teams, penalty differential, it might be complicated by who's gotten more time playing with matthews at 5v5, etc.
idk, in a nutshell i guess it seems better than what we had before; looking at corsi/fenwick and trying to read the tea leaves from that. we knew all along that the problem was this told us nothing about shot quality, and well i think expected goals speaks to that.
over a small sample it doesn't mean much. like when i do these game reports i'm not really trying to say anything, it's literally just saying what happened in that one game. over a larger sample it means more. have large a sample do you need for data to be meaningful in hockey? 10 games? 20? half a season? a full season? 3 seasons? who knows. mostly just trying to have fun here and generate some discussion.
Last edited by 72off; Today at 12:27 AM.