Quote:
Originally Posted by Flames Draft Watcher
I dunno I guess we just have to dive into the Corsi debate full on don't we? 
It's a stat inherently biased against defensive defensemen IMO. Coaches utilize certain players in defensive situations more than offensive situations. If I put someone out against Getzlaf and Perry then there's a decent chance they'll spend that shift in their own zone defending because Getzlaf and Perry are excellent at maintaing possession. Does that mean the players we put out against them suck? Not necessarily. If I put out Russell and Engelland vs them, are Getzlaf/Perry going to get a lot of shots or shot attempts? Probably they will. But if they didn't score then the defensive defensemen did their job despite looking bad in terms of shot attempts for/against. If Russell blocks a shot that was going in, he saved a frickin goal.
Now I'm sure the fancy stats crowd will say they adjust for that completely and such but I don't think they do really. And I don't really like Corsi as a stat anyway. I don't think it's a good measure of "possession" which is the exact thing its trying to measure. For possession wouldn't it make more sense to have a stat that measures how long you cycle for in the opposition zone? That's possession IMO.
Shots are a terrible metric for analzying a game. I'm sure we've all seen games where one team outshot the other but the prime scoring chances were actually in favour of the other team. Not every shot is the same. A shot from slot is a much more prime scoring chance than a shot from the boards. A shot off a rebound is the most likely type of shot to go in.
Corsi flat out sucks. It doesn't tell us what people tell us it tells us. These stats need to get a lot better at evaluating the things that really matter before I put much stock into them at all.
Maybe I'm an idiot and the fancy stats crowd can illuminate me about how these stats are better than they appear to be. I'd be interested to hear. |
I had all the same skepticisms but as it turns out, corsi is an acceptable proxy for possession. Paging CorsiHockeyLeague, he explained all of this last year
Not all analysis is corsi either.
@DTMAboutHeart on Twitter has an algorithm called expected goals which calculates odds on teams winning games based on a combination of team shooting percentage, possession, save percentage, and so on. It's basically cheating for gambling.