Originally Posted by Calgary4LIfe
And context matters. The one year blip (it was 1.5 seasons actually) was right during the rebuild.
Not having a competent goalie can also sink a team. Name a team in the NHL that did well in spite of poor goaltending. Heck, look at the Hawks this year.
At any rate, I think in reading ideas and pondering everything, what I think matters MORE than who has the most possession is what team actually does more WITH the possession they get.
The easy conclusion to draw is that possession = winning. The more possession you have, the more winning you do. I think that is only partially true, and there are exceptions which lowers the confidence interval. I think it matters more what you end up doing more often when you have possession.
Of course if you throw 100 shots in from the outside in low scoring areas, or even in high scoring areas but in which the goalie has line of sight to the puck, and is both square and ready for the shot, most of those pucks are going to be saved. The sheer number practically guarantees some squeak in.
So who outplays whom in this situation?
Team A gets 40 shots on net, and almost all has the goalie prepared for the shot.
Team B gets 30 shots on net, but they have a lot of cross-crease passes, breakaways and odd-man rushes.
Can we say that Team A REALLY outplayed Team B?
I think that is the point of this thread. Yes, Hartley's team failed in that last year, and the analytic community said: "See, we predicted this would happen!". Well, considering that the team was a rebuilding team, with a lack of experience, and didn't have stable goaltending - was it much of a prediction or did it just happen to align? Maybe a bit from both at least, but I do think a LOT had to do with the Flames playing in front of poor goaltending. The Flames actually played fantastic during that horrible year in front of Ramo, between the time when he got recalled until getting injured.
There is a tonne of 'noise' in hockey and analytics. It is evolving. CORSI isn't a garbage stat. None of those stats (with the exception of PDO IMO) are garbage stats that I am aware of. They are just not painting a nearly complete enough picture, and adding statistical analysis like Deluxe is trying to touch on here IMO is providing for a clearer and sharper focus. I am betting that in a few years time, there will be other metrics that people will be providing for additional information to raise the confidence interval and actually be more useful in both analyzing how a team is doing, as well as providing a bit more confidence in using them for predictions. With what Deluxe has in the opening post of this thread, it really seems that the work has been happening already, and it is great to see.
|