@Sean:

“It’s entirely possible to play an outstanding game, and have poor stats.”

And the opposite must also be true, which is where I feel Green is.

“You were using your favored metrics to discredit another metric (+/-).”

Metrics. Plural. Meaning, more than one to prove my point. I have yet to see any from your side except for “because I said so and don’t agree with you.”

Look, I understand that this new, progressive way of thinking is not for everyone, either because they don’t want to or can’t understand them. And that’s fine. But just because you “disagree with the notion that metrics can describe, accurately, how someone played in an individual game” doesn’t mean they can’t. People once thought the world was flat and the Earth was the center of the solar system. Eventually, they had to come around. You will too. Until then, thanks and keep reading RMNB.

You were using your favored metrics to discredit another metric (+/-). While corsi and other advanced metrics have their uses, they aren’t flawless and cannot be relied upon for 100% accuracy over the short term (such as a single game). It’s simply not a large enough sample size, and the error induced by unaccounted factors will play a far too large role. We’re not talking about systematic error that can be factored out (zone start, opponent quality, etc.), but rather random error induced on an event-by-event basis.

Hockey sabermetrics have a long way to go before they’re on par with baseball sabermetrics. And both of them deal primarily with “peripheral” stats. Ones that don’t actually impact the game, but tend to do a good job of predicting outcomes. But they’re peripherals for a reason, and there are guys that outperform, and underperform, their peripheral stats on a consistent basis.

Given the additional factors in hockey stats not present in baseball stats (which are almost always occurring in a much more isolated environment than hockey), single-game peripheral stats (such as corsi) just aren’t reliable to give an accurate description of an individual player.

I also disagree with the notion “metrics simply aren’t going to give any form of accurate indication of his play because it’s only been a few games.” While that is certainly true over the long haul, i.e where Green will end up at game 82 vs now, it gives us a VERY clear view of his performance for EACH game, which was the population you introduced with your statement.

Metrics are generally useful (in context) with adequately large sample sizes. Right now, metrics simply aren’t going to give any form of accurate indication of his play because it’s only been a few games. Especially given that Schultz (his typical defensive partner) is recovering from mono and has looked somewhat sluggish. Give it a few weeks and see what his metrics look like.

At 5v5 I have him as being outchanced -2,0,-1 in the 3 games and his defense arguably becoming worse as his Corsi goes from +10 to 0 to -6.

Show me a metric that illustrates “was a stud on the defensive end last night (and has generally played solid defense in the first three games)” and I will recant.

Or mind-boggling?

Or mind-boggling? ;->