Thread: Doughty leads the league in +/-

OK I've got to start asking some questions again.........

So is the "big picture" theory that's being presented about hockey as follows?

"It's a game essentially driven by luck, therefore you should adopt behaviors that put you in a position to capitalize on luck more often than not."

Is that why Corsi and Fenwick are the standard bearers for advanced stats?

Let me qualify this by saying that i'm not a super nerdy stat guy and I definitely don't speak for the community. However, it is definitely my impression that advanced stats in the NHL is still in an immature phase. People are starting to recognize trends, such as PDO, and that is pretty useful on it's own. If you can look at the stats, and recognize that a guy is playing below the expected line, you can make reasonable assumptions that he should eventually get back to the line. In fact, some of the stats are damn good at predicting certain trends. But so far, I havent seen any that get into the "whys". We can say a guy is doing well, or doing terribly, and that according to PDO he should eventually regress back to the mean. But we can't say how long it will take to get to the mean, or why he was so far outside the line in the first place. A stat like corsi, is already being disputed as being inadequate. Perhaps it will always be a stat that requires pairing with additional stats to tell the story, perhaps some other stat will come along and do a much better job.

With that said, luck has been getting thrown around quite a bit because, in my opinion, we still don't fully understand all of the stats (or at least what is driving them). Certainly luck does play a factor in hockey. It's too chaotic not to, but I think it might also be that our lack of understanding is causing us to over state the impact that luck has.

I just want to know if there's any conclusive evidence that links something like Corsi to victories in a more meaningful way than "remedial stats".

If not... then what are we even talking about?

The stats are an accumulation of actual happenings. They are helpful in making predictions and evaluations. They don't say "if this, then this". They are not definitive. Generally speaking, high corsi = high valuation. But it's sort of a total valuation, meaning the guy is likely doing a lot of the fundamentals correctly and is likely leading to a lot of positive outcomes for him and his team. The evidence comes after the fact.

agree with helvetica.
advanced metrics are really in the beginning stages, especially in the public domain. If you had the ability to calculate player value or any concrete measurement to winning, that would be very valuable to a franchise.
I believe that franchises have their own metrics, and keep them proprietary. No sense helping the competition.
It's widely known that this is true to the NBA.

Another reason i believe that people are still foggy on the value of the advanced stats, is because... they are still stats, and as such people can misinterpret, misuse, etc to fit the situation. Take a stat that measures a player's impact on team puck possession. We see that player X has a positive effect on his team's puck possession. If we look at a team that is doing well (winning games, scoring goals, etc..), we might look to this guy and say something like, "this guy drives puck possession for his team, and since we all agree puck possession is a good thing, and this is why the team is doing well with this guy in the line up". You take a different guy, with the same positive puck possession stat, and you say, "this guy is driving puck possession for his team, but they aren't winning, so it must be a result of bad luck". Thats a generalization, but those sorts of generalizations do get made when looking at stats sometimes. People look at the stat, agree that it's a positive thing to have, and then make assumptions. But the stats don't tell you the whole story. We still have a hard time linking the "why's" with what the stats tell us. Maybe the guy drive spuck possesion, but makes poor passes that result in turn overs... who knows. So, we have to look at more stats to make better guesses.

Originally Posted by Dr. Naysay

I just want to know if there's any conclusive evidence that links something like Corsi to victories in a more meaningful way than "remedial stats".

If not... then what are we even talking about?

Sidney Crosby- 2008-2009 season - Penguins win the cup
33g 70a =103 pts (3rd overall)
Negative Corsi score I don't recall.. something like a -.21

Bump.

Originally Posted by Alosha27

As of today, Drew leads the league in +/- with a -10..
I like Doughty but he doesn't belong on the ice for 30 minutes a game..

...

So why is Doughty playing 30 minutes a game average?

10 games into the season (3-5-2): -10

Since February 11th at St. Louis (14 wins, 5 losses): +11.

29 games into the season (17-10-2): +1.

Doughty Time on Ice/game:
Kings: 1st
NHL: 5th

Doughty Shifts/game:
Kings: 1st
NHL: 3rd

From +18 to -10. Not a bad turnaround. I am a lot more worried about his minutes if he's +18 than -10.