View Single Post
02-21-2013, 04:46 AM
garret9's Avatar
Join Date: Mar 2012
Location: Vancouver
Posts: 13,404
vCash: 500
Originally Posted by Inflict View Post
Well, there are two problems with comparing stats to last year. One, is that this is a lockout year with a shortened season (lots of different variables compared to a normal season). The second, is that we are not even halfway done with the season, and this is only a small sample size.

Other things to consider is that this is a new team and a new coaching system for Olli, so you would expect some growing pains. However, he has still been consistent in his lack of production even though his previous line mates (Kane & Wheeler) still found ways to score when they were on the ice together. So in my mind, this still does not excuse him for his lack of production. There is only so much you can blame on the team and on luck. Hitting the post may count as a quality scoring chance, but in reality, this doesn't help his production as it is a matter of accuracy rather than luck. All in all, I think it has to do with his lack of chemistry with his line mates more than anything.

Finally, I would like to ask why you think +/- is a bad stat to compare defense? I agree that +/- is a team stat more than anything, but how does that excuse Olli, or any other player for that matter, from failing to produce or defend 5 on 5 hockey?
Yes, the sample size is small, but the differences are quite dramatic in positive fashion, and in all levels, indicating non-randomness (also, don't want to get into it but Corsi/Fenwick substantially increases the sample size where 10-15 games does give a good indication). Albeit, due to sample size I would not make any assumptions yet to the level of difference.

Yes I would expect growing pains due to system. The linemates producing thing I've answered below, but I would like to mention it's been noticed by some that Noel's system seems to reduce production in his centres. I have yet to look really deep into it, but it does seem that the multiposition guys (Antropov, Burmistrov, Wellwood, Little) production rates tend to jump when they are on wing.

Now to answer your question.

There is validity to plus/minus but it is a terrible stat used alone without context and without looking into it deeper.

For usage/context: is it better being a slightly negative +/- when all you do is face EStaal, Crosby, Giroux, etc. or being a slightly positive +/- when you're mostly facing 4th line pylons game in and game out?

Another problem is it is very strongly affected by linemates, here's two examples from us.
2011-12 Byfuglien with:
PartnerTOI CF% GF% +/-
Enstrom 724 0.558 0.522 +3
Oduya 210 0.524 0.571 +2
Stuart 159 0.489 0.211 -11
Bogosian 41 0.532 0.500 +0
Jones 31 0.339 0.000 -3
Hainsey 27 0.298 0.667 +2
Flood 26 0.533 0.763 +1
2011-12 Bogosian with:
PartnerTOI CF% GF% +/-
Stuart 369 0.466 0.433 -4
Hainsey 272 0.496 0.609 +5
Oduya 247 0.499 0.222 -10
Enstrom 97 0.556 0.667 +3
Byfuglien 41 0.532 0.500 +0
Flood 19 0.536 1.000 +1
Clitsome 15 0.355 0.000 -1
Jones 11 0.217 0.000 -1
Kulda 3 0.583 1.000 +1
The real interesting one is that Byfuglien ended the season being a minus, even though he was rarely a minus with the teammates he spent the most time-on-ice from. Some would say Byfuglien is + with Enstrom/Odya, but - with Stuart because Enstrom is so awesome and Oduya is decent. But, look at how Bogosian is severely minus with Oduya... is Oduya being pulled down by Bogosian but not by Byfuglien?
A side problem comes from this that it isn't relative to TOI. Is Byfuglien's +2 with 210 mins with Oduya (where they outchanced their opponents) equal to his +2 with Hainsey for 27 mins.

And the biggest one is that it's proven that luck can dominate +/-, especially in short seasons.
You stated that Jokinen was not able to score while his linemates have. In fact, that's actually not 100% true. Jokinen's having by far the lowest On-Ice-SH% shows that he's been unlucky enough that his linemates points have came mostly from shots while he's not on the ice, but the teams shot density, the amount of shots, and the quality of shots hasn't changed. Also, SH% has been shown to fluctuate greatly in small samples while possession (Fenwick/Corsi) and shots are more stable predictor of future production.

Question which is easier to believe:
Jokinen is good enough that he is creating better scoring chances, possession and shots than our teams average, and preventing scoring chances, possession and shots than our team average... but bad enough that the points go in on their goalie and do go in on our goalie...
Or, there's some poor luck to it.

Last edited by garret9: 02-21-2013 at 04:56 AM.
garret9 is online now   Reply With Quote