The thing about ‘advanced’ stats in hockey is that they’re not really intended to be all that advanced…but compared to the crappy normal stats like +/- they are miles ahead. A lot of stats that look at usage–zone starts and quality of competition–are designed to illustrate a player’s deployment to provide context to his goal production. And the granddaddy of all fancy stats, Corsi, is not really that complicated once you wrap your head around it. The object of the game is to score goals and prevent them from being scored against your team, and since goals come from shots, the Corsi stats are intended to be simply a richer tool for analysis. So, when you think of a “good” hockey player, the tools and skills he owns should lead him to be a good possession player. Things like skating ability, hand-eye coordination, vision, anticipation, etc., these are the building blocks that make up a quality hockey player. Of course, those things don’t always come through in Corsi stats, and today I wanted to do sort of a case study on a Minnesota Wild player who is still very young, and seems to be developing nicely but is still a “bad possession player” by the numbers: Mikael Granlund.
Granlund has only 50 NHL games under his belt, so we can start to see trends develop, but obviously the Small Sample Size rule (SSS) is in effect. Plus he’s only 21 so he’s developing his skills every day. But any way you look at it, Granlund’s possession numbers are pretty, pretty bad. In the shortened 2012-13 season, the only forward who had worse Corsi numbers was Johan Larsson, and he only played in one game. This year, those numbers are a bit better but he’s still in the red, so to speak. One great tool that has developed in recent years is the WOWY analysis (With Or Without You) which outputs players’ stats when they are on the ice together and separate. Let’s take a look at the four wingers with which Granlund has skated the most, in order from least time to most (note, I gathered this data over the course of about a week so numbers may be a tiny bit off due to being outdated):
Matt Cooke (43:14 together)
So this is pretty clear…these guys don’t do well together. Each skater has better Corsi numbers when separate, and when together the ice is tilted considerably against them. I think team and player chemistry is one of the most fascinating things to think about in sports, but I think it’s lazy analysis to just say “ah well these guys don’t match up, Granlund is a finesse player and Cooke is a bruiser.” That said, I don’t want to spill a lot of pixels here because these two won’t be seeing consistent time on the same line. Suffice it to say I think each guy would be okay with that, there is a clear depression of their possession ability when on the ice together.
Dany Heatley (117:47 together)
Here we see the opposite trend–in fact, Heatley is one of the only players who really sees a boost in his Corsi numbers when skating with Granlund. We’ll get to that more in a minute. Heatley and Granlund played on the second line to start the season, and while Granlund played decently, Heatley played himself right down to the fourth line. This may explain some of the “Heatley separate” measurement, when you’re skating with Torrey Mitchell and Zenon Konopka, you’re not going to get a lot of good possession stats. There might be some chemistry at play here…the simple fact that Granlund is a set-up guy and Heatley is a shooter may be what’s driving the boost. In a recent study, I found that Dany Healtey is still shooting the puck a lot compared to his teammates. With Parise’s injury, I think there’s an outside chance that the lines get shaken up, so Granlund and Heatley could see some more time together, but I think long-term this isn’t going to be a pairing we’re going to see much of.
Nino Niederreiter (135:09 together)
Alright, now we’re getting somewhere. The Granlund-Niederreiter-Pominville line (henceforth and forevermore referred to as simply ‘Graninoville’) has caught fire like Katniss Everdeen lately. Sorry, that sort of wrote itself just there. Granlund separate from Nino is almost identical to his overall number, but Nino separate from Granny is an insane 64% and change. From an X’s and O’s perspective, the pairing of Nino and Granlund combines the finesse and strength elements…but ultimately I think what we’re seeing here is that when Niederreiter is not on the second line, he has been on the first line with Koivu and Parise…and anybody skating with those guys is going to get some ca-razy Corsi numbers. Anybody with two eyes in their head can see that the Graninoville line is clicking lately, so more observation is necessary. Besides, the real punch from this line has been with Granlund and another guy…
Jason Pominville (201:48 together)
It’s night and day…when these guys are on the ice together they have a very nice 54% corsi, but when Granlund is separate he flounders while when Pominville is separate his is still great. Granlund’s sub-36 corsi number is only in about 90 minutes so not really representative…he’s not great but he’s not *that* bad. As with Niederreiter, when Pominville is apart from Granlund he’s with Koivu/Parise so again, that explains that number. Getting back to the chemistry bit, these guys are just simpatico right now…they know where the other will be on the ice, they make excellent passes to each other, and they have turned that chemistry into 8 even strength goals between them (6 for Pommer and 2 for Granny.) Again, with the Parise injury they might get split up, but on the other hand Mike Yeo would be crazy to split these guys up based on what they’ve shown over the course of the last month or so. If it was up to me I’d keep them together–this could be the real secondary scoring threat for the Wild for years and years to come so let them develop their game together I say. Time will tell whether Yeo decides to follow my advice, but what’s clear here is that Granlund needs Pominville way more than Pominville needs Granlund.
So back to my original thought of what makes a “good” hockey player…Granlund has a lot of the things I mentioned: he’s a great skater, a fantastic passer, he is probably already the best passer on the team, he has good anticipation and vision, and he’s developing the strength that has been the biggest hole in his game. So why does he continue to be a negative Corsi player? There’s a lot of things that we could talk about…usage, system, teammates, etc. but any way you look at it, he’s not seeing the possession. Observe:
Alright so the first chart is simply Granlund’s Corsi stats when together with each skater vs those guys separate from Granlund. It’s easy to see that the red bars are almost all higher than the green ones. In fact, the only skaters who have a higher Corsi together with Granlund than separate are Heatley (already covered,) Koivu (2:45 together) and Konopka (3:18 together) so for those last two guys there’s nothing to say, they just haven’t spent enough time together to make any conclusions. The second chart is the same information, but just presents the difference between their together and separate numbers. But the bigger picture is that Granlund seems to drag down the possession ability of nearly every guy he skates with. But WHY? I don’t think I have the answer.
One thing that I think that contributes to all this is that Granlund needs to work on his defensive game. It’s too early to judge but I do worry a little bit that he will never be a great two-way center, and his value will be purely on offense. If he can’t make the good defensive plays to prevent shots, if he can’t win puck battles (in any zone,) if he can’t make good zone exit passes, all these things will contribute to negative possession numbers. Now, while Corsi is sort of the gold standard right now, it’s not always indicative of goal-scoring. Take a look at the possession numbers from the season Corey Perry scored 50 goals, and you can see that negative possession can still lead to goals (and obviously a 17% shooting rate helps.) But in the long term, clearly the team would prefer their second line center to be a good two-way player and a good possession player. I’m sure there’s research out there that looks at players developing over their career and whether it’s possible for a guy to switch from a negative corsi to a positive corsi, and yes, again a lot of it depends on linemates, systems, deployment, and all that, but for now all I can do is wonder.
To that end…there are some articles that are trying to prove a point and some that pose a question without an answer. This is certainly the latter. I would love to hear in the comments section what your opinions are–if Granlund has the tools to succeed and be productive in the NHL, why are his possession numbers so low? Do you think it’s just because of the gaps in his defensive game, or is there something else? Drop a comment and let me know. Thanks for reading and don’t forget to follow me on Twitter, @Hashtag_Hockey