Team MVP

[Stickytape]

I think it's great that a case can be made for several players.  This isn't a team that's riding on just one or two big names.

I'd say Phaneuf has the second hardest job on the ice, after Reimer, and there's no way to make that look easy; plus he has the C on his chest, so I think people expect more out of him than your average player.  However, if you set that aside and simply look at what he's contributed to the team, I think he's definitely in the running, and quite possibly the guy.

That said, Reimer has stolen a couple of insanely hard games, and has come away with wins despite the team being outshot as much as 3 to 1.  And this is on top of a lot of testing periods where he and Scrivens were swapping spots, which has to be hard on the nerves.  He's been amazing since the trade deadline and I'd say he's the biggest reason why the leafs have clinched their first playoff spot in 9 years.

I feel like it's one of those two guys, though there are a lot of honourable mentions.  Kessel and Kadri have both been responsible for a lot of points, and Kadri's in particular have resulted in a lot of almost single-handed wins.  Also, I have the feeling that if Lupul had been healthy all year, it'd be him, no contest.  And there are a few other guys who have had great years in general, who have been invaluable; McClement, for example, has been a huge reason for the way the Leafs have turned the 3rd worst PK into the 3rd best.

 

[GhostofPotvin29]

- logs 25 mins / night against the other team's best players every game.  (has the highest CORSI quality of competition numbers in the league for d-man .. Kulemin is #1 in that area ... someone who understands those stats better can explain better than I)

Well, I'm a bit of a novice to it too but my understanding of it is that the QOC number measures the Corsi of the players he played against and that dictates the "quality" of the competition. The problem with that, though, is that it seems to me as though Corsi itself doesn't really adequately compensate for the way players are used. Jay McClement, for instance, has the second worst Corsi numbers on the team behind only Holzer so, and again this is my novice's understanding of it, that means that when an opposing forward went up against McClement it graded him as being easy competition and Clarke MacArthur, who has the highest Corsi numbers, as being the toughest. I don't think that really matches up with how we see the game.

I think for McClement it tells us that he's been facing pretty tough competition (only Grabovski & Kulemin faced tougher), but that he's also been playing with a lot of not-so-good linemates for good stretches of the season.  Look at his offensive zone start % - 28% - the next lowest is 35%.  So he hardly EVER starts in the offensive zone, which is going to affect the number of shots he is on against.

So at 5 on 5, McClement is probably a good defensive C, and overall average.  Since those are 5v5 numbers, you need to look at how he's done on the PK to really see his contributions.  This link has a good breakdown of how to gauge his PK work: http://theleafsnation.com/2013/4/16/jay-mcclement-and-the-penalty-kill-part-i

That link also shows how he has started more shifts in the defensive zone than almost anyone else in the league, and that he gets killed when he's playing with Orr/McLaren.

 

[Kin]

Believe me I am the last guy who preaches these stats as being the be-all end-all. The quality of competition one is interesting though and McClement is way up there in that one as well.  I've heard people dismiss McClement as "a bad hockey player" because of those numbers.  So to me they are disposable in many cases.

But the thing about that is that while Corsi numbers may have questionable value they're at least a meat and potatoes sort of thing, right? It's just a raw calculation of shots when a player is on the ice vs. when he's off. What you might choose to glean from those numbers is an entirely separate conversation. Data, and having as much of it as possible, is a good thing. The problem I think with advanced hockey numbers is that people are drawing lousy conclusions from that data and QOC seems front and centre there.

The problem with QOC and why I think it's much, much less valuable than Corsi is that A) it's based on Corsi so it doesn't make sense to like it and not Corsi but more importantly B) it attempts to quantitatively express something they haven't quantified and lastly C) it's name is terrible, you can't express quality with numbers.

 

[GhostofPotvin29]

Corsi is seriously flawed if this guy is the worst on the team.

There are a bunch of different Corsi numbers and yes I think many are very flawed. They also suggest that Lupul is a terrible player too.  So yeah two of our best forwards are apparently garbage.  Puh'leez.

The quality of competition one is one of the few I reference because I don't see it as convoluted as some of the others.

Where are you seeing it suggests he is a terrible hockey player?  From what I've seen it suggests that he's a poor possession player, and throughout his career has not been able to drive possession on a line to the point where they are outchancing the opposition. 

I think it depends on the team whether you can live with that.  With good goaltending you probably can.  With last season's goaltending, it becomes more difficult.

 

[Kin]

I think for McClement it tells us that he's been facing pretty tough competition (only Grabovski & Kulemin faced tougher), but that he's also been playing with a lot of not-so-good linemates for good stretches of the season.  Look at his offensive zone start % - 28% - the next lowest is 35%.  So he hardly EVER starts in the offensive zone, which is going to affect the number of shots he is on against.

So at 5 on 5, McClement is probably a good defensive C, and overall average.  Since those are 5v5 numbers, you need to look at how he's done on the PK to really see his contributions.  This link has a good breakdown of how to gauge his PK work: http://theleafsnation.com/2013/4/16/jay-mcclement-and-the-penalty-kill-part-i

That link also shows how he has started more shifts in the defensive zone than almost anyone else in the league, and that he gets killed when he's playing with Orr/McLaren.

Two things:

1. I agree in general regarding McClement. I wasn't using his Corsi numbers to express that he's not good. Just why QOC is flawed if it represents McClement as easy or low competition.

2. You reference zone start a lot as being an important influence on stats and while I think there's some wisdom there I'm a little confused as to the mechanics of it. Don't the majority, or at least a significant minority, of a player's shifts start between faceoffs? Right? One team flips it deep, guys come on. So when you quote offensive zone % is that just referring to faceoffs? Or does that account for every single shift a player takes?

 

[GhostofPotvin29]

Yeah, I was just trying to piggyback onto your overall conclusion about McClement not being bad.

2. You reference zone start a lot as being an important influence on stats and while I think there's some wisdom there I'm a little confused as to the mechanics of it. Don't the majority, or at least a significant minority, of a player's shifts start between faceoffs? Right? One team flips it deep, guys come on. So when you quote offensive zone % is that just referring to faceoffs? Or does that account for every single shift a player takes?

It's off face-offs, I think it would be too difficult to track otherwise.  Apologies for not explaining that when I wrote it, I know I have other times but I know not everyone will know it, and forgot to this time.

So zone starts are the % of starts in the offensive zone from a face-off.

 

[Kin]

So zone starts are the % of starts in the offensive zone from a face-off.

Right. So my question then is more about how significant face-off starts are in the overall picture of a player's game when weighed against every other shift they start between face-offs.

 

[maplesyrup]

I'm comfortable with Phaneuf pick though I really liked the group as a whole more than a collection of it's parts. On a related note, this is the first year in a looooong time that I didn't actually hate having a guy (or guys) on the team.       

 

[armche123]

http://www.nhl.com/ice/threestarsstandings.htm

Great job by James!
He has my vote.

 

[oakl0008]

Reimer and Kessel should share the award! 

HM: Kadri  8)

 

[TML fan]

Colton Orr. He protects our stars.

 

[bustaheims]

I gotta go with Kessel. Down the stretch run in April, when the team needed offence, he raised his game and pushed the offence. Without him, the team very well could have fallen out of the playoffs. He may not have carried the team all year, but he carried them when they needed it the most.

 

[GhostofPotvin29]

So zone starts are the % of starts in the offensive zone from a face-off.

Right. So my question then is more about how significant face-off starts are in the overall picture of a player's game when weighed against every other shift they start between face-offs.

I think over the course of a season if a player has a low % of offensive zone starts, then you can factor that into your assessment of that player.  It's common sense to think if a player starts a low % in the offensive zone then it is going to lower their offensive opportunities, so it can help put things into context - either observations you have made, or other stats.

I can't speak to how significant it is, I suppose that will depend on your own conclusions of it.

 

[Corn Flake]

Corsi is seriously flawed if this guy is the worst on the team.

There are a bunch of different Corsi numbers and yes I think many are very flawed. They also suggest that Lupul is a terrible player too.  So yeah two of our best forwards are apparently garbage.  Puh'leez.

The quality of competition one is one of the few I reference because I don't see it as convoluted as some of the others.

Where are you seeing it suggests he is a terrible hockey player?  From what I've seen it suggests that he's a poor possession player, and throughout his career has not been able to drive possession on a line to the point where they are outchancing the opposition. 

I think it depends on the team whether you can live with that.  With good goaltending you probably can.  With last season's goaltending, it becomes more difficult.

I've seen it all over twitter from a particular site that uses those stats as the Bible on everything.

Driving posession as a 3rd line player and a heavy PK minute muncher to a positive degree is a ridiculous expectation. 

 

[Bonsixx]

Advanced stats bore the hell out of me.

Anyway, Reimer is the MVP for sure. Kessel and Phaneuf get the honourable mentions.

 

[Corn Flake]

Believe me I am the last guy who preaches these stats as being the be-all end-all. The quality of competition one is interesting though and McClement is way up there in that one as well.  I've heard people dismiss McClement as "a bad hockey player" because of those numbers.  So to me they are disposable in many cases.

But the thing about that is that while Corsi numbers may have questionable value they're at least a meat and potatoes sort of thing, right? It's just a raw calculation of shots when a player is on the ice vs. when he's off. What you might choose to glean from those numbers is an entirely separate conversation. Data, and having as much of it as possible, is a good thing. The problem I think with advanced hockey numbers is that people are drawing lousy conclusions from that data and QOC seems front and centre there.

The problem with QOC and why I think it's much, much less valuable than Corsi is that A) it's based on Corsi so it doesn't make sense to like it and not Corsi but more importantly B) it attempts to quantitatively express something they haven't quantified and lastly C) it's name is terrible, you can't express quality with numbers.

I guess all in all, Corsi looks like a crazy amount of conclusions drawn just on shot counts.

 

[Kin]

I think over the course of a season if a player has a low % of offensive zone starts, then you can factor that into your assessment of that player.

I agree. My question, though, is how much and, really, it's something that seems to me to be the sort of thing that would be easy to track. Face-off starts vs. other shifts. It seems to me that where the advanced stats movement is really coming up short is that they haven't done anything along the lines of DRS or UZR in baseball where numbers are combined with observation and informed judgment to give us a better sense of what a player is actually contributing on their shifts.

  It's common sense to think if a player starts a low % in the offensive zone then it is going to lower their offensive opportunities, so it can help put things into context - either observations you have made, or other stats.

Well, it is and it isn't. If the ratio of shifts started on face-offs vs. changes on the fly are 50-50 or even higher then clearly the number of offensive chances will be heavily affected by where those faceoffs are. If the split is 80-20 the other way though(and I can't imagine it is but bear with me) then where those face-offs are wouldn't play a huge role in a player's season. A more significant measurement might be, say, when players change on the fly whether it was their team who dumped the puck or not as that would usually measure possession when they get on the ice.

I can't speak to how significant it is, I suppose that will depend on your own conclusions of it.

Right. And I'm not so much asking you so much as I am expressing the opinion that the data being presented by these advanced metrics, while helpful, isn't complete enough for us to start saying things like "Phaneuf played against the toughest competition" or "McClement took only 34% of his faceoffs in the opposing zone which means..." whatever. It's information and information is good but people are rushing too quickly from information to conclusion.

 

[Kin]

I guess all in all, Corsi looks like a crazy amount of conclusions drawn just on shot counts.

And you know, you know I'm a super big numbers guy when it comes to Baseball so this isn't an issue of not liking numbers vs. human observation for me but rather just that the advanced metrics in hockey...they're not there yet.

I think it's pretty telling, for instance, that in this ERA where Baseball and Basketball, and even Football to an extent, are embracing advanced statistics on a professional level we haven't really seen that in hockey.

 

[GhostofPotvin29]

I guess all in all, Corsi looks like a crazy amount of conclusions drawn just on shot counts.

And you know, you know I'm a super big numbers guy when it comes to Baseball so this isn't an issue of not liking numbers vs. human observation for me but rather just that the advanced metrics in hockey...they're not there yet.

I think it's pretty telling, for instance, that in this ERA where Baseball and Basketball, and even Football to an extent, are embracing advanced statistics on a professional level we haven't really seen that in hockey.

Well from recent articles by people like Mirtle, it seems like a number of teams are using advanced metrics, and they already keep their own versions of "corsi."

I think it's a good thing to add to the conversation.  I agree with the central idea of corsi as it relates to puck possession correlating to success.  I don't think you can be a poor corsi team and be successful over the long haul.

 

[WhatIfGodWasALeaf]

Reimer

HM: McClement, his impact on the PK has been huge, I think his example has made others better too.

HHM: Phaneuf, just eats up minutes, has benefited from not having to do everything himself this year.

HHHM: Kadri and Kessel, dynamite offensive production.