After the explosion of Michael Grabner the last 3 games (and his long-run goal-streak of 15 goals in 15 games), the question has come on this blog about whether Grabner can win the Calder Trophy (Rookie of the Year). After all, he's now tied for the lead in rookie goal-scoring and tied for fourth in points. The answer is almost certainly yes: I'm pretty sure that the Calder voters only look at Goals, Assists, Points and +/-, and Grabner is high in 3 of those 4 categories (falling short only in assists). So Grabner certainly has a shot, perhaps even a good shot at being a finalist and maybe the winner, especially if he keeps up his torrid pace (or only slows down a little).
In this post I'm going to take a look at the play of 5 of the top Calder candidates and see if they deserve the award. I'm going to use some advanced statistics to see how each player is being used and how valuable their play has been. I would highly doubt the Calder voters would take such a deep look.
The five players I'm looking at are Logan Couture, Derek Stepan, Michael Grabner, Jeff Skinner, and Taylor Hall. This of course omits a few relevant players: Cam Fowler is being excluded because I'd rather not compare a D man to forwards and I don't think he's a likely candidate anyhow (#10 in points, horrible +/-, makes him nowhere near Tyler Myers last year). The inclusion of Stepan is also probably questionable, but when I first looked at the numbers Sunday night, he was top 5 in points among Rookies, whereas now that spot is taken by Sabres F Tyler Ennis. Similarly, the advanced statistic GVT (more on this in a bit), would think that BOS F Brad Marchand deserves a spot here (probably due to his +20 +/- rating). However, there's really no buzz for either of those guys for Calder as far as I've heard, so I'm sticking with these 5 for my analysis.
Below is a selection of basic and advanced statistics of the 5 players. These statistics are gathered from the great resource behindthenet.ca. All statistics below reflect each player's statistics after the games of Sunday, February 13, 2011. (Monday's stats are not included):
|NAME||GVT||GP||TOI/60||Relative +/-||Relative Corsi Quality of Competition ||Corsi Quality of Teammates ||Relative Corsi / 60||Corsi / 60||Normalized Corsi / 60||Penalties Committed||Penalties Drawn||Zone Start%||Total Offensive Faceoff Draws||Total Defensive Draws|
Table 1: The Advanced Statistics of the 5 Calder Candidates:
GVT= Goals Versus Threshold: The amount of goals created by a player's play above that would be created by a "Replacement Player" (Essentially, it tries to duplicate WAR for Hockey)
Relative +/- = A players +/- while on the ice MINUS their +/- while off the ice (with some minutes adjustment)
Relative Corsi Per 60 of Competition = A measure of the skill of the average opposing player on the ice. (Basically it's the average Corsi Relative of the opposing players on the ice while this player was on the ice.)
Corsi Quality of Teammates= A measure of the skill of the average teammate on the ice. (Basically it's the average Corsi of the linemates and D-Men who are on the ice at the same time of this player)
Relative Corsi/60= A Player's Corsi while they're on the ice MINUS the team's corsi when the player is off the ice.
Corsi/60 = Total Shots directed toward opponent's net MINUS Total Shots directed toward Your OWN Net (Per 60)
Normalized Corsi/60 = A Player's Corsi rating normalized to show how the corsi rating would be if a player started an equal amount of time in the offensive zone as the player does in the defensive zone.
Zone Start% = The percentage of faceoffs you're on the ice that are in the offensive zone (not including neutral zone faceoffs)
|NAME||Power Play TOI (Minutes) Per Game ||Penalty Kill TOI (Minutes) Per Game ||Goals||Assists||PPGs||PPAs||SHGs|
Table 2: The Total Goals and Assists of each Calder Candidate and the amount of Power Play and PK time for each player per game.
I've split apart these two tables simply for readability concerns, but I'll be taking each of these tables together. If you're confused by all the numbers I try to explain below:
The first number we should look at is GVT. GVT, or Goals Versus Threshold, is a measure created by Tom Awad of Hockey Prospectus to try and take every aspect of a player's performance and put a number on that player's value. Essentially, it's equivalent to baseball's WAR statistic (Sort of- Rather it's equivalent to RAR - Runs above Replacement). The number attempts to essentially say how many more goals a player has produced compared to the amount that would be produced by a "replacement player" called up from the AHL (this year an example of a "replacement player" would be Vancouver's Tanner Glass). The statistic is far from perfect and misses a large amount of context about how a player is being used, but it's a good starting point for determining a player's value to the team. Last year, Mark Streit led the Isles with 10.7 GVT, while Matt Moulson was number 2 with 10.2 GVT. Two years ago, Mark Streit led the team with 14.0 GVT (Streit was a good player). Last year, Alex Ovechkin led all non-goalies in GVT with 30.1 GVT. Tyler Myers led all rookie non-goalies last year with 14.6 GVT.
Of all rookies this year, Logan Couture is first with 10.4 GVT, with Grabner close behind at 9.8 GVT, and Skinner close behind in 3rd at 9.2 GVT. Functionally, under GVT, these players are all really close in value. Hall and Stepan are far behind with 4.7 and 4.1 GVT respectively and are not even the #4 or #5 rookies in GVT (the aforementioned Marchand is #4 in GVT at 7.5GVT, and Chicago's Bryan Bickell is 5th with 6.3). In other words, GVT would say that Couture, Skinner, and Grabner are clearly the top rookies and are a step above the rest (this includes Cam Fowler).
There's a clear other reason to dismiss Stepan as well from this list, in that he leads the rookies in zero of the above categories and has been totally babied by the Rangers, in that he's been used in offensive zone situations far more than defensive zone situations (as you can see by the faceoff starts he's been on the ice for). And even with this advantage, he doesn't lead the rookies in anything, and when you normalize his corsi rating, he turns out to be the worst of the 5 rookies in driving his team to a positive shot differential. Whereas there are some things favoring Taylor Hall in the Calder race, there's nothing back to Stepan's candidacy for the trophy.
Anyhow, I've actually gone over these numbers for a while now, and can't actually make a clear decision on who right now should win the award. So I'm going to present each of the four other players' cases:
The Case for Taylor Hall:
Hall's the least likely to win, or would seem so but for his high profile as the #1 draft pick. Using the advanced statistics above, Hall does however have two key points going for his candidacy. First, he has the highest relative corsi of the five players....what this means is that when he is at on the ice, his team's shot differential (the amount of shots directed at the opponent's net minus the amount of shots directed at a team's own net) was 13 better than when he was off the ice. This is the best among the rookies. The end result is that he has a positive ACTUAL corsi value: when he's on the ice, more shots are directed at the opponent's net than at Edmonton's own net. This is an amazing result, given the terrible team he's on, and he's done this despite having played with terrible teammates on the ice (as you can see from his teammates' corsi, the guys he plays on the ice with have corsis of NEGATIVE 7! Meanwhile, he's also playing against tough opposition.
In addition, that's not the only thing in Hall's favor: For one thing, he's been incredible at drawing penalties, having drawn the 2nd most penalties in the league (26) and only committing 6. Remember, every 5 power plays result in a goal, so by creating 20 more power plays than he's cost his team, Hall has essentially created 4 extra goals. That's very impressive and can make up some of the goals he's behind Skinner, Couture, and Hall.
Of course, there are problems with Hall's case as well. The GVT difference is caused by the fact that he is behind in goals and assists - which are obviously still important. Meanwhile, while he hasn't been babied, he has been given a whole bunch of opportunities to put up big numbers and he hasn't really taken huge advantage. He's seeing more offensive zone faceoff draws than defensive ones, and thus his zone-start % (which measures the ratio of offensive draws to defensive draws) than Couture or Grabner (and Skinner too barely). The end result is that his relative corsi edge is basically nil once you take into account the zone-start advantage he has.
In addition, his skill at drawing penalties is equaled by Skinner, who has the same +20 penalty differential, and actually is drawing penalties at a higher rate.
Meanwhile, Hall is getting basically 0 PK time, unlike Couture and Grabner, and is getting a good number of power play minutes. To be fair, only 7 of Hall's 34 points are on the power play, so it's not like his point numbers are inflated by this but well that's a negative too (we would expect his point total to be higher due to power play numbers).
All in all, I don't think his lower goal and assist numbers are overcome by his ability to direct shots at the opposing net or his penalty drawing ability, due to the fact that he's had a bunch of advantages that he hasn't taken advantage of.
Jeff Skinner, having a 9 point edge over the #2 Rookie (Couture), is probably right now the favorite for the Calder. And some of the advanced stats would tell us that he deserves this. Like Hall, he is a MONSTER at drawing penalties, drawing penalties at the best rate in the league and having already added 4 goals to his team through drawing penalties alone (making up the goal difference between him and Couture/Grabner). Meanwhile, while he gets the same power play minutes as Hall, he actually has taken advantage, with 16 of those points coming on the power play. Thus the extra points, good play on the man-advantage, and the ability to draw a ton of penalties (and get his team ON the man advantage), make Skinner a tough guy to beat for the Calder race.
That's not to say that Skinner doesn't have a whole bunch of things pointing to him negatively. He's played in basically the same situations as Taylor Hall - slightly more offensive than defensive faceoff draws, lots of power play time, and no penalty killing time, which we'd expect to result in favorable results. Meanwhile, he's been on the ice against the worst opponent's of any of the big 4 Calder candidates, another factor making us expect better results, and he's had better teammates-on-ice-with-him than Hall or Grabner. And despite this, Skinner really hasn't done a great job at keeping the puck directed at the opponent's net, with a barely positive corsi that becomes negative when normalized (meaning that his play has resulted in more shots going at his own net than his opponent's) and a relative corsi that is far below that of Grabner, Couture, or Hall. In short, Skinner's play has been far worse at keeping the puck going in the right direction on 5 on 5.....and he doesn't play on the penalty kill either.
Thus, Skinner has terrific numbers. But on 5 on 5 play, his play has clearly been inferior to Grabner or Couture...except in his ability to draw penalties. Skinner has taken advantage of a lot of power play time to get his better numbers. Does that mean he's played the best, or that he's simply combined good play with great opportunities? I think that's a value judgment.
Couture is tied for the Rookie lead in goals, has more assists than Grabner, and has the best +/- of any of these four. That's what will right now give him a shot to win the Calder. But in addition to his goal advantage (the +/- is at least a little the result of being the sole rookie to play on a good team), he has other things going for him. First, his corsi rating is the highest of any of these players by far: when he's on the ice, his team directs 15.26 more shots per 60 at the opponent's net than the opponent can direct at the San Jose Net. This is not wholly the result of his team being actually good (compared to the other 4 players' teams), as his relative corsi is actually +11.7, showing that San Jose's corsi is 11.7 shots better per 60 when he's on the ice than when he's off the ice. Meanwhile, Couture does this while playing against pretty good competition.
Unlike Hall and Skinner, Couture actually is called upon in defensive situations. Couture plays over a minute of penalty kill a game and actually is on the ice for more defensive faceoffs than offensive draws, if only barely. He is actually called upon to play a 2 way game and seems to do it well.
On the other hand, there are negatives to Couture as well. Like Hall and Skinner, he gets a large amount of power play time (over 2.39 minutes per game), but while he has capitalized by scoring goals on the power play, he doesn't set up his teammates on the power play (only 1 assist). Moreover, while part of his great corsi is caused by his own play, a lot of it is caused by his teammates.....he has the best teammates on the ice with him at all times of any of these four players...by far, and in fact has the best teammates on the ice with him (in terms of corsi) on his own team! Meanwhile, Couture hurts his team a little by committing more penalties (11) than he's drawn (9), essentially costing his team .4 goals. Moreover, not only has he cost his team 0.4 goals due to this, but by committing the most penalties of any of these rookies, he's deprived his team of his own PK services by doing this.
Essentially, the knock against Skinner is that unlike the other guys, he's on a pretty good team. Not really his own fault, but we have to take it into account somehow...how much has this boosted his numbers? It's hard to say.
And now we have Michael Grabner. Tied for the Rookie lead in goals (24) with a pretty good +/- for a crappy team (+10) - these numbers will get him noticed right now by Calder voters. But the question is: right now, has he been the best rookie so far?
Well Grabner has a whole bunch of advantages over the others. First of all, he has truly been called upon to be a defensive player a good portion of the time, as part of the Isles' defensive line (with Frans Nielsen, defensive specialist). This can be seen by the fact that hes' seen by far the least offensive zone faceoffs of any of these rookies and has seen the most defensive zone faceoffs of any player as well. Meanwhile, he barely plays any power play time, getting only half a minute of power play time a game. This makes the fact that he has 24 goals even more ridiculous. In addition, Grabner's relative corsi is nearly as good as Couture's and Hall's (+10), and his normalized Corsi, which takes into account his bad zone-start % puts him #2 among the rookies in Corsi to Couture. And this is all with poor teammates on the ice at the same time! (Before you point out that Frans and KO aren't bad teammates, I'd point out that the Quality of Teammates measurement also counts the DEFENSEMEN and Isles' D-Men have been horrible).
Meanwhile, Grabner plays a good bit of PK time, 1.16 minutes per game, and even has two short-handed goals. And while Grabner hasn't drawn a ton of penalties like Hall or Skinner, he has only committed 3 penalties all year, keeping himself on the ice and his team off the kill. Thus his net of +6 penalties has resulted in him creating 1.2 more goals for the Isles. A nice little cherry on the top.
Now Grabner has his own negatives too. He isn't a playmaker and has the lowest assist totals of any of these players. Moreover, while his corsi numbers indicate that he's not been a defensive liability, and if anything has been a defensive boon to this team, it's kind of hard to tell how much of these numbers is his own creation and how much was created by Frans Nielsen and Kyle Okposo. For example, since Okposo has been on Grabner's line (since he came back), Okposo's play (almost entirely with Nielsen and Grabner), the Isles have a shot differential of over PLUS NINETEEN! This has resulted in a large rise in Grabner's corsi numbers and has naturally coincided with Grabner's recent run of goals. While Grab's finishing is his own ability, it would seem that the numbers above are failing to take into account how awesome Grabner's two linemates have been over the last 12 games.
Conclusion and Thoughts: Right now, I'd probably give the Calder to Grabner, but Couture and Skinner have really close cases (I think Hall is a bit behind), if not better ones. And can Grabner keep up this pace? Probably not completely...no one can keep up the pace he's set the last 5 games and his shooting % is way too high for it to be entirely skill...it's bound to regress some. But even if he slows down a little, Grabner is right now on track to very possibly overtake Skinner and Couture in not only the conventional statistics, but the advanced ones that give a more complete pictuere of how well he's been playing.