Make sure to follow Travis on Twitter
On Monday, Tyler Dellow of MC79Hockey
opted to publish a detailed response to some of the local media rhetoric in Ottawa, who are currently painting Erik Karlsson as the next Bobby Orr during his stunning 2011-2012 statistical output. Fifteen goals and fifty-one dimes in sixty-four games played? Gaudy, gaudy numbers.
In response, Tyler Dellow's entire report - read here
- breaks down Erik Karlsson's assist totals at even strength, focusing on the validity of each. By the end of the article, you're left with legitimate questions as to how real the twenty-one year old's assist totals are.
Largely, Dellow - either directly or indirectly - has made a few points here. One, Erik Karlsson's numbers are a bit inflated. Two, Erik Karlsson's benefited mightily from some puck luck and/or randomness. Three(a branch off of points one and two), Erik Karlsson's recording assists on plays that other defensemen aren't.
My initial response to Dellow's report on Twitter was broken into three elements. One, assists are a poor statistical measurement. Two, assists are a macro-problem that's plagued the NHL for years. Three, Erik Karlsson's become a public enemy of sorts here because he's an outlier, and that his closest comparables would have the same kind of egregious ratio of true assists/false assists, relatively speaking.
It was the final of three points that Dellow disagreed with, noting that there was no quantified or qualified evidence to suggest that the true assist/false assist ratio existed across the board. A fair take, but one that could warrant a response.
Enter Brian Campbell of the Florida Panthers, who will now become the blog's whipping boy. Why did I choose Campbell? Pretty obvious. Like Karlsson, a good chunk of Brian Campbell's work comes on the power-play, and although his thirty-nine assists are second among defensemen in the National Hockey League, he has just fifteen at even strength.
Going back to the start of the 2011-2012 season, I made note of all fifteen even-strength assists, using the same exact measurements and formulas that Dellow used for Erik Karlsson. Further, I broke the assists up into four separate categories - the exact four categories that Dellow used to qualify Karlsson's ES productivity.
The four categories are as followed:
a) Plays where Brian Campbell was trying to pass for a scoring chance;
b) Plays where Brian Campbell was just getting the puck on net;
c) Plays where Brian Campbell was making a pass across the blue line; and
d) Plays where Brian Campbell made no appreciable offensive play
Although this blog will offer an alternative viewpoint, it's purpose is not
the same as past rebuttals I've drafted on HockeyBuzz or elsewhere. Usually, those are reserved for the steaming pile of shit mailed in by the likes of Damien Cox, Steve Simmons, et al.
Dellow's work here does bring an issue to the forefront, and although Ottawa fans won't love that Erik Karlsson moved into a brief negative limelight, it certainly has (some) substance.
My response - through screenshots, descriptions, and statistical measurement - will not
a) Reject the notion that Erik Karlsson's benefited from poor judgment re: assessed assists;
b) Reject the notion that the assist is a flawed statistical measurement; and
c) Reject the notion that randomness/luck hasn't had a measurably positive effect on Erik Karlsson's totals.
However, my response will
a) Emphasize that this is not an Erik Karlsson problem, but rather a league-wide issue among statisticians and their inability to properly understand the definition of an assist and subsequent wanton disregard for application of the statistic;
b) Show that offensive defensemen - especially those with reputation - benefit from absurd wiggle-room, which I think is fairly obvious to begin with; and
c) Highlight that Erik Karlsson's assist numbers, while high, are the product of his constant involvement in the play relative
to other offensive defensemen in the NHL, driving home the point that Erik Karlsson's performance in general is still as impressive as we originally thought.
: With no conclusion - teased by Dellow for a future column - I'll opt to reserve my rights regarding any potential and additional changes. It's a bit difficult to respond to something that doesn't come to any definitive end.
The above-mentioned 'Point C' remains the most important of the three. Dellow concludes in his blog that of Erik Karlsson's twenty-seven even strength assists, only ten were the direct result of him trying to find an open man for a goal.
Or, Erik Karlsson's true assist percentage is around 37%.
Since Brian Campbell has fifteen even-strength assists, the argument of relativity suggests that about 5.55 of his total marks should be valid dimes.
As you probably expected, Brian Campbell's alleged performance is probably[definitely] more egregious than Erik Karlsson's. Below, screenshots and descriptions for each of his fifteen assists:
a) Plays where Brian Campbell was trying to pass for a scoring chance:
[Primary] Buffalo, 10/29/11. Drop pass by Campbell to the entering Garrison, who rifles one on the one-timer. Good look here by Soupy.
[Primary] Winnipeg, 10/31/11. Drop pass by Campbell to Garrison walking the blue line, who one-times it home. Funny how often these two connect or come close to connecting on a game-to-game basis.
[Secondary] Dallas, 11/15/11. Hard one to slot, but after a bit of review I'm going to give him this one, especially since I went another way on a goal v. Minnesota. Campbell got rid of the puck like a live grenade with oncoming pressure to a semi-open Mike Santorelli. It was Santorelli who corralled the puck, waited for an open seam to Garrison up top, then set him up for the blast. Still, Campbell a) did keep the zone, and b) did keep the play alive cleanly.
[Secondary]Carolina, 12/18/11. Campbell makes most of four-on-four by skating in clean, around a sprawled defender, then finding an open man just outside of the crease. Quick scramble, Versteeg finishes. Perhaps one of his best dimes of the year.
b) Plays where Brian Campbell was just putting the puck on net:
[Primary] New York, 10/8/11. Pretty straightforward one here. Brian Campbell puts a shot on goal, rebound is deposited into a juicy scoring area, and Stephen Weiss beats Al Montoya clean.
[Secondary] Philadelphia, 11/13/11. Late game shot on net that winds up flying all over the place before Fleischmann corrals and buries.
[Primary] Minnesota, 2/23/12. Watch this replay a trillion times, and it's almost impossible to decide whether this was an angled-pass or a soft shot on net. At first look, it certainly appears soft enough - and directed enough - to warrant pass. But, there's a lot of - well,
shit - going on between Campbell and the net, and it certainly looks like he was hoping for some kind of screen or deflection as he threw it on. Here's one I'd consider moving around. Whatever it was, it was certainly effective - Bergenheim scored seconds after in tight.
c) Plays where Brian Campbell was moving the puck across the blue line:
[Secondary] Winnipeg, 11/10/11. Campbell floats one over to Garrison, who sends it down into the corners for Versteeg. Puck eventually makes its way back to Garrison, who scores on a dart past Pavelec.
d) Hilarious assists awarded to Brian Campbell for no appreciable offensive play:
[Secondary] Ottawa, 10/27/11. Campbell's exiting the zone clean, but it's not a stretch pass or anything of the sort. Versteeg actually receives the pass before mid-ice, carries in, and later hits Jason Garrison with the no-look drop pass for the goal.
[Secondary] Washington, 12/5/11. Clearing the defensive third. Weiss-Fleischmann run the two-man show through the neutral zone and finish brilliantly.
[Secondary] Boston, 12/8/11. Campbell frantically tries to clear the zone, hits puck sideways. Fleischmann also hits puck in defensive zone trying to clear, but his play finds an open Versteeg on the fly. Goal, KV. Note:
Goaltender was pulled and this is not considered an even-strength marker, but worth putting into the conversation anyway, if only to show how ridiculous some of thesereally are.
[Secondary] Toronto, 12/27/11. What you're looking at is the aftermath of a Brian Campbell poke check, which somehow turned into a secondary after Tomas Fleischmann put one through the sieve Jonas Gustavsson.
[Primary] New York, 1/5/12. Brian Campbell finds a forward-skating Mike Santorelli, who receives the puck before mid-ice, carries in, then puts a shot past Martin Biron.
[Secondary] New Jersey, 1/6/12. Brian Campbell passes to Michal Repik, who enters the zone on his own. Barch eventually scores.
[Primary] Colorado, 1/18/12. This one's unreal. Pictured: The puck coming off of Campbell's stick en route to Weiss, who corrals it before even entering the neutral zone. Weiss proceeds to single-handedly carve up the entire Avalanche roster - Adam Deadmarsh
included - and beats J.S. Giguere soon after.
[Secondary] Washington, 2/17/12. You'll notice that Brian Campbell's already passed the puck in the direction of Stephen Weiss(two strides deep in this picture), who enters the zone on the fly. Weiss takes the puck all the way around the back of the net, holds, then finds Tomas Fleischmann for the one-timer.
Now that I've essentially burned a hole directly through Brian Campbell, let's try and bring this together.
What we're seeing here is an alarming trend with offensive defensemen - or, defensemen - in general. Offensive defensemen tend to see higher assist totals and, in total only, more false assists. However, using Brian Campbell - and in the future, Alex Edler - as a parallel of sorts, it appears that the true/false assists are incredibly comparable.
So, why are Erik Karlsson's numbers flying off the charts? For starters, he's the best offensive defenseman in the National Hockey League right now. He's getting the puck up to his forwards in key scoring areas with regularity. At even strength, he's given his forwards ten pucks right to the stick where they can finish with regularity, per Dellow's own study.
Comparable Brian Campbell, however, has found his forwards with a direct pass just four times on the year. Remember, Brian Campbell has thirty-nine assists(!), and fifteen at even strength.
Further, Erik Karlsson's playing big minutes with better forwards. Jason Spezza, Milan Michalek, Daniel Alfredsson, and others are clearly feeding more off of EK65's play on the blue line than that of Tomas Fleischmann, Kris Versteeg, Stephen Weiss, and Brian Campbell, respectively.
With more finish from the forwards comes more gaudy assist totals. Seems like a reasonable correlation, no?
Above all else, though, what I can determine in agreement with Tyler Dellow is that assists - to the surprise of few - are fairly random in assessment. It's no different than the National Basketball Association's version of the assist, which is pretty much awarded whenever an isolation play isn't ran.
And, even then, I've seen it tagged on.
In conclusion, Erik Karlsson's total number should be just as impressive as you thought one day ago, if only because this is a problem that plagues the entire league equally. When adjusted for relativity, Erik Karlsson's still that far ahead of the pack.
Thanks for reading!