Photo Credit: The Province

A Tale of Two Penalty Killing Units

The Canucks have just one game in the books in their 2017-18 season, and you’d be surprised how many storylines that one game could generate. There was the Boeser snub, the new playing style, the even strength delpoyment, and the Horvat unit starting power plays, to name a few. Still, there are more things that caught my eye that I felt warranted deeper analysis.

One such aspect was the Canucks’ penalty kill against a powerful Oilers squad whose conversion rate was fifth best in the NHL last season. Since then, Connor McDavid and Leon Draisaitl have only gotten older and scarier. Still, the Oilers were held to but one goal on five man advantage opportunities, and it wasn’t even McDavid’s top unit that converted.

The real story here, for me at least, is the discrepancy in both look and results between two different groups of Vancouver penalty killers.

The PK units were broken into three forward tandems, and two defensive pairs – these stayed largely stable unless one of the penalty killers was the one serving the penalty. The units looked like this:

  • Forwards
    • Brandon Sutter & Derek Dorsett
    • Loui Eriksson & Markus Granlund
    • Bo Horvat & Alex Burmistrov
  • Defence
    • Alex Edler & Chris Tanev
    • Erik Gudbranson & Michael Del Zotto

Because there were more forward groups than defensive groups, the four-man units were often mixed up, though Sutter/Dorsett played most frequently with Gudbranson/Del Zotto, and Eriksson/Grandlund played mostly with Edler/Tanev. Bo Horvat’s time was scaled back and bumped up only when Sutter was sitting in the box, while Burmistrov was hardly used at all. Here’s a PK network, created by Micah Blake McCurdy of HockeyViz.com

The biggest discrepancy (and this should come as no surprise) was between the two defensive pairings. Alex Edler had an excellent night on the PK, while Gudbranson and Del Zotto struggled. The differences in the numbers are jarring.

Player TOI CA CA/60 FA FA/60
CHRIS.TANEV 4.07 4 59.0 4 59.0
ALEX.EDLER 3.37 3 53.4 3 53.4
MICHAEL.DEL ZOTTO 3.58 10 167.6 5 83.8
ERIK.GUDBRANSON 3.58 9 150.8 4 67.0

The rate at which the Oilers were able to generate shot attempts with Gudbranson and Del Zotto on the ice was able triple what they were able to do with Edler and Tanev on the ice. This becomes even more impressive when you see who they were each on the ice with.

Quality of Competition

We also have to be mindful of quality of competition when it comes to short handed time, because it is often the case that two different defensive pairings are out against two different units. This is certainly true in this case, and the results look even better for Edler/Tanev, and even worse for Gudbranson/Del Zotto.

You see, Edler and Tanev were charged with defending Edmonton’s top power play unit – the one containing Connor McDavid, Leon Draisaitl, Milan Lucic, Oscar Klefbom, and Mark Letestu. There are three players on this unit that hit double digits in power play goals last season, and McDavid isn’t even one of them; instead, he was fifth in the league in power play assists.

And yet, the Edler/Tanev tandem completely stifled McDavid’s unit, keeping them to just three shot attempts over three and a third minutes of 4-on-5 time. They did this with a variety of different forwards and, as you’ll see below, Alex Edler deserves a ton of the credit.

Meanwhile, Gudbranson and Del Zotto got shredded by the second unit Patrick Maroon, Ryan Strome, Drake Caggiula, and Ryan Nugent-Hopkins, who ended up scoring a goal against them. The Canucks pairing did manage to block a lot of shots, but were generally drowning in their own zone.

This opposition table (with data from NaturalStatTrick) uses Edler and Gudbranson as proxies for their pairings.

VS Pos TOI With CF CA CF% With FF FA FF% With   VS Pos TOI With CF CA CF% With FF FA FF% With
Connor McDavid C 3.35 2 3 40.0 2 3 40.0 Patrick Maroon L 2.95 1 8 11.1 1 4 20.0
Leon Draisaitl C 3.35 2 3 40.0 2 3 40.0 Ryan Nugent-Hopkins C 2.43 1 7 12.5 1 3 25.0
Mark Letestu C 3.15 2 3 40.0 2 3 40.0 Ryan Strome C 2.33 1 7 12.5 1 3 25.0
Milan Lucic L 3.15 2 3 40.0 2 3 40.0 Drake Caggiula L 2.18 0 7 0.0 0 3 0.0
Oscar Klefbom D 3.12 1 3 25.0 1 3 25.0 Kris Russell D 1.55 0 6 0.0 0 2 0.0
Adam Larsson D 0.22 0 0 0 0 Matthew Benning D 1.33 1 2 33.3 1 2 33.3
Patrick Maroon L 0.22 0 0 0 0 Leon Draisaitl C 1.17 2 2 50.0 2 1 66.7
Darnell Nurse D 0.20 0 0 0 0 Oscar Klefbom D 0.98 2 2 50.0 2 1 66.7
Matthew Benning D 0.05 1 0 100.0 1 0 100.0 Connor McDavid C 0.85 1 2 33.3 1 1 50.0
Drake Caggiula L 0.02 0 0 0 0 Mark Letestu C 0.52 1 1 50.0 1 0 100.0
Ryan Nugent-Hopkins C 0.02 0 0 0 0 Milan Lucic L 0.08 0 1 0.0 0 0
Adam Larsson D 0.03 0 0 0 0

Shorthanded Microstats

I have to say this in every article I write for the first couple of months of a season, and especially after only one game: there is a sample size alert! These are on-ice statistics, which means they are highly influenced by randomness and the play of teammates.

That said, I feel confident that even if these statistics don’t continue at these exact rates, they’ll continue to follow a similar pattern. To back that up, let’s look at something a little more granular.

Hand-tracking individual plays gives a greater degree of confidence in small sample sizes than on-ice stats do, because there are less extraneous variables involved that are likely to influence the results. That is, there’s less randomness in a play where it’s one player handling the puck than there is in one play where a teammate is handling it, in terms of the results we assign to the individual player.

What I did for this game (and hope to continuing doing whenever possible) is track all the PK puck touches for all the Canucks players. I marked down how they came to possess the puck (loose pucks, interceptions, steals), and what they did with it (passes, clears, turnovers). I did this for forwards and defencemen, but for now we’ll just look at the defencemen.

# Player Pos TOI Touches Takeaways Zone Clears Zone Passes Giveaways
23 Alex Edler D 3.37 8 6 3 5 0
8 Chris Tanev D 4.07 5 1 3 0 2
4 Michael Del Zotto D 3.58 4 1 1 1 0
44 Erik Gudbranson D 3.58 2 0 0 1 0

Alex Edler, who has the least 4-on-5 TOI among this group, also had the most touches of the puck. He also had an astonishing number of takeaways, either by intercepting passes, or outright stealing pucks from opponents. He had three zone clears and five clean in-zone passes, with zero giveaways. His partner wasn’t nearly as efficient, which may surprise some. This was likely just an off-night for Tanev, who you wouldn’t normally expect to have more giveaways than takeaways (you can see his most egregious giveaway in the video below).

I’ve gone through the trouble of compiling some of Edler’s shorthanded plays from Saturday’s game. Take the time to appreciate a couple of them, in particular the steals off of Draisaitl and McDavid. He really had a fantastic night, and this should provide some video evidence of the “little things” that Edler often does but doesn’t get credit for, because people often focus on his glaring mental errors.

Then there’s Gudbranson and Del Zotto. To their credit, neither had an outright giveaway, but that’s really only because they hardly ever actually touched the puck. Whereas Edler and Tanev hardly let McDavid’s unit complete a play, Gudbranson and Del Zotto were completely at the mercy of the second unit, and usually only found relief when they were helped by a forward, or the Oilers made a mistake and cleared the zone themselves.

Again, this is only one game, but this isn’t just about the results. It’s about how helpless Gudbranson and Del Zotto looked out there. Their inability to break up plays, or to reach loose pucks really hindered their effectiveness against the noticeably weaker power play unit. Travis Green seemed reluctant to allow either Ben Hutton or Troy Stecher on the ice short handed (Hutton’s only PK time came when Alex Edler was sitting in the penalty box), but if this is what Gudbranson and Del Zotto are going to look like on a regular basis, they might as well given Hutton and Stecher a look.

We’ll see what happens when the Canucks play against Ottawa tonight at 7 pm. I’ll do my best to keep compiling both microstats and video on this as the season progresses. It should provide a very useful window into who excels at penalty killing and who struggles at it, since that can often be hard to quantify.

Opposition stats provided by NaturalStatTrick.com. All other advanced stats provided by Corsica.hockey.

  • Killer Marmot

    Penalty killing is as equally important as the power play, but gets a fraction of the attention.

    The Canucks forward PKs took a big hit when Hansen and Burrows were traded and Gaunce got injured. Management must be scrambling trying to figure out what to do about that. If fans are bewildered when talented young players are not dressed for a game, this is one of the reasons why. The coach can not go into a game without a proper penalty killing squad.

    • My only comment pertaining to the lineup is this: forwards are not the issue for the PK, from what I’ve seen. Sutter, Dorsett, Granlund, and Eriksson are a strong core in that regard.

      The defence though, that’s pretty iffy. Gudbranson and Del Zotto were not good, and Green was reluctan to try Hutton and didn’t try Stecher at all. That doesn’t provide a lot of confidence after Edler and Tanev.

      • Killer Marmot

        The problem is that there’s only four players that are obviously capable penalty killers. Horvat isn’t a good penalty killer and Burmistrov I’m not sure of. This makes it tough to take Dorcett out of the lineup in favour of Boeser. And if any of those four get injured then things get really difficult.

    • SISMIM

      Very true. And perhaps part of the tale of the two PKs involves the WOWYs “with Granlund” versus “without Granlund?”

      Without Granlund, Edler and Tanev posted 0 CF% and FF% on the PK and bled scoring chances against at a similar rate to Gudbranson and Del Zotto.

      With Granlund, Guddy and MDZ posted a respectable 50% CF% and FF% on the PK, and didn’t surrender any chances.

      Small sample size issues abound in single game analyses, but still, it’s interesting that the Granlund WOWYs present a tale of two PKs with little separating the respective performances of our four main PK defensemen, but a whole lot going on with the results driven by the forwards.

      • This is a very thoughtful and well researched comment, thank you for that.

        You’re not wrong about any of that, and I’d noticed it myself. It’s a large reason why I used microstats and video to make my points about Edler having a great PK game. In on-ice stats alone, he was helped by Granlund, certainly. Anytime a player can make a foray up the ice shorthanded, it does wonders for killing time. Dorsett also had a great chance.

        With one of Granlund’s scoring chances (he had two, the post and then the breakaway), he was aided by Tanev breaking up an opposition play. The other he picked off on his own.

        For the most part, I try to keep shorthanded offensive chances separate from defensive analysis. The fact that Edler was 0% without Granlund is true, but not really descriptive. He had all of two shot attempts against in that situation, and a couple of the nice plays in the video in the article were made without Granlund on the ice.

        As far as forwards go though, Granlund was great. He worked very well with Eriksson. Sutter and Dorsett were actually better though, in terms of both takeaways and zone clears.

  • FireGillis

    #FIREGREEN playing your best player (boeser) is more important than penalty killing. I hope that inbred piece of garbage is gone before christmas


      • Vchiu

        I did not XD With him being a swiss army knife and playing any position in any situation I already had him pegged as Hansen 2.0 Thus, he is not allowed to start scoring breakaways until he’s hit the requisite number of posts and stonehand comments. And no powerplay time. Never.

        How’s Dorsett’s chirp game? He’s the closest we have to a pest right now too 😛

  • tyhee

    Even though Edler-Tanev ranking well above MDZ-Guddy as penalty killers would be what one might logically expect, I don’t have confidence in single game stats. Other than that, I think the methodology in this article was excellent and look forward to seeing more analysis from this writer-in two or three months.

    • Agreed on sample size (as I mentioned in the article). That’s why I tried to go as granular as possible with the microstats and video, but it is one game. What we have here is a baseline; now we’ll see whether it continues or not.

    • tyhee

      I wish there was an edit button.

      I overstated my objection when I said “in two or three months.” Analysis of this sort is useful immediately and Jeremy pointed out it is based on a small sample size.

      My only issue was with the reasoning given for the confidence that the small sample trend would continue. That’s the case even though I agree it probably will. My reasons for that aren’t based on stats but on what my individual subjective eye test has concluded about the style and abilities of the individual players involved.

      I also agree with Jeremy that Hutton and Stecher might be worth trying on the pk.

      Again, great analysis imo. I’m just quibbling about one small part.

      • Holly Wood

        Hutton and Stecher struggled in their own end twice in the first period of the Ottawa game, to the point I wondered if they would be broken up in the second. Then in the third Hutton turns it over, it’s in the net. They both can handle the puck so maybe it’s not gonna be as bad as tonight

  • defenceman factory

    Great article and great clips to back up your points.

    Players seldom get much notice on the PK unless they make mistakes. Edler’s somewhat underwhelming achievements in the offensive zone make it easy to underestimate his value. Thanks for providing a better perspective.

  • Beer Can Boyd

    Only 2 games in but Gudbranson looks to be exactly what we thought he was. Too slow for the modern game, and incapable of thinking his way around his limititations. To me, Pedan looked better than him in pre-season, and Gudbranson would have brought more in trade value. Oh well.