Jake Gardiner Is Really Good Defensively

article_aff32ac2-1fc9-4260-a783-2b9358caf4fc
Photo Credit: Kevin Hoffman/USA TODAY SPORTS

If you were to ask people around the NHL what Jake Gardiner is good at, you’d probably hear a lot of things related to his offensive abilities.  Earlier this year James Reimer said that Gardiner has a rare aptitude for passing.  Writing for the Canadian Press last fall, Stephen Whyno described Gardiner as a “young, smooth-skating, offensive-minded” defenceman but added that he “could be an adventure in the defensive zone.”  I think those comments are representative of the general opinion about Jake: he’s a talented skater who’s got great vision, but he hurts his team when he doesn’t have the puck.

Conventional wisdom on Gardiner, however, is wrong.  While he struggled a bit early on in his NHL career, over the past few seasons Gardiner has been one of the most reliable players in the league at reducing the burden faced by his team’s goalie.  The same skills that Gardiner uses to generate offence, like his skating and his vision, help him keep the puck away from his own net.

RELATIVE SHOT ATTEMPTS

The primary way that a defenceman can impact his team defensively is by limiting the number of shots that the other team gets.  By that measure, Gardiner has been extremely good over the past three seasons.  The best way to measure this is Corsi Against per 60 minutes of ice time at even strength, relative to a player’s teammates.  Relative Corsi takes the shot attempts when a player is on the ice and subtracts the shot attempts when he’s on the bench, so it gives us a good idea of how much better (or worse) a team does with a given player compared to his teammates.  For Corsi Against, a low number is better (ex. -5.0 would mean the team allows 5 fewer shot attempts with a player on the ice while +5.0 would mean the team allows 5 shots more).  Here are the top 10 players by CA/60 Rel over the past 3 seasons (minimum 100 games played):

Player GP CA60 Rel
Giordano 207 -8.84
Gardiner 238 -7.92
Ekholm 224 -7.74
Diaz 119 -7.37
Stralman 236 -7.28
Campbell 246 -7.25
Brodie 232 -7.18
Orlov 136 -6.77
Ekman-Larsson 237 -6.4
Tanev 203 -6.08

Only one regular defenceman in the entire NHL has done a better job than Gardiner at limiting shot attempts relative to his teammates over the past three seasons: Mark Giordano.

One of the things that you might have noticed about that list is that it seems to have a lot of players from some teams that have been pretty bad, like the Leafs, Flames, and Coyotes, and almost no players from top teams (Anton Stralman being the notable exception).  There’s a simple explanation for that, which you can probably guess: it’s easier to have a good rating relative to a bad team than a good team.

However, if you go down the list a bit further, you’ll find most of the top defencemen from some of the NHL’s better teams, with Marc-Edouard Vlasic, Kevin Shattenkirk, Victor Hedman, and Drew Doughty all in the top 30.  Keep going a bit further and the top 60 also includes names like Hjalmarsson, Muzzin, Lindholm, Jones, Subban, and Suter.

Because of this bias towards players on bad teams, we shouldn’t interpret CA/60 Rel to be an ordered list of the best defensive players in the NHL.  It’s much harder for a player like Drew Doughty to have a good result relative to the Kings than it is for Chris Tanev to do on the Canucks.  Jake Gardiner is certainly not the 2nd best defensive defenceman in the NHL.  But what the list does give us is players who are, on the whole, providing their teams with good defensive impacts.  Gardiner probably wouldn’t appear at the very top of the list if he played for the LA Kings, but in all likelihood he’d still wind up looking pretty good.

Indeed, not only is he one of the best in the league in aggregate over the past three seasons, Jake Gardiner is also one of the top players in each individual season.  Among players with at least 40 GP, Gardiner was 37th best last year, 3rd the year before, and 7th two years ago.  With 60 top-pair defencemen in the NHL, Jake is easily ranking as a top-pair player by this metric on a consistent basis.

17 players in total have ranked in the top 60 in CA/60 Rel in each of the past three seasons.  Here’s the full list: Matt Niskanen, Calvin De Haan, T.J. Brodie, David Schlemko, Kevin Shattenkirk, Jake Gardiner, Mike Green, Mark Giordano, Ryan Ellis, Marc-Edouard Vlasic, Oliver Ekman-Larsson, Chris Tanev, Barret Jackman, Anton Stralman, Mattias Ekholm, Brian Campbell, Brendan Smith.

You might complain that if we’re putting together a list of best defensive defencemen in the NHL, that list doesn’t look quite right.  While there are some players who are widely recognised for their great defensive play, like Giordano or Vlasic, there are far too many good players missing.  But if you add in the list of players who’ve ranked in the top 60 in two of the past three seasons, most of the names that you’d expect to see show up.  Among others, we can add to the list P.K. Subban, Jake Muzzin, Seth Jones, Victor Hedman, Aaron Ekblad, Niklas Hjalmarsson, Erik Karlsson, and Drew Doughty.

WHAT ABOUT GOALS?

One of the complaints about Corsi is that it doesn’t take shot quality into account.  I think there’s pretty good evidence that defencemen are not able to significantly affect save percentage.  That’s why shot attempts are the best measure of how a defenceman is contributing defensively.  But for the sake of argument, let’s say that defencemen are able to significantly affect SV% and that therefore we should be looking at goals against rather than shots against to determine who is playing well.  Let’s look at the top 10 players in GA/60 Rel over the past 3 seasons (once again limiting to a minimum of 100 GP):

Player GP Rel.GA60
Marincin 150 -0.71
Millar 159 -0.69
Bellemore 113 -0.64
Niskanen 245 -0.57
Campbell 246 -0.45
Larsson 172 -0.45
Liles 179 -0.44
Gryba 185 -0.41
Gardiner 238 -0.37
Tanev 203 -0.36

Well what do you know, Jake Gardiner appears near the top of the league again.  Regardless of whether you use goals or shots, Gardiner winds up looking great.

Looking at this list, though, I think you’ll agree when I say it doesn’t look anything like a list of the best defensive defencemen in the NHL at all.  And it doesn’t get any better if we go further down the list.  Here are the next 20 players:

Stuart 132 -0.36
Dekeyser 223 -0.35
Brodin 218 -0.34
Cole 190 -0.34
Hjalmarsson 244 -0.31
McNabb 164 -0.3
Zidlicky 218 -0.29
Kindl 145 -0.28
McBain 139 -0.28
Krug 238 -0.27
Greene 246 -0.27
Gonchar 124 -0.27
Alzner 246 -0.26
McQuaid 157 -0.25
Michalek 197 -0.24
Lovejoy 204 -0.24
Wiercioch 161 -0.23
Scandella 213 -0.23
Pardy 138 -0.21
Redmond 106 -0.21

This list looks almost entirely random to me.  You can pick out some good players here and there, but on the whole, there isn’t any clear order.  You can easily compare this to the list of players ranked by CA/60, and it’s clear that shot attempts are producing a much better list of good defensive players.

The case that Jake Gardiner is a valuable player defensively seems pretty strong.  Not only is he among the league leaders in reducing shot attempts over the past 3 seasons in total, but he’s also among the league leaders in each of those seasons individually.  He is likely getting a bit of a boost by playing on some pretty bad teams, but on the whole, the evidence says that Gardiner’s defensive impacts should not be overlooked.

  • MacTwoTimes

    I think him changing from wing to defence in his draft year probably didn’t help him as defence is based on experience and essentially learning from mistakes. He’s a solid damn now his defence is in order id love to see him unleashed offensively a bit more.

  • SWZY BSTD

    Can you explain why you used CA/60 REL rather than the just CA/60? Wouldn’t that help make a more direct comparison to defenders who don’t play on the Leafs?

  • DragLikePull

    Using Rel lets us adjust for team quality somewhat. If we used CA/60 rather than CA/60 Rel, we’d just get a list of players on good teams.

    For example, Luke Schenn and Rob Scuderi both had much lower CA/60 on the Kings last season than Giordano had on the Flames. That’s because the Kings are a really good team and Scuderi/Schenn are being boosted by their teammates. Giordano is far more important to his team’s results, which is indicated by his better Rel stats, but because he plays on a bad team his non-Rel numbers are lower.

    • Trevor5555

      Its a bit like the chicken and the egg though….you say using CA/60 just gives you a list of players on good teams, much like the criticizim of +/- , but the fact is that L.A. wouldnt be regarded as such a good team if it didnt have those good d-men.

      The problem is that we seem to want one magic stat that lists all the best d-men rather than accepting that each stat is just a small piece of info representing a players overall effectiveness.

      I think raw CA/60 gives a better idea of which d-men are truely the most effective shot suppressers as using the rel stats just shows how they compare to teammates. CA/60 seems more meaningful than CA/60 REL in that sense. I would go further to say that CF% takes into account positive and negative corsi events and gives a measure that represents both offensive and defensive minded d-men on a level playing field. Ottawas Karlson might allow more chances against but his offensive skills more than account for them and overall he has a positive impact on his teams possesion stats. But I agree there is value in seperating defensive impact and offensive impact as they are two distinct skillsets.

      Its nice to see that elements of our blueline seem to be improving. Im hopeful for the season to come.

  • JB#1

    I’ve really been coming around the past few years to appreciating Gardiner and thank God the Leafs didn’t jettison him a few years back in another very Leaf’ian ill-advised trade.

    Question.

    You said “Gardiner probably wouldn’t appear at the very top of the list if he played for the LA Kings, but in all likelihood he’d still wind up looking pretty good.”

    If I’m interpreting your article correctly, it seems to me that as the Leafs get better, Gardiner should drop down the CA/60 Rel list. If he does, this would be considered a good thing, yes?

    When the Leafs become better, if Gardiner stays near the top of this list, would this elevate him into very good to elite status? Or would this be considered a bad thing?

  • Stan Smith

    Question. When ranking dmen defensively would it not make more sense to use Fenwick instead of Corsi? After all, is play and positioning in your own zone not important in figuring this out?

  • DragLikePull

    If you use Fenwick the list re-orders a bit, but you still get all the same names near the top, so it doesn’t really make much difference once you get to a large sample like 3 seasons and a couple hundred games.

    Also, Corsi has generally been shown to have better predictive value than Fenwick, which means it is a more repeatable skill. This makes sense, because it’s essentially the same thing, but with an increased sample size.

  • Drapes55

    I don’t think you can say Gardiner is really good defensively because these numbers say so. I’m all for introducing fancy stats to help with evaluating players but you can’t rely on them unless the eye test also backs up you’re claim, and in this case I have to say you’re claim is wrong. Jake is a good middle pairing defender with lots of offensive upside. Defensively he’s average, the numbers may not agree but when you compare the competition he plays against on a nightly basis compared to the guys like Gio, Keith, Doughty, Karlsson, etc… Jake’s numbers don’t impress me as much. If he was putting these numbers up playing 25-30 mins a night against the top lines of each team then I’d agree but the fact is he doesn’t and when you actually watch him you would see that he is far from “really good” defensively.

  • Drapes55

    By no means am I ignoring the facts, I think Gardiner is a good defenseman, probably one of the best second pairing guys in the League hands down. I just don’t see the really good defensive play that the stats show. I think allot of this is due to the fact that when I think of defensive play I think of defending when the other team has the puck. Gardiner’s style of defence is more defence by offence in the sense that he gives up less shots by not allowing the other team to have the puck as long, and there is absolutely nothing wrong with that. When I say he’s average defensively I mean defending without the puck, I find he lacks a bit of physicality when defending and if he can implement that into his game I think he could be an elite defenseman in this league someday.

  • scotts2488

    can I have some of the drugs you are taking? Gardiner can’t even body check or display any sense of urgency in his own zone and you think he is good defensively….