Welcome back for the second in our looks at the Leafs’ players and their performances this season. Arvind has passed the baton to me for the next segment in our series, making this the world’s least physically demanding relay. Today we’re looking at the defenders.
As with the previous edition, which you can read here, we are heavily indebted to gentleman scholar and friend of the blog Alan (@loserpoints), who devised these awesome at-a-glance charts and who has generously assisted us in the production of this series. The explanation of the metrics used in this evaluation is below—same as with the forwards. If you’re already hip to that, look below for the article proper.
Most metrics are relative to team performance. They show how the team performs in that metric when the player is on the ice compared to when they are not. For example, relative shots per 60 minutes is shots per 60 when the player is on the ice minus shots per 60 when the player is off the ice. All data is 5v5 only and adjusted for score, venue, and zone starts via Corsica. Numbers are presented as percentiles comparing each player to others at the same positions. For forwards, this includes the top 390 players in ice time in 2016-2017. Percentiles indicate the percentage of observations within a sample that fall below a given value. For example, let’s say I have a friend named Joey, who owns a Rattata. Joey claims that his Rattata is in the ‘top percentile’ of Rattata’s. Essentially, he is saying that his Rattata is in the 99th percentile of Rattata’s, or that 99% of Rattata’s in the world measure below his. So similarly, if I say Auston Matthews is in the 98th percentile for Game Score, that means that 98% of players rank below him in that metric. Throughout this piece, being at the high end of the percentile is good, and being at the low end is bad.
(P1/60): goals and primary assists per 60 minutes of ice time
Rel.ShF/60: Relative shots for per 60 minutes of ice time
Rel.xSh%: Relative expected shooting percentage. This is a measure of shot danger.
Rel.xGF/60: Relative expected goals for per 60 minutes of ice time
Rel.ShA/60: Relative shots against per 60 minutes of ice time
Rel.xSv%: Relative expected save percentage. This is a measure of shot danger.
Rel.xGA/60: Relative expected goals against per 60 minutes of ice time
Rel.Sh.Share: Relative shot share. Similar idea to Corsi Rel.
Rel.xG.Share: Relative expected goal share. Same as above, but expected goals
So, let’s get rolling.
I love Jake Gardiner more than most people I actually know, and nothing in this chart serves to dampen those feelings. As you’d expect if you’ve looked at his numbers before, Jake Gardiner generates shots like a monster (keep in mind these relative rates of his are being compared to a team that as a whole shoots at the rate of a machine gun, so they may actually undersell just how great he is at that.) He’s also a very productive offensive defenceman. This chart shows his even strength scoring, but he’s also adept on the powerplay.
As expected, Jake is best at the offensive end. He gives up slightly more dangerous chances than we’d wish. He’s not a bad defensive defenceman by any stretch—at least, in net impact, he’s a clear and impressive positive, and by some metrics (Game Score stands out) he’s a #1. The popular rebuttal to this has always been his quality of competition, but towards the end of this year, he moved towards being the all-around first defender on the Leaf blueline, and acquitted himself well. You can say the Leafs need to try to add a genuinely elite top defenceman. But Jake Gardiner is a damn good one.
For a few years now, Morgan Rielly has seemed--at least to the analytics crowd—to be a glass cannon. Incredible offensive firepower, and terrible defensive fragility. Well, that still seems about right—that shots against number continues to be a disappointment, but the pretty offensive numbers more or less balance it out or even outweigh it (again, game score).
Maybe the oddest thing that immediately jumps out about Rielly: his atrocious expected shooting percentage. The easiest guess is that Rielly is suffering in quality of competition terms: he spent much of the year on the shutdown pairing, and so the other defence pairings got featherbed usage from which they could generate quality opportunities. Rielly’s actual personal shooting percentage this year was atrocious (3.5%), but his even strength production was overall still quite respectable. I don’t sweat this too much.
Putting that aside, we come back again to the glass cannon issue. Leaf fans have hoped for a long time that Rielly would become that elite #1 d-man his athleticism seemed to suppose, and...well...he isn’t. He’s good, and I’m very happy having him in my top four, but him taking the toughest defensive minutes is a product of necessity and hope rather than a natural fit. A team that can play Rielly safely on its second pair is well-positioned to wreck its opponents. A team that has to play him on its first is going to have some bumpy road ahead of it.
Jeez, I dunno guys.
Nikita Zaitsev visibly has a lot of talent, and he was always playing the very toughest minutes, all year long. He had to come over from Russia this year. He’s a tremendous passer, and this chart (since it’s focused mostly on 5v5) doesn’t really show his quality work on the power play.
But there’s no getting around it: all the numerical evidence suggests that Zaitsev is getting killed defensively. Playing with Morgan Rielly does no one’s shots-against numbers any favours, and it’s never fair to blame one player for defensive issues like these. To borrow a quote from Katya, it takes a village to be this bad. But I don’t think you can make a tenable case that Nikita Zaitsev is a successful top-pairing defenceman based on the numbers, and you can make a case—one in which that xGA figures prominently—that he’s barely even an average one in net value. That’s a scary thought going into a seven-year deal.
Okay. Let’s dispel the gloom a little. Zaitsev’s skills are obvious, his point production is good, he shoots right, and the Leafs’ forwards have a lot of room to improve defensively. There’s probably some hope that a top four of:
Jake Gardiner - Unicorn 1D
Morgan Rielly - Nikita Zaitsev
Would work quite nicely. So please, someone, lift the weight from Zaitsev’s right-shooting shoulder, and maybe this will all turn out okay.
If you’re wondering whose numbers shocked me the most in the course of this piece, I’ll tell you now: it was Connor Carrick, by a huge margin.
Connor Carrick bounced around the right side of the lineup at some points this year. Carrick also clearly got a bonus in his relative defensive stats by virtue of playing on a team with, but not on a pairing with, shots-against victims Rielly and Zaitsev. So that’s my best guess.
Because this number profile suggests he produces virtually no offence, and is actually well above average defensively. This is compared to his team, remember—that team that was so potent in scoring and so terrible in being scored on—so it makes some sense that these numbers are the reverse of his HERO chart values, which are based on the whole league’s average. Those show him as producing a lot of shots for and allowing a lot against.
Still, this is very interesting. Maybe most of all, it suggests something interesting about that super-potent Jake Gardiner and Connor Carrick pairing from early this year; Carrick may have been doing more than expected on the defensive end. Or put another way—Carrick and Rielly would seem like good bets to balance each other out a little. It’s a little hard to think of Carrick as a defensive d-man, but he’s feisty and he’ll fight for space. Maybe there’s something more there than we thought.
It’s time for me to own up: I was very harsh on Matt Hunwick earlier this year, and I was wrong.
Hunwick is genuinely doing worthwhile defensive work, cutting down heavily on shots against, while producing points at a surprisingly decent clip (although whether you want to rely on the future of that odd shots for/expected shooting percentage split is up to you.) The Leafs don’t shoot much when he’s on the ice, compared to their normal machine gun rate, but they also put up a much better defensive showing. In defence of my earlier condemnation of him, Hunwick’s shot numbers got markedly better as the year went on. But he’s also one of those players that expected goals clearly marks out as doing genuinely useful d-zone stuff. As long as the Leafs aren’t giving up any kind of term to do it, I would be happy to have Matt Hunwick return to our third pair next season.
I should preface this by saying: Roman Polak seems like a mensch. By all accounts he’s a good guy, as well as a tough and hard-working one, and he was once a pretty solid defenceman. His cringe-inducing injury in the playoffs came when he was playing his best hockey of the year, and it sucks. I hope for the best for him personally.
I just don’t think you can do as badly in shots against as he usually does and hold down an NHL job. Expected goals still thinks he’s okay, and certainly shot differentials aren’t the whole story. Polak is a walking shot quality argument (despite those heavy slapshots he wires off that don’t seem to ever achieve much.) But he gets submarined in shots against, and it’s much, much worse than anyone else in the defence group, including his regular partner Hunwick. That’s pretty tough to deal with.
Having said that; as with Hunwick, Polak improved as the year went on. Unfortunately, that was before an injury that threatened his already questionable mobility. If the Leafs are operating coldly, they shouldn’t re-sign Polak (at least as a player—he might make a pretty fine consultant in the Robidas style, if they want another one.) I’ll understand why they might want to bring him back, but at best, I don’t think he’s more than a 7D now.
I have long been a card-carrying member of the Marmar Fan Club—actually, at this point, I’m not sure if there’s anyone else left in it besides me and Arvind—and it’s because of numbers like these. Martin Marincin does everything at an above average level except produce points, and yet people who go by the eye test loathe him. Mike Babcock seemed unimpressed with his work for much of this year, but Marmar reappeared a penalty-killer when push came to shove. His work against the guillotine-like Washington power play was bordering on heroic at times.
There are really only two ways to go with Marincin: either he’s a sneakily good defensive defenceman on net value—I’m not defending the occasional brain fart he has where he gifts the puck away—or the numbers are all wrong. Maybe the stats overrate Marincin, but I find it very hard to believe he can do so many things that lead to consistently good underlying numbers and be totally useless. Hell, his relative shot share is the best on the team. And Marincin is a left-hand defender who can survive playing right side. If the Leafs still have Marincin next season, and I think they will, they might make some decent use of him yet.