The most easily predictable outcome for the Leafs as they started this season was that Mike Babcock would play his top four defenders exactly like last year. The only change would be to add in someone with Travis Dermott on the third pair. It wasn’t a popular prediction in some circles, but things played out as expected.
The next most easily predicted outcome was that every mistake Ron Hainsey made would loom big, while Travis Dermott would be forgiven nearly everything, and the demands would start immediately to have them switched. That’s playing out as expected as well, reality be damned.
Reality, however, is often very interesting. If you look, you will see a carefully managed usage of both players that I expect to result in Dermott gradually walking from the wading end of the pool to the deep end, and somewhere in the middle, he’ll meet Hainsey coming back the other way. They are already closer than you think, if all you’ve been looking at is the lineup chart.
Let’s jump to last year and see how they were used during the period of the regular season that Dermott was on the roster. He started in early January and played most of the games. He started out playing a lot, but his time on ice diminished at the very end of the season as the team was gearing up for the playoffs.
2017-2018 from January 6 to the end of the regular season
What the average time on ice shows over that time is that Gardiner and Zaitsev (even though he was recovering from an injury) were the top pair, Hainsey and Rielly played less and the third pair, no matter who was on it, played 15 minutes and change per game, on average. Dermott has some 17 - 19 minute games in there.
For the results over those games, I’m going to use Offside Review’s data, which includes their own Expected Goals model. I’m planning on using this site for several projects over the season.
In this time period, Dermott has a Relative Expected Goals percentage, a Relative Corsi For percentage and a Relative Goals For percentage well above any other defender. It’s not really even worth assembling the numbers to see. He was stunningly successful at shot share, shot quality and goal results.
This overwhelming success in results has led more than one person, many of whom should know better, to decide Dermott is the best defenceman on the team, and should totally have been top pair from day one. If we dig into the numbers before we move on to this season, we can see why that’s a touch premature.
If you break down the Expected Goals to its For and Against parts, something interesting is revealed. Both Carrick and Dermott have really high relative Expected Goals Against per 60 minutes. The jump between them and the next best, Jake Gardiner, is large. This half of the equation accounts for most of Dermott’s excellent overall percentage results.
The Relative Expected Goals For per 60 has the two defenders you should expect to see at the top. And no, it’s not the two you think are best or you like most. It’s the pair who played the most with the most successful offensive lines: Morgan Rielly and Hainsey. Dermott is next, but he’s so far below Hainsey, he’s basically a hair over team average with Zaitsev, Carrick and Gardiner a hair below.
Roman Polak, by the way, is about at team average in Expected Goals Against, and is abysmal at Expected Goals For.
So, this should open up a whole host of questions about Dermott’s results, and how much of his defensive success was playing with two other defensively successful defenders with forward lines who were not the hopeless-in-their-own-end sorts that Gardiner and Zaitsev spent quality time with.
Teammates have the biggest effect on results over any other external factor. The fact that Hainsey played much, much tougher competition than Dermott did is not really that meaningful, although the skew is big enough to have some effect. It’s the rest of the Leafs that matter more. But all the aspects of usage have some effect.
Let’s talk about sheltering. I hate that term because it’s used to moralize against players. Oh, he was sheltered, he’s not tough enough for the hard minutes. Because every top-scoring player needs to start in the defensive zone against the Sidney Crosby line every night, or he’s not really elite, right? Bah.
Sheltering has been traditionally seen as related to zone starts. And traditions are hard to shake. Most shifts start in the neutral zone, and zone starts tell you more about what the coach thinks the player is good at than anything else.
The 2017-2018 zone usage for Leafs defenders was very simple: All of the top four got more defensive zone usage than any of the third pair, and all about the same as each other. The third pairing players were all mostly given neutral zone starts, with Dermott and Carrick in particular shading a little to some offensive usage.
Babcock thought his third pair wasn’t all that hot in critical defensive situations, and his top four played the hard shifts. He also thought all of them not named Polak might have something to offer offensively in the right situation.
The real issue with sheltered usage is that the third pair, particularly Carrick and Dermott always played with forwards who were better than the opposition forwards. And Dermott was given a lot of ride-along shifts with the Tyler Bozak line or the lines centred by Patrick Marleau while Auston Matthews was out hurt. Those lines were carefully deployed to keep them on the right side of the ice.
My take from all of that is this: Dermott showed signs of incredible talent and at a level far above Carrick. Both players, when given all the help in the world to succeed offensively weren’t adding much to the mix. When given some help to succeed defensively in fairly easy situations, they really took it and ran.
Hainsey, on the other hand, might as well have been on a different team. He played with the worst defender defensively in the toughest situations, but he was spared some of the most defensively inept forwards. The overall Expected Goals percentage was just over team average for Rielly and just under for Hainsey. We’ve conclusively proved he’s not quite as good offensively as Rielly, or as bad defensively, and not much else.
One other thing to note about the heavily positive perception of Dermott: He happened to be lucky enough to be on the ice for the best goaltending of any Leafs player, and that always helps you look good.
That’s where they were, now let’s have a quick look at this year so far, where, so far, the perception is that Dermott is being held back and Hainsey is really bad.
Leaving aside Martin Marincin, who only played one game, let’s have a full set of stats for the defenders so far, again from Offside Review:
2018-2019 defenders to October 9
|Player||GP||TOI||TOI/GP||Rel xGF%||Rel CF%||Rel xGF60||Rel xGA60|
Just judging by ice time, it looks like Dermott is getting some top four minutes and they’re coming from Hainsey’s share of the pie. One goal of tracking these two this year is to monitor that to see if it changes and if Dermott’s ice time increases.
In terms of Expected Goals, Dermott is great, but so is Hainsey. And this is something else to monitor: Will Hainsey’s results improve if and when his usage changes? Hainsey is getting most of his good results from the Expected Goals For side, but his Expected Goals Against is also good (these are relative numbers so in the Against column, a negative is good). The only really bad defenders defensively are exactly who we should expect: Ozhiganov, who isn’t ready to have his stats looked at, and Rielly.
Dermott is showing very well on both sides of the ledger, particularly if we recognize that no one is likely to touch Rielly in Expected Goals For. But the bulk of his success is on the defensive side of the ledger, just like last season.
In terms of zone deployment and other usage, Rielly and Hainsey are getting some more offensive starts this season, so Babcock thinks they can be most effective that way. Dermott is right about where he was last year, so he’s not spending much time in difficult or high-leverage situations defensively.
Dermott is getting much more time with the higher-class forwards than Ozhiganov, and spending time with top four defence partners. He has had the best success behind Nazem Kadri, and has been on the ice for some good results with Auston Matthews, although the shots against shoot up with him compared to with Kadri, which is to be expected. Dermott hasn’t played with the fourth line much at all *see correction below.
Hainsey’s usage is separating from Rielly’s in a way it did not last season. He is getting less time proportionally with the top forwards, and there has been some interesting pairing mixing going on. One constant is that the worst forwards on the team always get played with Jake Gardiner and Nikita Zaitsev.
Right now, Hainsey is getting the job done and still looks like he’s not as good as Rielly at most things, but is better defensively. Dermott looks like he’s just barely moving into the top four deep end of the pool with all the indicator lights showing green — another step forward would be a good idea. One thing to note is that while Dermott is slowly being immersed in the NHL, he’s also being relied on to teach Ozhiganov everything he needs to know on the ice. That level of coach’s trust is not something every guy his age gets.
At some point after a few more than just four games, we’ll look at this again and see what if anything has changed and how close together they are.
After I published this, Travis Yost published an analysis of ice time usage that contained this chart:
So, just by looking at the on-ice together numbers at Natural Stat Trick, I underestimated the percentage of fourth line minutes Dermott was seeing. But the more interesting thing is that you can run down the cart and see where Hainsey’s time with a given forward is less than Rielly’s and see a corresponding jump in Dermott’s percentage time with that player.
In selected situations, Dermott has been taking his shifts. He’s also seen a bit of time with Hainsey himself as well as Nikita Zaitsev. And you can also see how much Tavares time Hainsey and Rielly are getting.
Read the full story because he doesn’t think the defensive pairings are going to change this season much, and I, obviously, think he’s wrong.