On Saturday night in Detroit, the Leafs won a game almost entirely because of their depth forwards. This is not something we expected coming into the year. The Leafs under-invest in their bottom six, simply because of the fact that they invest so heavily in their top forwards. Cuts have to come from somewhere, and Kyle Dubas has made the conscious decision for those to come at the expense of the Leafs depth forwards.
So was that win a one-off, or has the expensive upper end of the Leafs lineup failed to deliver?
There is now a microscope on the quartet of John Tavares, Mitch Marner, Auston Matthews, and William Nylander. They’ve all gotten paid, and now it’s time for them to deliver. With that in mind, I figured it would be useful to contextualize how the top two lines have performed this year, relative to their Leafs counterparts from last season. Obviously, it’s a tiny sample size thus far, so we’re very much looking what the lines have done, as opposed to who they are from a ‘true talent’ perspective. There’s a bit of room for interpretation in determining the best 2018/19 line (or set of lines) to compare the current year’s lineup to. In general, we’ll try to match them up by centre, since the identity of Leafs lines last year were inherently tied to their pivot. Let’s get started.
EDIT: I wrote this article before this thread from EvolvingWild, who run the immensely valuable evolving-hockey.com. In short, the shot location data the NHL is providing this season is, to put it politely, messed up. The upshot of it is that there are essentially no shots recorded in the net-front area this year, which would partially (or perhaps entirely) explain the results I discuss in the article. At this point, it’s unclear to what degree this issue is explained by data quality versus actual changes in the Leafs play. The rest of the article is unchanged, for posterity.
Line 1: Kasperi Kapanen - John Tavares - Mitch Marner
Line 1 Comparison: Zach Hyman - John Tavares - Mitch Marner
Last year, the trio of Zach Hyman, John Tavares, and Mitch Marner was among the most productive lines in the league. This year, Zach Hyman’s injury has forced Kasperi Kapanen into a role on the left wing, alongside Tavares and Marner. Nonetheless, Hyman was (and is) seen as the most dispensable part of that trio; the salary structure of that line suggests as much. As such, the expectation was that this line with Kapanen would more or less pick up where they left off last season with Hyman (though the smart money was always on them scoring fewer points this season than they did in 2018/19). Even if one did expect regression in their point totals, this is supposed to be a top-10 line in the league, one that consistently outshoots, out-chances, and outscores its opponents. However, they have not recaptured the glory of 2018/19 thus far. It’s easy to draw a line between Hyman’s absence and Kapanen’s presence and make conclusions based on that.
Hyman certainly has attributes that Kapanen lacks. Hyman’s board work, discipline, and athleticism synergized perfectly with his two more traditionally talented linemates. From a fit perspective, it doesn’t look as though Kapanen can replicate those tools, and the value he provides above Hyman in terms of speed, puck skills, and finishing talent have not manifested themselves on the scoresheet. However, we can compare these lines in slightly more sophisticated ways than point totals. First, lets take a look at the shot results for these two lines, from last season to this season.
Perhaps surprisingly, the 2019/20 iteration of this line is generating a similar amount of shot attempts, and surrendering fewer. If we look at unblocked shot attempts, they are taking fewer and surrendering fewer, but their proportion of unblocked shot attempts is higher this year compared to last. So what is the problem?
Well that makes it pretty clear. Despite their genuinely solid Corsi results, Line 1 has quite simply, gotten their ass kicked on the scoreboard. They’re generating no offense and are getting scored on for fun. There is obviously a component of bad luck here; we can see that their GF/60 vastly trails their xGF/60 and their GA/60 vastly exceeds their xGA/60. However, there are also clear issues that arise when we look at their expected goals rates. While Line 1 has cut down on shot attempts against, their expected goals against is identical to last year. And unlike last year, they’re not generating offense to compensate. For context, last season, Tavares’ on-ice expected goals for rate put him 16th in the league. This year, the rate he’s achieving is close to 4th line territory.
Given that they’re generating the same amount of shot attempts as last year, but nowhere near the amount of expected goals, the obvious conclusion is that they’re shooting from bad locations. That turns out to be exactly the case.
Note the absence of net-front shots in comparison to last season. Getting to that area is a Hyman specialty, but he is not solely responsible for it. John Tavares has made his livelihood scoring goals from that area of the ice, and last year was his best in that regard. For whatever reason, that’s not happening this season.
Contextually, this line is being used in a fairly similar way to last year. They’re still getting a lot of minutes as well as the toughest competition of any line on the Leafs. They’re just far less successful at generating offense. Considering this line employs two players who we’re paying to be elite offensive contributors, that has to change, quick.
Line 2: Andreas Johnsson - Auston Matthews - William Nylander
Line 2 Comparison: All Auston Matthews lines from 2018/19
On the other hand, this line has been incredibly strong throughout the early parts of this season. Last year, Matthews had an underwhelming year from an on-ice results perspective, a function of a rotating cast of linemates (none of whom were as talented as William Nylander) and the incomplete nature of Matthews’ game. This season, the Matthews line has seen improved on-ice results everywhere.
In the first six games of the year, the Matthews line has increased their shot output and drastically cut back on the amount of shots they give up, to the tune of a vastly improved shot share (both for all shots, and unblocked shots only). It’s worth noting that last year, Matthews did not have the linemates commensurate with the expectations placed on this line. As such, the comparison to his groups last year is a bit of a low bar. It’s the bare minimum for the Johnsson-Matthews-Nylander to exceed what Matthews did with the motley crew of wingers he had last year.
When we look at actual results, Line 2 has converted shots to goals at a much higher rate than expected (some overperformance is typical for Matthews, but not to this degree), and has also seen opponents do the same. However, like Line 1, Line 2 has seen a significant decline in their ability to generate quality chances. To put this in context, Matthews’ expected goals for rate last year ranked 23rd league-wide (again, this is with relatively mediocre linemates). The expected goals for rate of Line 2 this season would put them around 170th among forwards league-wide. It’s a very significant difference. What works in Line 2’s favour is that unlike Line 1, they’ve also cut down on chances against, and as such, their expected goal share is still very strong (about 58%, depending on the model used).
It’s unclear to me whether this offensive reduction is coached, a product of small samples, or something else entirely. Over their first three years in the NHL, Matthews and Nylander has meant elite offense, both in terms of expected goals and actual goals. From 2016/17 to 2018/19, Matthews and Nylander together have put up 2.84 xGF/60 and 3.32 GF/60 at 5v5. Both are strong figures, particularly the expected goals tally. In fact, it’s higher than other famous NHL dynamic duos like Steven Stamkos and Nikita Kucherov, and is about the same as the expected goal rate of Connor McDavid and Leon Draisaitl (the gold standard of offensive powerhouses). It is worth noting that both of those duos score (actual goals) at higher rates than Matthews/Nylander, which is primarily due to Nylander’s lack of individual shooting talent on the level of the other five players mentioned in this paragraph.
Regardless, to see Matthews and Nylander in the low 2’s for expected goal rate is a little alarming, especially given the competition that they’re facing thus far. Part of this is tilted by the Leafs having four home games, but Line 2 has faced weaker forwards than the Leafs’ nominal third line (see image below).
Time will tell if this is the new normal for this particular line. As it stands, their results have been strong, but in a way that we wouldn’t have expected coming into this season. If they can maintain this level of defensive performance while reverting to prior years of offense, then they’ll be one of the best lines in hockey. However, that is a lot to ask of this group.
All statistics are from Natural Stat Trick, and are 5v5 score and venue adjusted. The heat maps are from HockeyViz.com.