What’s wrong with Bozak?

I had this wild thought in the middle of the game against the Flyers on Saturday: Should they even take Tyler Bozak on the road trip?

It was a bit shocking to see Mike Babcock play Bozak for a few shifts in the third period with Dominic Moore and Kasperi Kapanen on the fourth line.  But the earlier demotion of Mitch Marner was, to me, as much about Bozak as it was Marner.  Bozak saw his minutes drop drastically for a couple of games.

The problem with dealing with a struggling Bozak has been lack of centre depth.  There are more centres than there used to be, sure, but the training camp drama of who was going to get the fourth line job obscured that there are a lot of fourth-line centres, and none of them have shown the ability to play up higher than that, at least not yet.

To knock Bozak down Saturday night, Babcock had to play Patrick Marleau at centre, something he said afterward he didn't want to do.

But my frustration with Bozak's play so far this year, and that game in particular, was enough that I formed the thought in my mind: “Why play him at all?” When I asked the PPP chat if that idea was as off the wall as I thought it was, I got less fight than I expected.  And then as we tend to do, we tried to argue ourselves out of the idea. Or rather, test that assertion to see if it held up. Let's do that here.

Disclaimer: An 11-game sample of play can have a lot of random variance in it. There is no guarantee that this accurate measure of what has occurred is the road map for what is to come.

The Measure of the Man

I don't want to measure Bozak against Auston Matthews, an imaginary average player, or some ideal of perfection. He's an unusual centre in a lot of ways. There are not a lot of pass-first centres as extreme as he is at getting masses of points without shooting much.  He's also fairly extreme in his defensive shortcomings.

Normally, I'd like to compare him against himself.  The problem with that is that the Leafs, the only team he's ever played on, are so radically different now, it doesn't seem all that useful to compare Bozak’s results now to what he did playing for Randy Carlyle.

One of the things I said Saturday night was that I was beginning to wonder if he had a twin brother who had played the last two years under Babcock. Was that all a fluke, I wondered. Okay then, there are three Tyler Bozaks: his historical pre-Babcock record, his last two years, and his frustrating first 11 games of 2017-2018.

Tyler Bozak’s Career Corsi Results

MeasureHistorical BozakBabcock's BozakCurrent Bozak
CF%46.251.046.7
CF6056.962.653.4
CA6066.460.261.0

The obvious thing is that his Corsi For percentage has crashed back to his historical average.  But the odd thing is how it's that bad. His Corsi For has tanked, not his Corsi Against.  In fact, the against, his on-ice defensive results in a way, is exactly as good as it was the first two years under Babcock.

Relative to his team, it gets weirder. Bozak has never had a negative Relative Corsi For per 60 in his career.  That is, he has never been below his team's average in rate of shots for.  This year, he's a -7.79 and the only forwards worse than him are Eric Fehr, Matt Martin, Dominic Moore and Connor Brown.

James van Riemsdyk is just at the average mark, or zero, and Mitch Marner is at -3.83, around where the Kadri line is all grouped.  So Bozak isn't just struggling to produce a good volume of shots offensively, he's worse than his linemates by a considerable amount.

Looking at shots in total, there's just less going on on the ice around Bozak right now.  His total pace of play (Corst For per 60 + Corsi Against per 60) was the same historically as under Babcock.  Now it's markedly lower.

That's his on-ice shot environment described. Let's now look at his personal shooting. His individual shots per 60 minutes (Corsi, not SOG) went from 7.5 to 9.5 under Babcock, and is 8.2 this year.  That's not much change, and that makes it seem like at least in terms of the shots he does take, he's consistent.

Look Out Behind You!

I'm going to avoid looking too deeply into anything else he's doing or experiencing defensively. There's been no change there, but I will point out one reason why it seems like he has been horrible in his own end: On-Ice Save Percentage.  That is: the Save Percentage of the goalies while he's on the ice.  And while Save Percentage isn't 100% the goalie, an On-Ice Save Percentage is also never the "fault" of one skater.

Bozak's  career average On-Ice Save Percentage is .913, and it spiked in the Babcock years because the goaltending in general improved. Right now it's .842.  That kind of blip in on-ice goalie performance can skew the eyetest in dramatic fashion. And coaches are not immune to it.

Here's a graphic representation of the Leafs' on-ice shooting and save percentages.

Note that Bozak's line is in the basement, but for the fourth line, in shots that go in the net for the other team. But their own scoring rate is better than league average. Bozak's On-Ice Shooting Percentage is over his career average right now.  If he's been unlucky, it's been defensively.

Expectations Need to be Met

Just to confirm that Bozak’s issues are not really defensive, I took a quick look at his Expected Goals Against. (This is all done filtered to a minimum 50 minutes on ice, so the players who've barely played are not included.)

Bozak is in the middle of the pack for the Leafs team for xGA60, about equal to the fourth liners, just a touch worse than Kadri's line and better than Matthews and Nylander, who are the worst forwards on the team. The best player on the team by this measure? Mitch Marner. So let's look at that on-ice goaltending chart again and think some more about the eye-test and runs of bad luck.

But while we're in expected goals territory, let me just point out that Bozak's Expected Goals For per 60 is second lowest, with only Connor Brown worse.  That's not personal expected goals, that's on-ice.   So when Bozak is on the ice, you're as likely to see a scoring chance that should go in as you are when the fourth line is. But on the bright side, he's been better defensively than almost all of them.

And to dive just a touch deeper into this expected goal weighting for shots, this time looking at offence, Bozak's Expected Fenwick Shooting Percentage (the chance of an unblocked shot being a goal, on average) is less than his actual. Again, this is on-ice amounts, but that means when he's on the ice, the Leafs are doing better than a set of average skaters would at scoring goals with the same types and locations of shots.

Shot Location

What has declined in Bozak's results is the pace of his shots for, and nothing else. Or at least nothing else findable in these numbers.  Let's be clear here, he's not good defensively. He isn't gifted at it, and if you've seen him make a hash of some play in the defensive zone, you are not alone.  But he's no worse than he ever was.

So why aren't those shots hapening?

Let's start with the Leafs as a team. This is their team shot heat map. Red is more than league average, blue is less.

This is the Leafs’ power and their glory made into red and blue blobs. This is every win this team has. This is every time you thought they were actually great.  All that shooting in the slot and the circles is how you win hockey games.

Auston Matthews' on-ice chart matches the team’s the closest. As his line goes, so goes the Leafs. But he and William Nylander and Zach Hyman play this way — shots from in tight, not much from the points, nothing from the outside.

Tyler Bozak's looks like this:

It's fair to say this is dramatically different. You can almost see van Riemsdyk camped on the right corner of the net and Marner shooting from the right circle.  This leaves things open on the left side, so it's not a huge surprise that the left defender seems to be swooping in to shoot more.

That means a couple of things. If the defender is shooting more, pinching in more to do it, then the blue line is being covered by his partner and likely Bozak or Marner.  It also means a lot fewer quality shots are taking place.

This is fairly consistent with the patterns from last season when Bozak had the same linemates most of the time.  This is last year's:

(Note: this is subscriber content. If you want to regularly see past year’s charts, subscribe to Hockey Viz.)

While the team plot from last year is broadly the same pattern as this year’s, although not as good, Bozak’s pattern is similar too.  All that left defender shooting is still there.

Which left defender exactly?

Glad you asked.  Last year Bozak was on the ice for almost 400 minutes with Jake Gardiner.  He also played about 300 minutes each with Matt Hunwick and Morgan Rielly.

This year, Bozak’s most frequent left defender is Andreas Borgman. If you add on the minutes he was on the ice with Calle Rosen to that total, you get 68 minutes with the rookies.  Gardiner joined him for 41 and Rielly for 31.

In order to give his young, rookie defenders some lighter competition, Babcock sent them out behind the forward line that he gave lighter, offensively-focused minutes to.  In addition, Gardiner and Zaitsev have put up some very uncharacteristically dull play this year, not just with Bozak, with high-quality defence offset by invisible offence.

Is Tyler Bozak Actually Broken?

Maybe, but not nearly as badly as it looks or as I thought. It might be better to say that he's been put in a position in which it's much harder for him to succeed. He has less experienced teammates on defence with him, the goaltending has been hilariously bad when he's been on the ice, and the offence has not been clicking.

Some of this offensive malaise seems like it might be Marner.  He's not the artist in the offensive zone so far this year that we have seen him be. No one is surprised by Marner anymore, and they have, likely, figured out how to effectively check him.  He needs to sort out for himself how to counterattack and get it going again.

Some of this might be van Riemsdyk.  Although the part of one game sample with Patrick Marleau on that line didn't instill confidence that the problem was solvable by swapping out the left wing.

Some of this certainly is Bozak himself. He took several games to figure out Marner last year. If Marner isn't the same guy, then Bozak might need some time to get it going with whoever Marner is now.

Bozak is a passer who sets up a lot of scoring chances, so he needs some scoring wingers. The selection on the Leafs is not exactly sparse in that area. It's possible a lineup change would help. But it's more likely a higher percentage of time on the ice with one of the top two pairs of defenders is what the doctor ordered.

Bozak's periods of greatest offensive pace have come with Rielly and Ron Hainsey.  And if you want an offensively gifted left defender and a steady guy on right who sticks to the blueline to behind the defensively fragile line, they sound ideal.

This might be one of those times where you need a coach with enough spine to just ride it out while everyone returns to their normal style of play, rather than start trying radical lineup redesigns.  The timing is bad for radical moves; you don't want those on a road trip.  But even a coach as secure in his position as Babcock wants to win, and if the Goals For percentage doesn't start matching the Expected Goals For percentage with that line, he will have to do something.


Acknowledgements:  All Expected Goals data is from Corsica Hockey, unadjusted. All other numbers are from Natural Stat Trick, and any on-ice data is Score and Venue Adjusted.  All Hockey Viz content is labelled, but some textual concepts are informed by that site’s graphics.