As the trade deadline approaches, teams that want to acquire players are waiting until prices drop. While Kyle Dubas now has to wonder about his goaltending situation, he still says he’s looking at forwards. Word is getting out that the Leafs may consider Alex Galchenyuk their top-six forward acquisition after all, which could be negotiating tactics. Since arriving on the Leafs and moving to the second line with John Tavares and William Nylander, Galchenyuk has stuck in the role, but is he truly enough?

First We Need To Talk About Tavares

John Tavares has nearly identical first and second assists to Auston Matthews. They have 13 and 12 primary assists respectively and seven secondary. The difference is that Matthews has 2.25 times as many goals. No one expects Tavares to score like Matthews, but they do expect him to be further away from Jason Spezza on the Leafs goal-scoring ranking than he is.

In the prior five years, three years on the Islanders and two in Toronto, Tavares has 1.22 and 1.57 Goals per 60 minutes on each team in all-situations. It’s not surprising his goal rate went up in Toronto, and that’s helped immensely by 2018-2019 where everyone he was on the ice with had a great year. This year, in only 39 games played, he’s at 1 Goal per 60.  Something is up.

Tavares’s results so far are not completely unprecedented. His 2016-2017 season on the Islanders had broadly similar points and assists rates over 77 games. The culprit that season was a power play that fizzled and 100 minutes of terrible hockey played with Andrew Ladd instead of regular winger Anders Lee.

This season is very different in that his power play points are right where they should be versus his career norms, but his five-on-five goals have dried up like the proverbial ice in the spring sunshine. Looking at the effect a small period of time with Ladd had on his scoring back on the Island, it’s hard not to ask if this problem is really just the revolving door of wingers he and Nylander have played with.

Beyond finding a better quality player for the top-six, there is an unstated expectation that filling this obvious hole might “fix” John Tavares. It’s not enough to just look at Galchenyuk, the bigger issue so far has been Tavares himself.

The Revolving Door

If you go to Evolving Hockey and look at Tavares on their forward lines tool, you will get a very short list of winger arrangements from the default minimum of 30 minutes played together. If you foolishly set that minimum to zero, you get 88 entries. That’s actually fairly normal, though, and Auston Matthews has a similar number. Three one-hundredths of a minute played with Travis Boyd and Zach Hyman aren’t any explanation on why Tavares is scoring so little at five-on-five this year, but the number of players he’s had meaningful minutes with might be.

Matthews has two wingers he’s played significant time with along with Mitch Marner — Joe Thornton for 166 minutes and Zach Hyman for 276 minutes. Joe means good Corsi, Zach means good Expected Goals, and no matter which, there’s always goals scored. Tavares’s life has not been that simple.

The most Tavares has played with another winger (along with Nylander, he’s a constant, so I won’t constantly mention him) is very surprisingly Alex Galchenyuk. In only nine games played, he’s at 86 minutes on the second line. Alexander Kerfoot is next with 80, and Ilya Mikheyev comes third.

Before I show you the results with these contestants on Who Plays with Tavares, CBC’s hot new reality show on every Saturday night, it’s time to talk about a thing that seems very contradictory about hockey.

Quality of Competition Doesn’t Matter (Until it Does)

We know that the single biggest impact on a player’s results are his teammates. And yet since time immemorial as soon as discussion of player value comes up, out come the clichés about playing hard minutes and facing top lines. It’s the go-to excuse. Or it’s a complaint if a player is said to be “sheltered” which is a way to insult them for their lack of toughness and to maintain the “skill player vs. grinder” false dichotomy. Quality of Competition, which can be quantified in a host of different ways, does not skew results meaningfully in nearly every case. It’s included in all models because it is measurable in effect, and because in the rare cases where a player really does play extremely hard or easy minutes it has a small impact. But it is not the answer. It’s never the answer. Except when it is.

Quality of Competition rockets right up to being one of the most meaningful inputs into a player’s results when you are looking at a single game. It should be obvious to anyone who has watched a game that playing against McDavid all night is different to playing against whoever is centring the Sabres’ top line these days. It’s so obvious that it is understandable why people extrapolate that up to season-level stats and use it as the go-to excuse. Somewhere between one and 82 is the number of games where competition starts to wash out to league average for everyone. And that number varies by how many minutes a player plays, so it’s easier to find skewed competition for a player like Spezza right now than for Tavares.  That’s what you’ll find if you look it up. Spezza has a disproportionate amount of time spent vs fourth lines, and less vs top lines, and Tavares has a regular average deployment where he sees all lines and defence pairings at league average amounts.

Never Use WOWYs (Unless You Have to)

That average nature of Tavares’s competition holds true right up until you get the knife out and start chopping his minutes up by line. Some of his 30-86 minute segments with a given winger might well be very skewed in terms of competition, both by line and by team faced. Which is why no one uses WOWYs as an analysis tool much anymore. But given that we want to know — now, not in a month — if Galchenyuk is really better than other options, we’re stuck with slicing and dicing.

All numbers are now five-on-five:

Results of each winger with Tavares and Nylander

Alex Galchenyuk985.9571.4155.0863.9
Alex Kerfoot1980.4510060.1164.04
Ilya Mikheyev2768.8834.9540.5136.73
Jimmy Vesey1543.1548.4558.748.7
Zach Hyman2842.93054.9968.17
Wayne Simmonds1033.5310058.768.6

Starting off with just the gross percentages, I’m showing you Fenwick (unblocked shots) because Expected Goals is based on that and not Corsi.

It should not come as a surprise that a line tasked with scoring is well over 50% in the share of unblocked shots, particularly not on a team that has a team rate that is 52% and eighth in the NHL. Remember this when you look at amazing stats for prospects like Filip Hållander or when you see a high SOG % from your favourite top line AHLer.

The abysmal results with Ilya Mikheyev stand out starkly when you consider that we should be seeing something more than 50% better than they are. I’m not a fan of his play in the top six, and while I agree that getting frustrated with a low-quality player with a bad shooting percentage is called being a typical hockey fan and not doing real analysis, I think there is an obvious conclusion to be drawn about him. He’s just not suited for the game at the level of complexity playing top-six on the Leafs requires. Bearing in mind that the Tavares line with Mikheyev may have hit some Montreal or Ottawa dampening fields, that’s still an alarmingly bad result, and you have to wonder how he lasted there so long.

Alexander Kerfoot has been good in this role, and could end up playing there more as the season progresses. Alex Galchenyuk has been good as well, and survived a few periods of dull play recently without much dent in these numbers. Jimmy Vesey has the distinction of looking almost as bad as Mikheyev once you move over to Expected Goals, but the surprise is Wayne Simmonds. As an emergency replacement, he’d top my list to try again if Galchenyuk and Kerfoot are unavailable.

Who Should Play with Tavares?

Looking at the component parts of the Expected Goals and actual goals scored can give a hint as to who was lucky, who got good offence, and who got good defensive results.

Top 3 Wingers For and Against with Tavares and Nylander

Player 1GPTOIGF%xGF%GF/60GA/60xGF/60xGA/60Sh%Sv%
Alex Galchenyuk985.9571.4163.
Alex Kerfoot1980.4510064.041.5502.911.635.12100
Ilya Mikheyev2768.8834.9536.730.881.641.883.244.195.22

Considering Galchenyuk has only one goal himself, his version of this line has been getting the thing most approaching decent finishing results as seen in the actual goals for, but they are hitting just their expected amount, which makes them league average performers so far. That still makes them the best version of this line. There’s no big red flag explaining Tavares’s personal results with any contestant, other than the lack of opportunities generated with Mikheyev.

For all Mikheyev had not bad defensive results last season, he is not good with Tavares or in general this year. I would swallow an entire salt shaker while looking at the Expected Goals against for Galchenyuk, because I think Sheldon Keefe is being very, very careful with that deployment until the last two games of the road trip.

All three have been lucky in the level of goaltending they’ve experienced, extremely so for Kerfoot.

The problem, if it’s even fair to call it that, with the second line production has always been Tavares.

Individual Results — all 5-on-5 minutes

William Nylander39516.687738.477.228.73
John Tavares39560.76638.485.888.31
Alex Galchenyuk9106.381121.894.358.22
Alex Kerfoot39462.275634.269.628.19
Ilya Mikheyev39463.275235.615.956.68

First, let’s zero in on Alex Kerfoot, because he is the centre playing wing when he is on the Tavares line, and that shows up here in his full results. He has a high Expected Shooting % (the far right column, based on all unblocked shots). And he has scored over expected (the column next to it, again taken as a percentage of unblocked shots), unlike anyone else on this list. But he is a very low-volume shooter. He has the lowest rate of shots of these players by a meaningful amount. He picks his moments, and his personal shooting percentage has always been high to very high. Unfortunately, there’s no reason to expect that to stay high if he shot more, since he’s not a passer as a character flaw, it’s how he plays the game, and he’s judicious in his shooting. But like all players dependant on someone else to score — like Joe Thornton — he’s only going to look as good as the shooting of his linemates allows.

Galchenyuk, who shoots a lot as a Leaf, has a very encouragingly high Expected Shooting % compared to what it was on the Senators. The concern with him was his location of shooting, and he’s doing better there on the Tavares line. Mikheyev is the worst, as expected, because he really is the classic indiscriminate shooter. But the truly worrying number is Tavares because he’s supposed to be the star, right? Score all the goals?

His expected shooting percentage is near to his recent five-year average, so there’s no big sign of age-related decline here. He is not the best shot in all of the NHL, but he usually shoots over expected by a little. This might be easier to understand by looking at Individual Expected Goals (ixG) vs actual goals.

Nylander is one under, so that’s a bit less than his usual shooting quality, Kerfoot is one over on a fairly low overall number. Galchenyuk has played too few minutes here to expect reality, full of variance, to have caught up with expectations. Tavares is over two goals below expected. It was more than three a week ago, so he has had some improvement lately which is encouraging.

Tavares has said things about not finding good opportunities and the classic “needing to be better”, but his overall shot map is reasonably close to his normal ones, and it’s likely we are simply looking at bounces not going his way, a few less tip opportunities than he would have had playing in the past with master-passer Marner, and too many meaningless minutes with Mikheyev and Vesey. I have (while annoyed) expressed an opinion that his tip-style of net-front play is stymied by Keefe’s cycle system, but I have definitely not proved that here. And I always believe first that the answer is variance before whatever story I or you might make up that is unverifiable.


I don’t think there’s anything wrong with Tavares. I think Galchenyuk is a promising looking prospect for that line. Kerfoot works well too. With both of them as winger options, the situation in the top six looks a lot less of a problem than it was, and if Dubas has to pivot to some other need in a deadline trade, that’s likely fine. Someone better would be, well... better! Improving in a big-minute position is always going to have a larger impact than tinkering with the depth, but there’s really no downside to standing pat as long as Mikheyev stays in the bottom six.