There is no Hunlak conspiracy, there is just Hunlak

Updated: December 14, 2016 at 11:27 am by Jeff Veillette

A month ago, Leafs fans were introduced to a defensive pairing that, well, sounded like a worst-case scenario. Matt Hunwick on the left side. Roman Polak on the right. All of Toronto’s veteran presence on the point concentrated into a single deployment. What we knew heading into it wasn’t encouraging. 

Now, they’ve played for 133 minutes this season at 5 on 5. Their next game together will make them the third most used pair for the team this season, solidifying Morgan Rielly & Nikita Zaitsev, Jake Gardiner & Connor Carrick, and “Hunlak” as the three pairs. Here’s how each of those three pairs looks in terms of shot-based percentages at the moment.

Pairing Corsi Fenwick Shots Chances Expected Goals
Rielly-Zaitsev 51.1 50.2 51.1 46.6 49.7 40.7
Gardiner-Carrick 56.6 58.0 59.1 65.3 65.5 66.7
Hunwick-Polak 42.9 47.3 44.4 56.1 55.4 55.6

As you can see, the Hunwick-Polak pairing has been the worst regular pair in three of four shot-based metrics. Interestingly, there are some things you can pull out of this table that aren’t terrible.

For example, the fact that they have a sharp increase from CF% to FF% implies that they’ve done a better job at dictating shot blocks for their team (or a lack thereof from the opposition) than the other pairs have. The drop off again in shots percentage is curious, though; it would imply that shots that they don’t get a piece of are more likely to hit the net, or that offensively, they’re driving more missed shots towards opponents.

Chances, expected goals, and actual goals are where this gets interesting. While the Gardiner-Carrick pairing beats them handily in all three, they’ve outperformed Rielly and Zaitsev here. I think a lot of that comes from goaltending; goals hang in the balance of a goaltender’s ability to stop a puck. Scoring chances come off of close range shots, rebounds, and rush shots. Expected goals are weighted but those weights like the same type of shots that scoring chances do.

The trio of Frederik Andersen, Jhonas Enroth, and Antoine Bibeau have a 0.943 save percentage with Hunwick and Polak on the ice; obscene, but also a reminder that Andersen had a specific heat up point that came a week or so before the formation of that pairing. As such, Gardiner-Carrick are getting a not as great but still stellar 0.932, while Rielly-Zaitsev have 0.908 to work with throughout the year.

Perhaps some of the SV% difference simply comes from that pair being formed earlier. But whatever the case is; that difference exists, and on top of obviously playing into GF%, it could play into scoring chances and expected goals. If Andersen and company are making clean saves and freezing or moving over the puck right away, this lowers the chance of rebounds (which would weigh higher in SCF and xG as well), and gets the Leafs going on the rush faster.

Could those two be helping the goalies a little to create those chances? I mean, it’s possible; maybe they clear sightlines just enough for them to track pucks better. But, at the same time, they were 0.891 with an otherwise similar Corsi/Fenwick/Shot numbers last year. Similar happened with Gardiner-Carrick; above-55% shot based numbers, lower (0.907) save percentage, lower chances/expected/goals percentage. It’s also possible that both pairings are simply playing against less skilled offensive opponents now that Rielly and Zaitsev take the lion’s share of those minutes; for example, Hunwick’s second most played against forward this year is Danon Heinen, who I had to immediately google upon seeing on the list.

Another theory is that, as stay at home defencemen, the Hunlak pairing might be naturally inclined to minimize scoring chances but give up more general shots by collapsing sooner, allowing attackers in, while the other pairs are more prone to give up their shots in the form of drives to the net. That’s strictly an eye-test thought, though; and even then, you wonder if in the long run, that cut down of a couple chances is worth it for the average odds of a “less dangerous” shot going in.

With this said, the net conclusion is still that this is Toronto’s worst pairing; hardly a shocking conclusion. They’re leagues behind Gardiner and Carrick statistically, and while they undercut Zaitsev and Rielly in a couple of ways, that pair has been improving (visually and statistically) by the week, plays more significant minutes, and makes up for a lot of its deficiencies with their offensive contributions; something Hunlak will never claim. 

Even if they’re helping Andersen a little, the bulk of the pairings success comes off of sheer goaltending luck; if it were possible for a defenceman to be able to bring a goaltender’s performance up by 29 points consistently, no team would invest significant money in a starter. To what degree that run slows down (or picks up for other pairings, or both), is unknown, but it’s all but certain that the goals for gap will close based on current play; in believing these two will continue to pull a lopsided positive goal differential based on net-negative play, you’re basically holding into a live grenade, praying that it continues to be a threat to your opponents until it explodes in your hand.

Many have noticed this, and even prior to this pattern beginning, suggested that the Leafs would be better off looking at a couple of their other defencemen in the same spot. Martin Marincin (who is now out of the equation for a bit with an injury), and Frank Corrado (who has played just one game this season) are often brought up alternatives, as their bodies of prior work look a bit more promising and both players are just young enough to be able to grow with the team while still having a bit pro experience.

It’s lead to repetitive arguments across the Blogosphere, Twittersphere, mainstream media, water coolers, and everything in between. “The third pairing isn’t going to make the Leafs competitors!”, will come out the speakers, as one person yells that just a few minutes won’t jump them 20 standings spots while the other yells that every gradual move towards sustainable success will make a difference in the long run.

Or the idea gets floated that the pairing is clearly bad, but are clearly being played to be sold at the trade deadline, as if there isn’t more value to be gained in pumping up their 23 and 24 year olds that are sitting in the press box and can be sold as players with “potential upside”.

Or the idea that they’re sheltering Marincin and Corrado from the expansion draft, as if losing games is worth less than losing a depth defenceman.

Or the idea that the team is quietly trying to lose, as if Mike Babcock, lineup decision maker, is trying to make Mike Babcock, man who has repeatedly publicly presented this team as competitive, look dumb.

Or, most fun, the idea that the two (Hunwick in particular) have become key to the back end and that “the nerds” are seeing it all wrong, pointing to the previously discussed Goals-For percentage as the answer.

Even with all of that, here’s where we really stand:

  • Mike Babcock looks to Matt Hunwick and Roman Polak because he thinks they’ll win the game
  • Right now, Hunwick and Polak’s goal-based results have worked out
  • It’s very likely that they’re benefactors of goaltending luck of the draw
  • Until they aren’t, them being caved in in shots doesn’t matter
  • Even once they aren’t, it might still not matter
  • The usual suspects will remain in the press box for the foreseeable future
  • Hunlak will live on until they fail to meet Babcock’s presently unknown standard

It’s as simple as that. I don’t think it’s the right call for them to be deployed. The net damage they’ve caused so far has overstated, but the mistakes that they do make are massive and I’m not sure how much longer the bounces can go their way. I’m also not going to try to justify their appearance with some thought of “they must know better, however…”.

Hunlak exists because the coaching staff presently believes they are the third best pairing available. That’s it. We’ll see how long that belief remains true.