Is PK Subban really any “risker” than any of his contemporaries?

By now you’re all well aware of the narratives surrounding Mike Babcock, P.K. Subban, and Team Canada’s personnel decisions. And let’s face it, there was going to be controversy surrounding them no matter which players were scratched and which were dressed. We Canadians panic a lot about small things like this. It’s just what we do. 

But at the same time, the reigning Norris Trophy winner as last year’s consensus best defenseman in the known universe was scratched yesterday in favour of three inglorious minutes from Dan Hamhuis, as Canada just barely squeaked by a Finnish team that’s probably not even as deep as the Buffalo Sabres by a 2-1 score in overtime.

It’s not exactly a stretch to say that they probably could’ve used the services of one Pernell Karl..

Offense was at a premium yesterday, and while Canada’s possession numbers show that they were pretty clearly the better team and would crush an 82-game season, single elimination tournament play leaves you a razor-thin margin of error. You can’t afford to be playing too many tight-checking, close games because the fewer goals you score, the greater chance there is for a single unlucky bounce to knock you out of the tournament. 

Logically then, you should want one of your most talented players – and arguably the second best offensive defenseman in the world – in the lineup to tip the scales in your favour just that much more. Unfortunately, this probably won’t be happening any time soon:

We all know what P.K. Subban can do on offense. He leads all NHL defensemen in scoring over the past two seasons, and is an elite shot generator on the powerplay. The fact that he’s better in the offensive zone than most everybody on Canada’s defense isn’t really up for debate. 

Instead, the issue with P.K. Subban seems to be trust, or a lack thereof from Lindy Ruff and Mike Babcock. “He’s too risky in his own zone,” the pundits say. “He has far too many giveaways!” The thing is, this sentiment is completely wrong. It’s garbage. There is not a shred of valid evidence to support this opinion. 

In fact, P.K. Subban stacks up as one of Canada’s very best defensive blueliners, both at even strength and on the penalty kill, and far more proficient defensively than not only Jay Bouwmeester and Dan Hamhuis, but Duncan Keith and to a lesser extent Shea Weber as well. First off, let’s establish a little context to keep in mind in terms of how each Canadian defenseman is deployed by his NHL team:

As opposed to the usual O-Zone Start% metric we use when looking at deployment patterns, this chart shows the true proportion of offensive zone starts and defensive zone starts each player has seen over the last 3 years. The upper left hand quadrant indicates a player is starting many of his shifts in the defensive zone and few in the offensive zone, whereas the lower right quadrant is the opposite.

As this is the case, we’d expect to see players in the lower right hand quadrant (mainly Duncan Keith and Drew Doughty) have a larger net positive effect on possession than guys buried in the defensive zone (Shea Weber, Marc-Edouard Vlasic, Dan Hamhuis) because they’re in a better position and therefore more likely to register the next shot attempt than they are to give it up. However, this isn’t necessarily what we observe:

In this chart, we’re looking for a big blue bar pointing up (meaning that the player’s team controls a larger portion of total shot attempts when this player is on the ice), a big red bar pointing up (meaning that teammates get more shot attempts when this player is on the ice), and a big green bar pointing down (meaning that opponents get fewer shot attempts when this player is on the ice). Of all the defensemen selected to team Canada, P.K. Subban not only has the biggest positive effect on his team’s offense, but the biggest positive effect on his team’s defense too. 

Interestingly, the Chicago Blackhawks have yielded a ton more shots against with Duncan Keith on the ice. I’m hesitant to label this a deployment thing because of what we looked at in terms of zone starts, or a Quality of Competition thing because Joel Quenneville has increasingly relied on Niklas Hjalmarsson and Johnny Oduya to handle tough assignments, and Quality of Competition as we usually measure it doesn’t really have a significant impact on long-run results. 

This would seem to indicate that of all the Canadian defensemen, Duncan Keith is the least defensively reliable player, which is surprising to say the least. This sentiment is further reflected when just looking at raw Fenwick against numbers:

P.K. Subban once again stacks up nicely. Despite playing for a woeful Montreal Canadiens team, opposing players have been able to manufacture extremely few shot attempts against him over the past handful of seasons; only Drew Doughty and Alex Pietrangelo have been stingier. 

Duncan Keith on the other hand gives up more shot attempts than anyone outside of Shea Weber (who is buried in his own zone by Barry Trotz) and Jay Bouwmeester (whose time in Calgary is included at the very beginning of this sample). This isn’t limited to 5-on-5 either, as Keith has struggled immensely to kill penalties as well.

Between 2011 and 2013, there were 100 defensemen in the NHL that spent 300 or more minutes on the penalty kill. Of these 100, Keith was 94th in shots against per 20 minutes of ice time, giving up nearly 2.6 more shots per 20 minutes than an average player in this group. Not surprisingly, Chicago’s penalty kill suffered, toiling in the bottom-5 in the NHL in both shots allowed and PK%. That is, until Joel Quenneville switched to the Hjalmarsson/Oduya pairing as the lead PK unit in 2012-13. The Swedish pair put together a monster season, ranking among the NHL’s top shot preventers. As a result, Chicago cut their shots against by nearly 20%, and finished 3rd in the NHL in PK%.

By contrast, P.K. Subban was 9th in the NHL in PK SA/20 over the same time, and has been one of the league’s elite penalty killers for years:

The graph means that opposing teams generate around two fewer shots per 20 minutes when P.K. Subban is on the ice killing a penalty compared to when he’s on the bench, or about 24% fewer shots against P.K. Subban than they do against Duncan Keith. Once again, P.K. Subban is not a defensive liability.

First of all, there are no bad players on Team Canada. Every single one of these players has elite strengths, and I could only really see an argument that Jay Bouwmeester doesn’t belong on the team. That being said, the evidence points to P.K. Subban being Canada’s best offensive weapon from the blueline, and probably the second best defender behind Alex Pietrangelo too. 

The argument that he’s “risky” or untrustworthy is fueled mainly by biases and misconceptions about what “good defense” is. When Subban is on the ice, he’s more likely to get a chance to score the next goal than he is to give it up, which is basically all that matters in this discussion.

At the end of the day, the drop off in defensive ability from Subban to Keith is marginal in small samples, as is the drop-off in offensive ability between Subban and Bouwmeester. We’re talking like maybe one shot per game. This adds up over a long season, but keeping Subban out of the lineup isn’t going to make-or-break Canada’s chances to win the gold medal.

With that in mind, it’s still reasonable to say that scratching Subban is poor decision making, as you’re incurring the opportunity cost of not having a super-elite blueliner in your lineup, instead replacing him with a good-to-very-good one. And in doing that, Mike Babcock and Lindy Ruff are incurring more risk of losing a tight game than P.K. Subban has ever carried.

You can get on the coaching staff for not icing the best lineup possible while still liking the team’s chances to win it all, because.. CANADA!