The Canucks have a problem on defence. They’re currently giving up 2.89 goals per game, tied for 12th worst in the NHL. Part of the problem has been their goaltending, which has been outstanding one game, disastrous the next, and mediocre otherwise, with Ryan Miller and Eddie Lack combining for an atrocious .900 save percentage.
But there are defensive issues as well and two players in particular who have underwhelmed in that regard: Dan Hamhuis and Kevin Bieksa. For the first time as a Canuck, Hamhuis is a negative possession player, while Bieksa is making a mess of relatively soft minutes.
Both Hamhuis and Bieksa were expected to bounce back from pedestrian seasons under John Tortorella, returning to a partnership that was extremely fruitful as a shutdown pairing under Alain Vigneault. Instead, they’ve struggled and seen their partnership broken up once again. What’s the problem and how can the Canucks fix it?
One way to look at this is to use a relatively new statistic called Delta Corsi or dCorsi. Essentially, it’s a way to put the Corsi statistic — which is just attempted shot attempts for and against — in context. Whenever I write about Corsi, I try to account for context, such as if a player is starting a lot of shifts in the defensive zone or if he is facing the best players on opposing teams on a regular basis.
What dCorsi attempts to do is figure out an Expected Corsi for a player based on those contextual factors, then compare a player’s actual Corsi to that expectation. This is useful in a couple different ways: it can tell us is a player getting feather soft minutes is making the most of them or a guy facing tough opposition is getting shelled more than the average player in that situation, but it can also suggest whether a player is actually being used correctly. A player touted as a defensive defenceman who gets absolutely destroyed in tough minutes probably isn’t as good a defensive defenceman as he’s purported to be.
While I’m always a little skeptical of statistics that try to account for context to this extent, the results corroborate what other advanced statistics and the eye test have suggested about Hamhuis and Bieksa this season: they’re just not playing as well as we would expect.
Stephen Burtch, who has been working on dCorsi for several years, posted the dCorsi results for all players who have played at least 200 even-strength minutes this season. That gives us 12 Canucks to look at this season.
Quick key: the first six columns give the expected Corsi For or Against per 60 minutes of ice time, and the resulting dCorsi For or Against. The dCorsi column is the player’s Delta Corsi per 20 minutes of ice time. Essentially, if that player played 20 minutes, that’s the difference in shot attempts the team would have compared to a completely average player. The last column is the player’s cumulative impact on the team’s dCorsi.
Unsurprisingly, the Sedins are dominant possession players and Radim Vrbata is up there as well, but not necessarily at the offensive end like we would expect. They actually get fewer shot attempts than their usage would suggest, while preventing far more. This is likely a result of their style of play that gives them long stretches of possession without shots attempted on goal, preventing the opposition from getting any of their own.
Meanwhile, Chris Higgins, Nick Bonino, and Richardson all get more shot attempts than average, but allow far more than average as well, giving them a negative overall impact in possession.
On defence, Alex Edler and Chris Tanev have been truly outstanding this season, particularly by this metric. This is a big switch for Edler, who has generally underperformed by dCorsi, not living up to the expectations of his sky-high offensive-zone start percentage that he enjoyed under Alain Vigneault.
This season, however, Edler is actually starting more often in the defensive zone and has thrived, putting up the best Corsi of his career. This is perhaps a sign that he is best used starting in the defensive zone where he can use his transition game to drive possession. It’s also a sign that Tanev is very, very good and should get a contract extension post-haste.
The two players we’re really concerned about here, however, are Hamhuis and Bieksa. Their dCorsi impact is among the worst in the league.
This is especially alarming for Hamhuis, who has always been a solid puck possession defenceman, putting up great underlying statistics despite tough usage. Up until last season, Hamhuis had a major positive impact on possession, with the best average dCorsi impact of any Canucks defenceman.
Bieksa, for his part, has generally had a near-even dCorsi over the last several seasons, apart from a dip under John Tortorella, essentially playing as well as an average defenceman would do given his particular usage. He has made up for those average possession numbers with some decent offensive ability, putting up three 40+ point seasons in his career.
Hamhuis hasn’t seen his usage change much at all this season compared to previous seasons. His zone starts an quality of competition are similar to his previous seasons with the Canucks. Bieksa, on the other hand, has seen a switch to mainly offensive zone starts and is still posting a negative Corsi. He’s been better in previous years with more difficult usage.
Here’s the thing, though: Hamhuis and Bieksa might not actually be the problem.
The two defencemen have pretty much evenly split their even-strength ice time together and apart, according to Hockey Analysis. When together, they have posted a Corsi percentage of 51.7%. Apart, Hamhuis has posted a Corsi% of just 42.2%, while Bieksa’s is 44.6%. They are far better together.
That’s because when they have been apart, they’re having to prop up lesser defence partners. Bieksa, when not with Hamhuis, has played mainly with Ryan Stanton, who has had an atrocious start after a solid rookie season last year. Their Corsi% when together on the ice has been a horrid 38.2%. Amazingly, Stanton has been even worse away from Bieksa, with a Corsi% of 34.4%, albeit in just over 17 minutes of ice time, so that’s a very small sample size.
Hamhuis, when he isn’t playing with Bieksa, has played with Luca Sbisa and Yannick Weber. With Sbisa, Hamhuis’s Corsi% is an ugly 35.5%. He’s better with Weber, but still below 50% at 48.1%.
Weirdly, both Sbisa and Weber have been positive possession players when away from Hamhuis, so there’s something that really isn’t clicking when they play with Hamhuis. Fortunately, that also makes a fairly simple solution clear.
Once Sbisa is healthy, he should play with Weber on the third pairing, while Stanton takes a seat in the press box and possibly gets a stern talking-to from Tom Sestito. Sbisa and Weber, when paired together, have put up a solid Corsi% of 55.5%. That’s worlds better than their numbers apart from each other and particularly better than their numbers with Hamhuis.
I never thought I would say this at the start of the season, but the Canucks need Sbisa in the lineup. While he’s been legitimately atrocious with Hamhuis, he and Weber form a legitimate third pairing, with the added bonus that Weber can play on the power play and Sbisa on the penalty kill.
That frees up Hamhuis and Bieksa to reunite with one another and hopefully put their early possession statistics behind them.Tags: Charts, Dan Hamhuis, fancy stats, Kevin Bieksa, Statistics, Stats