There were plenty of interesting statements in Mike Gillis’s epic season wrapup press conference Tuesday morning, but one of the most jarring came in response to the very first question posed to him by the press. To kick things off, David Ebner of The Globe & Mail asked Gillis when the “issues” that ultimately led to the Canucks’ early exit first began to surface. Here’s how Gillis responded:
I really felt that the game in Boston – for some reason – was such an emotional and challenging game, it was almost like playing a Stanley Cup Final game in the middle of the season, and from that point on I don’t think our team ever really collectively got their emotions together. We had some injuries that disaffected us, and we just didn’t seem to play consistently as well from that point. There were points where our goaltending was so good it got us through, but, as a group I didn’t think we executed as well or played as well from that point through the remainder of the season.
Heading into the playoffs, again, we won a lot of games I thought our team was somewhat indifferent in. Met a team in the playoffs that was very well coached, they played hard and they won some games that could’ve gone either way and suddenly you’re down three nothing in a series.
As a disciple of the extremist “Church of Hockey Math” (trademark, Blake Price), I’m always skeptical of a statement that lends this much power to an “intangible” force like “collective team emotion.” It’s a pretty dubious claim when you stop to think about it: a veteran team, one of the league’s best over the past two seasons, saw their season derailed by a regular-season win in early January?
Let’s channel Seth Meyers on this one: Really? Really?
The thing is, using the numbers I trust to judge “real” team quality, Gillis’s statement seems to be corroborated by the facts. Using “Fenwick”, an advanced metric that counts up all goals, shots, and misses for and against, to produce an expanded plus/minus number that we suspect closely corresponds with the team’s own in-house scoring chance tracking, it’s pretty clear that the club wasn’t nearly as good in games 43 through 82 (after Boston) as they were in games 1 through 42 (before Boston). Here are the basics:
|Shot%||Fenwick%||Shot% Tied||Fenwick% Tied||shooting%||sv%|
For whatever reason, possibly that the team hit their competitive emotional peak in mid-January, the Canucks regressed at this point. Their goaltending stayed stellar — in fact it improved, as Gillis rightly pointed out — but otherwise, Vancouver entered Boston a dominant five-on-five team and exited only a good one.
So should we accept Gillis’s explanation that the club blew a flat tire because of some emotional letdown or diminished compete level or some such hocus pocus after the Boston game? Did Cody Hodgson exhaust all the collective team clutchiness after beating Tim Thomas with that slapper off the goal post?
Or are there other more analytically sensible explanations that might help us contextualize the Canucks’ slide from great to meh over the latter half of the season? To math!
The Hodgson showcase
Later in the same presser, Gillis admitted that the team created a showcase for Cody Hodgson this past season, deploying him exclusively in the offensive zone against inferior competition in an effort to inflate his trade value and rid themselves of an apparent human resources headache. Cam Charron pinpointed the moment that the Canucks implemented this plan earlier this week, and summarized the impact of the Hodgson showcase on team possession numbers:
So Games 39 to 63 represent 25 games where Hodgson was sheltered. Were the Canucks objectively better in that time frame? Let’s see what TimeOnIce says:
Fenwick Tied PDO Pre-Sheltering 55% 101.7% Those 25 Games 47.4% 101.9% Post-deadline 58% 99.7% Total 53.1% 101.3%
Not at all. In fact, the 25 games wherein Hodgson was sheltered represented some of the worst possession hockey the Canucks played on the year. They were 16-4-5 in that run, but were awful with the puck, and their average goal differential was worse (+42.6 per 82 during those 25 games to +54.7 during the rest), so they got a whole lot lucky.
The Hodgson showcase was a qualified success in that the team indisputably upped Hodgson’s perceived value. But Gillis didn’t net a huge a return on the asset — Kassian and Gragnani are more like a couple of strategic lottery tickets — and his club’s performance suffered in the interim.
That said, the team’s possession numbers rebounded in a big way following the Hodgson for Samme Pahlsson swap at the third line centre position. Considering that, I’m unconvinced that the Hodgson showcase was a big “reason” explaining the team’s struggles following their early January win in Boston. It’s important context, and sheltering two lines in that extreme of a fashion certainly deflated the team’s puck possession game, but since things turned around once Hodgson was gone, this issue was effectively resolved.
Ah, but here’s an issue that wasn’t. One of the major events in Vancouver’s season-derailing victory over the Boston Bruins was Brad Marchand’s predatory low-bridge on Sami Salo. The hit concussed the veteran Finnish defenceman and he wasn’t quite the same afterwards. So You’re an Expert already demonstrated that Salo’s performance fell off in a big way in terms of the chance data.
So. Did Brad Marchand break Sami Salo?
I’ve broken the season up into 8 segments and broken down Salo’s performance in terms of his adjusted fenwick per game in each of those segments. I’ve also put together a bar graph to help us visualize how Salo performed as the season went along. (A quick disclaimer: the samples below aren’t uniform. Salo only played 7 of the games between 11-20, 4 of the games between 43-52 and 9 of the games in most other quadrants):
Looking at the data, the theory that Marchand is to blame for Salo’s dropoff isn’t quite supported. While he was never less effective this season than immediately after returning from his concussion, the numbers indicate that Salo was trending downward long before the fateful game in Boston. I’d argue that his advanced age and overall injury history were responsible for his increasingly submerged possession numbers. The Marchand hit just exacerbated things.
I would point out that getting Salo out of the top-4 seemed to become a priority as the season went along. Going into the postseason, the Canucks ran an ultimately unsuccessful experiment playing Edler with Bieksa as the team’s top-pairing, and Chris Tanev with Dan Hamhuis on the team’s ostensible “shutdown pairing.” It’s pretty obvious that Alain Vigneault was desperately trying every trick he could think of to keep Salo out of the top-4 and to manufacture a situation in which Alex Edler could be effective in tough minutes without the calming influence Salo provided through the first two months of the season.
It never happened. Ultimately, the Edler-Bieksa top-pairing experiment was a failure and it quickly became apparent that, as promising as he looked, Chris Tanev wasn’t ready for top-four minutes.
Furthermore, frankly, Alex Edler isn’t ready and may never be ready to drive play against top competition the way you’d expect a true top-pairing blueliner to. He needs a steady partner to be effective, and the decline of Sami Salo, which was never more apparent than following the game in Boston, effectively left the Canucks with one reliable defensive pairing rather than two. By the playoffs, their attempts to get back to two left them with zero.
In short, when Gillis says, From that point on, I don’t think our team ever really collectively got their emotions together, he’s not wrong, if by “their emotions”, he means “Salo.”