Anatomy of a midseason coaching change

In the last two days, I’ve written posts in this space combatting some of the discussion about the debate over whether Alain Vigneault is the right man to lead the Canucks. This is a topic on which I take a much stronger tone because it seems like a lot of people think that the coach really matters in the grand scheme of things. A coach can put guys on the ice and make in-game adjustments, but ultimately the style of a team will be determined by the kind of players management acquires.

The first post is a defence of trusting the process. The second is about score effects and sitting back. This one is ultimately my thoughts on the positives and negatives of a midseason coaching switch. Read on the jump.

I think Alain Vigneault is a good coach and a sharp guy, but most of his knowledge is inherent in the institution. One of the things about Mike Gillis’ tenure with the Canucks is that the front office seems to have more input into roster and lineup decisions than most regimes. Gillis has often talked about lessons learned from Moneyball, and seems to prefer a hands-on approach where every part of the organization works in lockstep.

Here’s the reason the Canucks should keep Alain Vigneault until the offseason: because firing him solves nothing. Coaches firings, especially midseason ones, I find are usually done to distract the press into convincing the masses that the organization is doing something to fix a larger problem. Mark Cuban said at Sloan that the goal of a general manager is not to manage a club, it’s to keep his job. That’s paraphrasing, but also very true.

This season has claimed the coaching life of Lindy Ruff, which buys Darcy Regier a few more hours, but has done nothing to fix his club filled with lousy contracts. Last season, Brian Burke fired Ron Wilson in an attempt to stop the bleeding of the Toronto Maple Leafs, and eventually his own tenure as general manager came to an end when the sale of MLSE was complete. Pierre Gauthier fired and traded everybody he could and was saved until the summer. Scott Howson replaced Scott Arniel with Todd Richards, but in these situations, the coaching replacement was just indicative of a carousel and an attempt at a quick fix, not a calculated decision.

Some more famous coaching changes:

This is the famous one. The Penguins fired Michel Therrien midway through the 2008-09 season a year after going to the Stanley Cup Finals. At the time of the firing, the Penguins were 27-25-5. They finished the season on a 18-3-4 run, dispatching Philadelphia, Washington, Carolina and Detroit in the playoffs enroute to a third Stanley Cup for the franchise under new head coach Dan Bylsma who has been with the organization ever since.

Thing is… just how good were the Penguins in 2008? A lot of empirical evidence suggests that the team was lucky to make it to the finals against Detroit that season. In 2008, the team’s Corsi Tied % was just 44.5%, which is unconscionable considering they were first place in the Eastern Conference. That was Marc-Andre Fleury’s career season, with a .921 overall save percentage. Him and Ty Conklin had .940 and .938 save percentages at even strength that season for the third best tandem in the league.

But… that goaltending didn’t last in the first half of the 2008 season. The Penguins goaltending fell to .918 at evens in 2009, but the team’s Corsi Tied % was still low—the team was getting outshot every night. In the first 56 games under Therrien, it was 44.8%. I don’t know exactly what Bylsma did when he stepped in, but the Penguins Corsi Tied % changed dramatically, going 53.9% in the final 25 games of the season. (I’m using Corsi Tied % because a commenter brought up this old JLikens post. It’s more predictive over the long-run than wins and goal differential).

The key difference between the 2009 Penguins and 2013 Vancouver Canucks is that those Penguins were empirically bad teams who were no longer being propped by great goaltending. Corsi Tied % is a very predictive statistic, and it does show that the Canucks are likely playing better than the results indicate. They’re getting good puck possession but have yet to turn the possession into shots and chances.

Mike Gillis made a huge mistake at last season’s trading deadline. It’s not that he traded away Cody Hodgson, but it’s that he did not trade for Jeff Carter. 27 other general managers made that mistake. Dean Lombardi didn’t.

The LA Kings under Terry Murray at the start of last season had a Fenwick Tied of 50.2%. After replacing Murray with Darryl Sutter in December, they went up to 54.9%. They still had a low shooting percentage as a team, and the team really came together when they traded Jack Johnson for Jeff Carter, increasing the team’s Fenwick Tied % to 61.2%. The Kings remain on top of the league in both Fenwick Tied % and Corsi Tied %.

It’s obvious that the Kings were a better team under Sutter than Murray, but the big change was acquiring Jeff Carter. That team became an instant contender, and the Kings have been by far the best team in the league since that move.

But coaches don’t always increase a team’s possession numbers.

With this one, I’ll thrown in possession numbers, team records and PDO numbers to let the reader jump to the obvious conclusion:

The Capitals started their 2011-2012 season 7-0-0, then hit a PDO slump between games 8 and 22 and threw the baby face out with the bathwater and brought in Dale Hunter. The change killed the Capitals from a puck-possession perspective and the end result was no greater than anything the team accomplished in the playoffs under Boudreau. They lost in the second round, Hunter left to go back to the London Knights, and now under Adam Oates the Caps are still a team trying to regain their footing.

The numbers are a little deceiving, because the Habs, who have historically had a pretty good powerplay, weren’t converting on their PP opportunities. Martin actually had the team in plus territory for most of the season, but they fell off somewhat after Scott Gomez (yes, that Scott Gomez) went down to injury after the first month:

The end result was a last place finish in the Eastern Conference and a lottery pick. Rookie GM Marc Bergevin has made some moves that have paid off to turn the ship around this season, but after a 19-24-9 season under Cunneyworth, the Canadiens have become the poster boys against the quick fix.

(Two charts generated using timeonice scripts like this and this)

The irony is that I recall being convinced that Jacques Martin and Bruce Boudreau were problems in their respective markets, employing defensive strategies that took away from team talent. While a lot of coaching switches have no effect, there are some instances where positives or negatives swing wildly when a change occurs midseason.

As it stands, the Canucks are a good enough possession club to not warrant a coaching switch, lest you end up like the Capitals. That said, I’m not a strident defender of Alain Vigneault. I just object to bad narrative and I firmly believe that a coaching switch isn’t what the team needs. The team needs either a scoring winger to help the powerplay or a two-way centreman to replace Ryan Kesler.

There’s no reason for Gillis to be attached to Alain Vigneault, just as Billy Beane wasn’t attached to manager Art Howe. Moneyball describes how Beane is frequently at odds with Howe to implement his own system, and eventually he ends up trading Howe and replacing him with bench coach Ken Macha, but during the offseason.

I think a midseason coaching switch is random and there’s no way to predict where it will go. The sample size is pretty low both ways. If you go with the coaching switch to satisfy Dave Pratt or Tony Gallagher, then you’re thinking like a fan in the stands. When you think like a fan in the stands, well, eventually you become one.

Making a move for the sake of making a move doesn’t seem like a positive.