Photo by “Dan4th Nicholas”, via Wikimedia Commons; altered by author

Michel Therrien has an interesting distinction in the research I’ve been doing about NHL coaching changes: he’s given me 4 instances where he and his replacement have coached 20+ games within the same season. He’s also replaced or been replaced in three of those instances by legit coaching talent – he replaced Alain Vigneault for the Montreal Canadiens in 2000-01, was replaced two years later by Claude Julien, and lastly was fired in favor of Dan Bylsma for Pittsburgh in 2008-09. What’s incredible about these three cases is that, in every single one of them, there was a drastic change in outcomes for the teams involved. Using 2pS%, or possession measured by two-period shots-for divided by two-period shots-for and against together, the numbers tell a story:

*Note my y-axis here (40-60%), reflective of the extremes of 2pS% in the NHL – in other words, not intended as any kind of graph trickery.

The average bump was nearly 5 percentage points, which, in a league that usually ranges 10 percentage points, is monumentally bad. That’s like going from a sure 5-seed to well out of the playoffs. The only exception to this Therrien Shift was when he replaced Ed Olczyk for Pittsburgh in 2005-06, where he oversaw a whopping 0.44 percent improvement, presumably because Edzo was just that bad (further suggested by Edzo not getting any subsequent coaching work). Therrien is a prime example of a clear signal of futility, something that we don’t always get with this exercise. That said, when you do get a clear signal, it’s worth paying attention, and in my next post (viewable here) I’ll look at other coaches who have likewise caused the statistical spidey senses to tingle.

For this post, I want to look at replacements that either proved to be good or bad moves, in dramatic 2pS% (and sometimes, PDO) fashion. Prime case in-point would be Montreal’s firing of Vigneault in 2000-01, where the loss of his top player, Saku Koivu, exposed a thin, injury-prone lineup backed by an end-of-his-rope Jeff Hackett. The result was a 952 PDO and 5-13-2 record in the first 20 games, despite a decent 52.7% control of possession. Therrien would be able to offset the drastic drop in 2pS% (down to 46.8%) by a regressing 1009 PDO; he self-described his approach as bringing “passion and emotion” to his players. Maybe the Jets ought to rethink their “Fuelled by Passion” slogan…

As we know, Vigneault went on to bigger and better things, and Michel Therrien has continued to be hired by teams who apparently are big into self-flagellation. Interestingly, Montreal is no stranger to these drastically bad decisions, having gone through three of the worst 2pS% drop-offs from a coaching change in the past 60 years. The nearly 6% drop of the Vigneault-to-Therrien switch came in 3rd-worst…4th worst was Montreal’s decision to bring in Randy Cunneyworth in 2011-12 as replacement for Jacques Martin. The move was a big deal at the time for a stupid reason (Cunneyworth was the very rare Habs coach to not speak French), but the bigger deal became the 5 1/2 percentage point drop in 2pS% (from 52.1% to 46.8%), which culminated in a well below-.500 record after the Canadiens’ 13-12-7 start with Martin.

The worst coaching change found Montreal and its faithful cannibalizing its coaching talent. Claude Ruel, by late 1970, was physically ill from the rigors of the Habs coaching experience. The replacement for legend Toe Blake, Ruel’s biggest sins were not winning the Stanley Cup every year and, for some reporters and his players, his struggles with the English language. Not a strong disciplinarian on a team of players with typical hockey-player maturity (recall Dryden’s The Game, plus what you already know about 20-something men), Ruel was the butt of jokes and scrutiny over much of his tenure (and beyond). When he resigned in December 1970 to a team advising role (one he would hold over the entire Bowman Montreal dynasty), the Canadiens’ possession went into freefall, dropping from 56.4% to 48.5% under replacement Al MacNeil. A PDO that bounced to 1039 (up from Ruel’s league-average 1003, likely due in some part to the post-Ruel acquisition of Frank Mahovlich) helped offset some of the issues, though, and somehow the Canadiens were able to scrape out a Stanley Cup despite a heap of criticism piled on MacNeil. MacNeil stepped down after the season and was replaced by Scotty Bowman…and you know the rest.

And what about the positive choices? There simply is no comparison to the Washington Capitals’ 1993-94 decision to fire Terry Murray for Jim Schoenfeld. A popular coach among his players, Murray had the distinction of replacing his brother, Bryan, as Caps coach a half-decade earlier. Schoenfeld had been an ESPN analyst – before Melrose poisoned that well – and a strong coach for Buffalo and New Jersey in the previous years. The latter stint included his infamous shouting match with Don Koharski, where he called Koharski a “fat pig” and told him to “have another doughnut.” Schoenfeld took a 49.9% possession team and made them into a 59.2% juggernaut, lifting Washington from the basement of the Eastern Conference to playoff contention (though some of the surge was stifled by sub-par goaltending from Don Beaupre and an injured Rick Tabaracci). Schoenfeld and his squad would upset Pittsburgh in the opening round only to fall to the eventual Stanley Cup champion New York Rangers.

So what changed with Schoenfeld? Penguins coach Eddie Johnston observed that the Caps players seemed more confident and organized on the ice; the addition of Joe Juneau in exchange for Al Iafrate gave the Caps a talented center in his prime, a playmaker who dictated play by setting up others. Juneau helped offset the fading talents of Dave Poulin and Dale Hunter. Schoenfeld was also tough on Peter Bondra and especially Dmitri Khristich, the latter seeing a drop in playing time that affected his offensive output. Schoenfeld also increased Sylvain Cote’s playing time, along with whomever they paired him with (a couple of players filled that role), to the point where he was as important as the Kevin Hatcher-Calle Johansson pairing. Schoenfeld’s shuffling of players received criticism, but ultimately brought fantastic results.

As for the supposed “system” he installed, it’s tough to see much of a shift if you look at the splits of forward and defense share of shooting (Schoenfeld is coach from Game #48 onward):

Despite the loss of Al Iafrate, Cote took over the defense shooting load and ultimately the players seemed to play similarly in the zone. It’s possible that a player like Juneau helped Washington control the puck in the offensive zone better (demonstrated by more 2 period shots-for per 60 – 29.2 up to 33.2), as did a more defense-conscientious Cote (Iafrate’s former partner on the blue line). This also led to fewer shots-against (29.3 to 22.9), though the Caps goaltenders recorded a dismal 86.2% save percentage. Ultimately, it was a few player personnel and usage shifts, rather than an obvious systemic change, that helped the Caps turn their season around under Schoenfeld.

Speaking of positive coaching changes, Michel Therrien’s removal in Pittsburgh was another of the most extreme positive moves made by an NHL club, wherein in Dan Bylsma oversaw a lineup that went from 46.8% possession to 54.3% (good for 4th all-time).

Probably the most interesting positive move was a predictable one: the return of Jacques Lemaire to New Jersey, in relief of John MacLean in 2010-11. The shift of possession from 48.4% to 54.4% was impressive, but watch the famous Lemaire approach on the shots-for/shots-against:

That’s shots-for dropping over 5 shots per 60, and shots-against dropping over 10 shots per 60. Woof. The old man still had it.

Whether it’s due to player confidence in their coach, better player personnel choices, or tactics, I already demonstrated in the previous post that switching coaches, on average, has a slight positive impact. Now looking at these individual accounts, though, you can see on a case-by-case basis the opportunity to completely change your season is possible. This certainly emphasizes the need for teams to recognize poor coaching and cut bait quickly, which is difficult to do…but possible. That will be the focus on my next post on this topic.

For those interested, a quick table of the top 10 positive and negative changes in 2pS% possession measures (denoted by “P%”) – plus “Off” (2-period shots-for per 60 mins) and “Def” (same, but for shots-against):

Don’t you just love how the tables show up on here? Sorry about that.