The Colorado Avalanche are in need of a new coach, and may
be on the verge on turning a new leaf. Patrick Roy resigned his post yesterday
as head coach and vice president of hockey operations due to disagreements with
general manager Joe Sakic. In a statement, Roy said “I have thought long
and hard over the course of the summer about how I might improve this team to
give it the depth it needs and bring it to a higher level.
To achieve this, the
vision of the coach and VP-Hockey Operations needs to be perfectly aligned with
that of the organization. He must also have a say in the decisions that impact
the team’s performance. These conditions are not currently met.” Sakic, for his
part, described being caught off guard by Roy’s resignation, saying that he
believed the two of them were on the same page as far as decision-making went.
In any event, a new era is about to begin in Colorado, so let’s take a look at what that might mean.
Roy’ tenure as head coach of the Colorado Avalanche began
seemingly auspiciously, with the head coach going 52-22-8 in his first season,
winning the Central Division, and capturing the Jack Adams award for the
league’s best coach. Informed fans knew better however, noting that the team’s
46.85 CF% and 101.75 PDO meant their performance was unsustainable, and
unsurprisingly the team bowed out the first chance they got in the playoffs,
losing their first round series 4-3 to the Minnesota Wild. The following season
the Avalanche went 39-31-12, and last year they went 39-39-4, good for 7th
and 6th in the Central Division respectively but not good enough to
make the playoffs either year.
Patrick Roy’s unorthodox style is fairly well documented: he
brings a pugnacious attitude and willingness to pull the goalie earlier than
most, but also a dismissive attitude towards hockey analytics, not surprising
given what they say about his time behind the bench in Colorado. All of this
had us wondering over at NHL Numbers what the Avalanche would look like if they
were simply an average puck possession team. To do this, our very own Ian
Fleming crunched some numbers and found out exactly what this would look like
by taking the Avalanche’s CF% for last season as a percent of the league
average CF% and then adjusting each player’s numbers by the percent difference,
which was about 9%.
What the results demonstrate should be unsurprising: the
Colorado Avalanche would be a much better team if they had just average
possession numbers, let alone above average. The team picks up 13 goals and 22
assists total by Ian’s count, for an average of .4 goals and .69 assists per
player. Most players pick up only a goal or a couple of assists, but Colorado’s
top end players are the real winner here, with Gabe Landeskog picking moving up
from 13 goals and 18 assists to 14 and 20, Matt Duchene going from 21 and 19 to
23 and 21, and Nathan Mackinnon rising from 10 and 18 to 11 and 20. This seems
to indicate that Roy’s system was not just a detriment to the team in general,
but to the team’s blue chip talent in particular.
The more interesting results are found perhaps at the team
level, where Ian took both the teams CF% and CA% as percentages of league
average in order to adjust their goals for and against for the season. While
Colorado’s CF% was about 9% lower than the league average their CA% was about
15% higher, meaning there was significant influence on their goal differential
when adjusting to league average. Following Ian’s model the team’s goals for
went from 138 to 151, their goals against from 151 down to 131, and their GF%
from 47.8% to 53.6%, a pretty substantial increase.
All of this takes the
team’s goal differential from -13 to +20, and while this model doesn’t regress
each individual’s shooting percentage or account for changes in goaltending
suffice to say that the jump in goal differential would still be significant
even if it did. While we can’t say for sure that the team would make the
playoffs if they performed as such, in the Central Division both Nashville and
Minnesota secured wildcard spots in the playoffs with goal differentials of +13
and +10 respectively, and in the Eastern Conference Detroit made the playoffs
with a goal differential of -13 while Philadelphia’s was -4. It doesn’t require
too much of an imagination then to see that if the Colorado Avalanche had the
kind of coaching that would bring them to league average in possession instead
of Patrick Roy, they would very well have a chance at the playoffs.
Will the Avalanche have a bounce back season? It’s hard to
say, as the lineup still maintains serious deficiencies, despite the true
talent that is there. Whether the team makes the playoffs or not is equally
hard to determine, and perhaps the improvement under a new coach is not as
dramatic as our model would suggest. Still, playing around with the underlying
numbers a bit seems to suggest that Patrick Roy hurt rather than helped the
team during his time there, all while winning a Jack Adams award. For fans of
the Colorado Avalanche, this could be an exciting time to turn a new leaf.