NHL: Canucks Projections – Team Totals

PLENTY TO CELEBRATE: The Canucks are in a position where they could light the lamp a whole bunch in the shortened NHL season


Well we’ve come to the end of the road with our Canucks Projections project. Having previewed every relevant Canucks roster player from the Sedin twins, to the top-six forwards, the third liners to the grinders, and from the young roster players to the defenseman. So we’ll now dive in and look at the team’s projected scoring as a whole. And on that front, my simple projection model has a lot of good news for Canucks fans as it only projects the team to fall off slightly from last season’s goal scoring pace (in fact, it improved a modest improvement over last season until I re-calibrated it in the wake of Booth’s injury news).

We’ll start with the categories that I don’t really project all that rigorously because they’re essentially random: shootout winners (which for god knows what reason count as “goals for” in the NHL) and short-handed goals. To “forecast” these totals all I did was look back five years and take the Canucks’ average over those seasons (which, in both cases, was only tenths higher than the league average) and pro-rate that number over a 48 game season.

Over the past five years the Canucks have averaged eleven shootout games per season and, on average have won five of those contests. Modified to fit the shorter schedule, the model simply assumes that the Canucks will win three shootout games this season.

In terms of short-handed goals the Canucks have demonstrated uncanny consistency, averaging a hair over seven short-handed goals per season over the past five seasons – and never fewer than six, and never more than eight. In a forty-eight game season the model assumes that this output will stay constant, and assumes that the Canucks will manage four shorties on the (shortened) season.

It’s when we get to sorting out power-play minutes and even-strength minutes that this
gets a little bit tricky. Luckily for me the Canucks have had the same coach for seven seasons and roughly the same core for most of that time. I know, for example, that the Sedins will eat up somewhere between 31-32.5% of Vancouver’s even-strength ice-time, and that Burrows will come in slightly below that (between 29-31%). Ryan Kesler, who is used situationally because of his stellar two-way play, will usually hover just above Burrows eating roughly 30% of even-strength ice-time while his most common line-mates will fall somewhere between 27-28%. Vigneault juggles his wingers a lot but he always seems to have a designated third-liner who averages between 12:15-12:30 per game (it used to be Kyle Wellwood, more recently it has been Jannik Hansen). The fourth liners tend to play between 16-18% of the club’s even-strength ice-time depending on their situational value and the health of their teammates ahead of them on the depth chart. Replacement level fodder is assumed to play just a touch over seven and a half minutes per game.

Vigneault’s defensive deployments are actually even more predictable: top-pairing gets close to 40% of even-strength TOI, the second pairing gets roughly 35% and the third pairing eats up the remaining 25% of ice-time.

Basically I just do my best to calibrate the time-on-ice based on the proportion of ice-time similar types of players have received from Alain Vigneault in the past, and considering the team’s only major new addition this season was Jason Garrison – that wasn’t as arduous a task as it would’ve been to build a similar model for at least 25 other NHL clubs.

Obviously the big Rumsfeldian “unknown unknown” you have to account for is injuries, and there’s no good way of doing this (if I knew how to, I’d be making a lot of money working for a professional sports team – also I wouldn’t have spent a bunch of time scrambling on Tuesday to re-write all of these articles and account for Booth’s injury in my model). So I just use a durability average based on the players past record of health.

Basically I apply a player’s average number of games played per season over the past five or so, but with a few exceptions. I tend to be more forgiving when the injuries are of a “freak” or “cheap-shot” nature (like David Booth’s knee injury last season, which was clearly an error on my part considering his injury news on Tuesday), but I try to keep it as objective as possible. Obviously no model can account for freak groin strains.

The exception to that rule is a guy like Ryan Kesler (and now David Booth), who is out but will be returning from injury at some point. Kesler’s return date is less clear than Booth’s, but in both cases I just do my best to guess on a return date based on available information (like a public diagnosis) or absent that, a combination of common sense and conversations with beat writers.

For the Canucks at even-strength in 2013, my model is projecting them to score 104 goals
in a 48 game season. That’s a pretty bullish projection as it would equate to 178 goals over an 82 game season (or one more even-strength goal than the Canucks managed in 2011-12). I wouldn’t have said so before I ran the numbers but even if Kesler misses the first quarter of the season and David Booth misses the first 17 games – my analysis suggests that the Canucks could be a better offensive team than they were a year ago at five-on-five so long as the likes of Daniel Sedin and Chris Higgins can be healthier than they were in 2011-12.

The same would’ve basically been true about the power-play where I was projecting theCanucks to improve upon their goal totals from a year ago until Booth went down with an injury. The

Canucks lost three power-play goals because of Booth’s absence (the model was  projecting him to lineup with the twins in Kesler’s absence), and removing 55ish minutes of power-play time from Booth’s projection cost Vancouver’ three goals. Instead of 34 (which would’ve represented an improvement over the season previous) I’m now projecting the Canucks to score 31 goals this season (Over/Under 35.5). Over a full 82 game season that would pro-rate to 53 goals or four goals less than the team managed a year ago. That should probably cost the team a full win in the standings, and in a shortened season: that’s important.
While the upgrade from Sami Salo to Jason Garrison helps stem the bleeding, losing out on David Booth’s power-play efficiency on the second unit and missing Kesler – the team’s most efficient power-play goal scorer – to begin the season is projected to result in a slightly more pedestrian power-play than the Canucks iced last season.

Put it all together and I’ve got the Canucks down for 139 goals over the full 48 game (Over/Under 104.5) season (not including shootout winners (Over/Under 4.5)). That’s a rate of 2.9 goals per game (not including shootout winners) which would have been good for eighth in the league last season and represents a meagre step back from the 2.94 goals per game the team managed a year ago.

Of course, our model relies on problem solving and basic common sense rather than any sort of sophisticated algorithm. It’s intended as a snapshot of what the roster is capable of based on the way the team’s personnel is likely to be deployed and the recent past performances of those players. Hopefully you enjoyed going through it, and be sure to let me know what you think I got it wrong. Here’s the spreadsheet I used to put it all together.

As they’re liable to shout occasionally in the upper bowl at Rogers Arena: Go Canucks go!