When Harrison and Daniel invited me to join PITB for a weekly stats and analysis post, I was excited at the chance to showcase my intellectual side. So, when they subsequently insisted that the title of the series be a silly pun based on my last name (other suggestions included Drances with Wolves and Drance in your Pants), I died a little inside. But, once I had worked through the disappointment and sadness, I pledged to myself that this would still be a smart series. I couldn’t be happier to be filling this space with words and numbers. I’d like to thank the good Messrs. Wagner and Mooney and welcome all of you Bulies to “Drance Numbers.” If you like what you read, I urge you to check out my regular haunt over at CanucksArmy.com.

I’m a firm believer in the predictive power of numbers in hockey, and of their utility. Numbers and sports go together like Birdman and Weezy (we do keep score after all) and advanced stats in hockey allow us to back up our impressions with something quantitative and objective. Your eyes (and mine) have a tendency to be dishonest, and often our opinions, though held in good-faith, are clouded by bias. Numbers, on the other hand, don’t have a favorite player, they haven’t pre-formed an opinion about a particular goaltender being a “choke-artist,” and they have no capacity for cognitive dissonance.

There’s an old joke about the statistician who drowned in a lake with the average depth of three centimeters. The kernel of “truth” in this joke is supposed to be that “numbers don’t tell the whole story,” a common refrain to which many life-long hockey fans’ retreat when presented with advanced metrics in hockey. Now, I’m not sure that numbers have ever claimed to be thorough, detailed yarn-spinners (except for the time that 7 ate 9), but I do understand this reaction. Frankly, we’re very much entitled to enjoy the sport of hockey on a diversionary level, and form any opinion we’d like about any player on the ice without looking into it more deeply. To each his own, in hockey fandom especially.

Just know that metrics like Corsi, Fenwick and Scoring Chances aren’t new — Roger Neilsen was using variations of them a generation ago. At least six NHL teams are currently using advanced metrics to supplement their decision-making processes, and the actual total is probably much higher than that. It’s virtually assured that the Vancouver Canucks use advanced stats to make decisions about which players to dress, and who to target in trades or free-agency. If you’re like me, and feel compelled to know as much as possible about what’s going on with your team (on the ice and off of it), you’ll most certainly find a basic understanding of advanced metrics helpful.

What’s most important to keep in mind is that fancy-stats aren’t divorced from what you’re observing on the ice on a game-to-game basis. You know how Chris Higgins holds the puck under his body when he gets into traffic, how he kicks the puck onto his own stick and, using this method, often manages to retain possession? We can measure that. You know how hockey often seems like a game that is overwhelmingly decided by “puck-luck”? We can measure that too.

As the season goes along, and as we watch every game with the simmering intensity that defines Canucks fandom – I want you, you faithful PITB readers, to let me know what you’re seeing. Let me know if you notice anything (i.e. if you notice that “this random second line winger is AWFUL at finishing”, or that “Aaron Rome sucks, why does he get so much ice-time!?”), speak up, and I’ll do my best to quantify your impressions.

****

With that, we’ll move into the meat of today’s post. I wanted to know what we can expect from the Canucks this season in terms of their offensive output. So I crunched the numbers, and I’ve come up with a thorough goal-total forecast for every Canucks skater. My forecast is based on each player’s scoring rate over the past three seasons, and their expected ice-time this season. I’ve promised Harrison to limit the use of graphs and tables in my posts, but if you’re interested in seeing the spread-sheet you can do so here. Yes, I’m a nerd.

Basically I went to NHL.com and calculated total even-strength goals and total time on ice for every Canucks skater over the past three seasons. If you want to see an individual’s scoring rate for a single season, you can do so quite easily at Behind the Net. I did the same with powerplay ice-time, and I then calculated each skater’s scoring rate (goals per sixty minutes) and made an educated guess at the ice-time they’re likely to receive this season.

The first thing I had to do when projecting ice-time was to project the number of games each skater will play. The way I did this was to average out the number of games each skater has played over the past five years and use that number as a baseline. As a result, players like Sturm and Salo — who have not been particularly “durable” over the last few seasons — are projected to play fewer games than Henrik Sedin, who hasn’t missed a game in well over five seasons.

Once I tabulated each “expected games-played” number, I reconfigured it a bit according to logic and common sense. For example, I assumed that Daniel Sedin is likely to play all 82 games, even though his average was slightly lower because of a freak injury that cost him 21 games in 2009-10. I also assumed that Keith Ballard will be scratched for a few games at some point this season and took a few “expected games played” off of his total. Finally, I accounted for current injuries: I projected Kesler to miss the entire month of October, and Raymond to return in mid-January. Obviously, it’s impossible to accurately forecast which players will miss time over the season, but I’ve accounted for every man game.

Once I had my figure for “expected games played,” I used average TOI to project players into their expected roles. One assumption I made was that the presence of a really good fourth line centre in Maxim Lapierre will lead to an increased even-strength usage rate for the Canucks fourth line. Last season, the fourth line averaged just a shade under eight minutes per game at even-strength; this year I expect that to go up somewhat. So I gave the fourth line an extra thirty seconds per game, and to compensate for this adjustment, I took away twenty seconds per game from the 2nd line and ten seconds per game from the Sedin line. I plugged Chris Higgins into Raffi Torres’ role last season, and I swapped Marco Sturm in for Mason Raymond. For Cody Hodgson, I projected him in Kesler’s role for eleven games and in Tambellini’s role (an offensive contributor who plays up and down the lineup) for the balance of the games I expect him to play. I used Tambellini’s TOI for Mason Raymond upon his “expected” return as well.

Because the fourth line is likely to rotate personnel throughout the season, I combined the fourth line wingers and they are accounted for on the projection as “X Fourth Line Wingers.” I gave them a combined goals/60 rate of .45 (originally it was lower, but Dale Wiese’s alone should be about .41 based on his AHL translation numbers). As for Chris Tanev, and Cody Hodgson – they lacked a long enough track record to project using the same method I used for the more veteran skaters. Instead, I estimated their AHL ice-time (that information isn’t available) based on their roles with Manitoba last year, and used Gabe Desjardins’ NHLE numbers to estimate their goals/60 rate at the NHL level. Ultimately, the forecast for those two is probably a bit conservative – and I personally expect more from them than the numbers show.

Finally, NHL teams average 5 shoot-out wins per season, and those counts towards a team’s Goals For totals, so I included them. In the past three seasons, the Canucks have scored 22 short-handed goals so I included that into the total as well, but I didn’t assign those goals to any individual skaters.

So how many goals are the Canucks going to score next season? According to my detailed forecast: the Canucks should manage in the ball park of 247 Goals For in 2011/12. That total would have put them in the top five in both of the past two seasons. Though it represents an 11-goal regression from last season’s league leading total, it should be enough to keep the team in contention for the top spot in the Western Conference. Here’s a chart that includes the projected expected output for each individual skater.

Name Projected EV Goal Totals Projected PP goal totals Total Goals
Daniel Sedin 24 12 36
Alex Burrows 25 3 28
Ryan Kesler 14 13 27
Mikael Samuelsson 16 8 24
Henrik Sedin 17 5 22
Marco Sturm 11 2 13
Manny Malhotra 10 1 11
Jannik Hansen 10 10
Alex Edler 4 5 9
Chris Higgins 9 9
Mason Raymond 4 2 6
Maxim Lapierre 6 6
Kevin Bieksa 4 1 5
Sami Salo 2 3 5
Keith Ballard 4 4
Dan Hamhuis 3 1 4
Cody Hodgson 3 1 4
Andrew Alberts 1 1
Chris Tanev 1 1
Aaron Rome 0 0
X Fourth-Line Wingers 10 10
Totals: 178 57 235

The leading goal scorer among Canucks forwards? Daniel Sedin, with 36 goals. Among defenseman? Alex Edler with 9. My forecast has Ryan Kesler leading the team in power-play goals, but finishing with 27 goals total, 14 less than he scored last season. By my forecast, the Canucks can expect 29 goals from their blue-line, and should finish the season with five twenty goal scorers.

 

Thomas Drance lives in Toronto and works in social media and communications. He is the managing editor of Canucks Army, and an opinionated blowhard to boot. You can follow him on twitter @artemchubarov.

Tags: , , , , ,

11 comments

  1. Paisley
    October 6, 2011

    Excellent article. I hope to read more by Thomas Drance in future.

    VA:F [1.9.16_1159]
    Rating: 0 (from 0 votes)
  2. Mike
    October 6, 2011

    Loved the analysis. Great article.

    VA:F [1.9.16_1159]
    Rating: 0 (from 0 votes)
  3. James W
    October 6, 2011

    Drance Drance Revolution?

    VA:F [1.9.16_1159]
    Rating: 0 (from 0 votes)
    • Daniel Wagner
      October 6, 2011

      Personally, I was lobbying for “Private Drancer.”

      VN:F [1.9.16_1159]
      Rating: 0 (from 0 votes)
      • beninvictoria
        October 6, 2011

        Drance’ylvania isn’t an option?

        VA:F [1.9.16_1159]
        Rating: 0 (from 0 votes)
  4. Flint
    October 6, 2011

    Way too many assumptions are made in this article for me to think it is a prediction, rather than purely a reflection of past performance.

    VA:F [1.9.16_1159]
    Rating: 0 (from 0 votes)
  5. Eric Blacha
    October 6, 2011

    I can see some being higher, but you used math; and much like naked Kesler, it can’t be denied. Good job.

    VA:F [1.9.16_1159]
    Rating: 0 (from 0 votes)
  6. Noodle
    October 6, 2011

    Good stuff. I think a few of the goal totals are a bit conservative, but your methodology looks sound. Nicely written article too. Welcome!

    VA:F [1.9.16_1159]
    Rating: 0 (from 0 votes)
  7. aidanbc
    October 7, 2011

    Nice start, tiny drancer. Looks like a lot of work went into this. Did you prepare the same predictions last season? How did they compare to reality? I worry that Kesler’s numbers might be deflated because the three year average doesn’t account for the improvements he’s made offensively and his evolving role with the team. Likewise, I feel like Samuelsson might be trending downward, and that his totals might be overestimated. Can you explain why three seasons is a valid frame of reference for predicting performance, other than it provides a large sample size?

    VA:F [1.9.16_1159]
    Rating: 0 (from 0 votes)
  8. annie
    October 8, 2011

    Drance Music was clearly the best option: at least the singer/songwriter will argue with you about Jeff Skinner on Twitter.

    I am really glad to see this column – I am a huge fan of advanced stats, even if I occasionally have “nope, don’t care, still like that kid” moments. I am also a huge fan of not doing work. Best of both worlds, then, and I envision many happy days referring back to this column in the heat of extremely vicious and irrelevant (why is there an inverse correlation between how much people on forums care about something and how much it actually matters: there is a stats question for you) internet debates.

    VN:F [1.9.16_1159]
    Rating: 0 (from 0 votes)
  9. Anonymous
    October 8, 2011

    Love this, it gives me a good idea for my pool ;)

    VA:F [1.9.16_1159]
    Rating: 0 (from 0 votes)