Filed under:

# Offensive Line Efficiency of the Canucks

A look into a new formula that shows what lines get even strength times versus their production to see what lines at best at even strength.

Hello all,

Let’s take a break from individual autopsies as the Canucks are a team and the team as a whole can sometimes tell a better story than individual pieces. I think one of the weaknesses of this team this year was its even strength play. The Nucks were ranked 22nd in the league. Let’s put that into perspective. 15 out of the top 16 teams in this category made the playoffs. (LA was 6th, but had so many other issues that led to their demise). So basically we sucked and still made it in. Thank you 2nd ranked PK and a PP that was over 19%.

A lot of bloggers and MSM have talked about the weakness of the Nucks at even strength, the use of lines and their productivity. I think I have come up with a statistic that can show the productivity of lines. I chose to focus on even strength as it was the weak area for the Canucks and most of each game is played 5 on 5 and will give the greatest sample size. My raw data comes from dobberhockey.com where it can give you line production numbers and line combination frequency based on NHL real time data. So what that means is that when a goal is scored the forwards on the ice at the time of goal are part of the point, even if they didn’t record a point. Also, the line combinations have way more variations as real time data takes into account partial line changes. So some line combo frequencies are really low as players may have been on the ice together because one got trapped during a change.

So the math works like this:

Eg: A line of Zandberg/Missy/Kent are on the ice at EV 10% of the time. This line has 9 EV points attributed to it out of a team total of 125 EV points.

Now we have a number, what the hell does it mean? I am using my own logic here and I am sure this will be a point of debate, but if a line’s frequency at even strength is 10% and we want to look at lines in terms of offensive productivity, then hypothetically they should contribute 10% of even strength scoring. That would mean the guideline for efficiency would = 1. Anything above it would be a more productive line and anything below it would not.

Now….take a breath and let it sink in.

Ok, when looking at line combinations I used a cutoff frequency of 2%. This was arbitrary on my part but the reasoning for it is simple. As mentioned above, some lines were recorded because players got trapped on ice for a noticeable time. And if a point was scored while that line was on the ice, their OEN would have been way above 1.00. Secondly, I chose 2% frequency as it gets rid of some lines that were intentional but only played a few games together (probably due to injuries) or ended up together several times because the coach shook up the lines during a game. The 2% frequency and higher shows the lines that were intentional played the most.

I feel that this stat gives us a better look at the mindset of coaches as it doesn’t look at the player determined stat of ice time. I say player determined because the coaches make the determination of the frequency of who goes out on the ice, but the players have a larger say on when they come off. (Especially when there is no whistle.)

Now let’s look at the Canucks during the regular season.

There are a lot of conclusions that can be made from this table. First off, the first line with Vrbata was more productive than with Burrows, but line with Vrbata also had twice the ice time. Secondly, and more surprisingly, Nick Bonino was on a line that produced the highest OEN. But that line was not kept together long enough, which was too bad as it had speed and some size with Matthias. You can also see that the line that had magic last year (Kass, Matthias and Richie) was not clicking at all this year. Overall we can see that the Canucks lines were balanced.

OEN is limited in what it can show. It is not a good tool to compare even strength lines between teams as even strength ice time and point totals are different for every team. Although you could draw comparisons of value between top lines of teams if their ice times were similar.

For comparison here are Anaheim’s numbers:

Anaheim’s even strength ice time was more evenly distributed, but the top 2 OEN lines had Perry and Getzlaf on them. And having Beleskey with them led to higher offensive efficiency than with Maroon. And a line of Jackman/Maroon/Thompson, which was used frequently, led to no even strength results.

I think the one area we can compare between teams is the usage of lines at even strength. The Sedins dominated ice time 5 on 5 while the Ducks had more balance. The reasoning for this variation is the interesting one and it will surely lead to discussion over how strong the 3rd and 4th line of the Canucks really were. Or maybe the Canucks had to come from behind too often, leading to the Sedins being used more frequently in shorter shifts. I will you guys drive that topic.

Having spent hours reading through data, I hope for two things. One, I hope math is right and two, I hope I have created a statistic that is relevant to understanding the game better. I also hope that any flaws are brought up and discussed to see if the stat can evolve. I also look forward to my Nobel Prize in hockey advanced stats.