Comparing lines by situational usage and results

Eds. Note: Bumped to the front page for awesomeness.

After all the energy expended on here lately discussing usage, role, and primary responsibility, I felt like it would be nice if we got something productive out of it. This post is my effort to come out of that focus on forwards' role with something that is both useful and quantitative.

My work was made a lot easier because someone already identified a good methodology for me to use. Remember a month ago when we looked at the Heavy Lifter Index? That was a scheme for measuring which players performed well despite being put in tough situations, which sounds like a good start for what I'm after.

My goal in this post is to use the methodology they developed to answer more questions -- in addition to knowing who does well despite being put in tough spots, I want to be able to ask who gets put in the toughest spots or how HBL's performance ranks among lines that get similarly beneficial situations.

The HLI methodology in a nutshell: They use three statistics related to situation (QualComp, Corsi Rel QoC, and offensive zone start percentage) and three statistics related to results (Corsi differential, scoring differential, and penalty differential). For each statistic, they ask "how many standard deviations above or below the league average is this guy?", and they add those six values together to get a composite score. So if a player was one std dev above average in two statistics, right at league average in three, and one std dev below average in one, he would have a HLI score of (1 + 1 + 0 + 0 + 0 - 1) = +1.

What am I doing differently? Instead of just adding all six values together, I will add the three situation values together to get a situation score, and I will sort all players by that situation score. Then once this gives me a rough grouping of how players are used, I will look at how a given line's performance compares to other lines that were used similarly.

Also, I don't like how they counted results -- in their scheme, being one standard deviation above the mean in penalty differential counts the same as being one standard deviation above the mean in goal differential. But the standard deviation on penalty differential is about one penalty per 60 minutes, while the standard deviation on goal differential is about 0.8 goals per 60 minutes. Since a penalty isn't worth 0.8 goals, I don't want someone who draws an extra penalty/60 to get as much credit as someone who scores an extra 0.8 goals/60. And I wasn't crazy about using simple scoring differential, since that will favor forwards who play in front of better goaltenders. So I made some adjustments and came up with my own results/performance score.

How does my results formula work? I still used their basic idea of combining goal differential, Corsi, and penalty differential. But I made some tweaks to the way the results score was calculated.

First of all, instead of simple goal differential, my formula uses a goaltending-adjusted goal differential that assumes a .917 save percentage on the other team's shots on goal. That way a player's performance score won't be influenced by the quality of the goaltender behind him. Secondly, instead of normalizing the three results components against their standard deviation, I normalized them against their impact on goal scoring. Over a long term, it would probably be adequate to just use adjusted goal differential and penalty differential, but since this small sample size will have a lot of shooting percentage luck involved, I included the Corsi term to credit people for directing play towards the other net even if their shots haven't been going in lately.

One big caveat: although I've attempted to correct for goaltending, the results metrics are team-based, so they will obviously be influenced by the quality of the teammates. The situation metric should be fine for looking at how an individual has been used, but I'll try to mostly look at a line's performance instead of an individual's when talking about the results metric. Even then, the results will be affected by the quality of the defense behind the line, but I don't see any simple way out of that -- I'm happy to have suggestions, but my hunch is that when we try to compare players on different teams, their teammates' skill is always going to have some impact.

Enough yapping, what are the results?

[Edited 1:30 on 2/13 to fix a calculation error and resulting analysis]

First, let's look at the individuals who get deployed in the toughest situations to get a feel for how the numbers work:

 Name Situation score QualComp* Corsi Rel QoC* Zone starts* Dave Bolland 8.0 3.3 2.3 2.5 Fernando Pisani 6.0 2.1 1.6 2.2 Bryan Bickell 5.3 2.2 1.2 1.8 Patrick Marleau 5.0 2.6 1.9 0.5 Adam Burish 4.9 0.5 1.6 2.8 David Steckel 4.7 2.5 1.0 1.3 Pavel Datsyuk 4.7 2.1 1.8 0.7 Henrik Zetterberg 4.1 2.1 1.8 0.2 Manny Malhotra 4.0 0.4 0.3 3.3 Boyd Gordon 4.0 2.3 0.8 0.8

*Remember that these numbers are in standard deviations above the mean -- Bolland started in the defensive end 2.5 standard deviations more often than the average player.

I won't bother listing the players put in the easiest spots, because there's a pretty similar list already available elsewhere. (Here's a hint: the three most sheltered players are George Parros, Tim Jackman, and Kevin Westgarth.)

And since this is a Flyers blog after all, let's take a look at our boys:

 Name Situation score QualComp* Corsi Rel QoC* Zone starts* Andreas Nodl 2.1 0.3 1.0 0.9 Mike Richards 2.0 0.7 0.6 0.8 Dan Carcillo 2.0 0.1 -0.2 2.1 Jeff Carter 1.7 0.2 0.6 1.0 Darroll Powe 1.7 0.6 -0.2 1.3 Claude Giroux 0.3 -0.4 0.5 0.2 Blair Betts -0.3 -1.8 -1.4 3.0 James Van Riemsdyk -0.7 -1.1 0.5 0.0 Nikolay Zherdev -0.7 -0.7 -0.3 0.3 Daniel Briere -2.3 -1.2 -0.4 -0.7 Scott Hartnell -2.4 -1.2 -0.5 -0.7 Ville Leino -3.6 -1.5 -0.6 -1.5 Jody Shelley -4.5 -3.2 -2.9 1.6

Nothing shocking there -- we've been talking for a while about how both Richards and Carter have been taking a lot of tough spots, allowing Briere to get great scoring opportunities. What I want to do from here is to put those lines in context by comparing them against lines on other teams that face similar situations. To do that, I'll give a line a score for situation (by averaging the situation scores of the three players on that line).

Because the Flyers have spread the tough assignments around a little, they don't have a line that's in the top 30 of tough situations. The Richards line is #35, with an average situation score of 1.2. So let's compare them to the other lines that have situation scores between 1.0 and 1.5:

 Line Team Situation score Results score Dupuis-Crosby-Kunitz Pit 1.48 0.55 Iginla-Tanguay-Jokinen Cgy 1.42 0.32 Cole-Staal-Samsonov Car 1.33 -0.41 Chimera-Fehr-Gordon Was 1.25 0.02 Horton-Krejci-Lucic Bos 1.22 0.59 Plekanec-Kostitsyn-Cammalleri Mon 1.19 0.20 Nodl-Richards-van Riemsdyk Phi 1.16 0.02 Schremp-Comeau-Sim NYI 1.08 -0.72 Franzen-Bertuzzi-Filppula Det 1.05 0.30

Richards' line hasn't been dominant, but they're holding their own. They're in the lower tier of this list lines, but a neutral results score in tough situations isn't something to be ashamed of. If JVR can develop into more of a scorer, they'll be right up there among the best of the defense-oriented lines.

Carter's line has an average situation score of 0.4, which is 52nd in the league. Let's compare them to the other lines between 0.3 and 0.6 in situation score:

 Line Team Situation score Results score Laich-Semin-Perreault Was 0.59 0.66 Sutter-Dwyer-Jokinen Car 0.57 -0.25 Morrow-Ribeiro-Benn Dal 0.57 0.00 Gionta-Gomez-Pacioretty Mon 0.54 -0.09 Gallardi-Stastny-Stewart Col 0.50 0.05 Samuelsson-Kesler-Raymond Van 0.47 0.51 Halpern-Moen-Pyatt Mon 0.46 -0.52 Carter-Giroux-Zherdev Phi 0.44 0.62 Kostitsyn-Goc-Tootoo Nas 0.40 0.17

Geoff has been singing the praises of Carter this year, and rightfully so. His individual scores for situation and results are the best on his line, and he is carrying that line to a superior performance for their role. They have the sixth-best results score of any line in the league, despite facing above-average difficulty situations. The HBL line has gotten a lot of the attention, but this line has been at least as responsible for the Flyers' success.

The HBL line gets the ninth-easiest situations in the league, with an average situation score of -2.7; here are the ten lines with situation scores between -2.25 and -3:

 Line Team Situation score Results score Nash-Brassard-Voracek CBJ -2.25 0.35 Adams-Rupp-Asham Pit -2.33 -0.03 Bernier-Matthias-Hordichuk Fla -2.44 0.19 Gaborik-Frolov-Christensen NYR -2.47 -0.40 Stortini-Fraser-Jones Edm -2.55 -0.73 Tlusty-Bodie-Carter Car -2.55 -0.49 Burrows-Sedin-Sedin Van -2.56 0.88 Briere-Hartnell-Leino Phi -2.74 0.64 Eaves-Helm-Mille r Det -2.75 0.17 Hossa-Kopecky-Sharp Chi -2.91 0.41

This is tougher to compare, because a few of these are offensive lines being given their best chance to score and a few are stiffs being hidden in safe situations. But I'll note three things: 1) their results score is fourth-best in the league for any line, 2) that's not because they're the only scoring line getting that kind of offensive opportunity, and 3) let's all share a good chuckle at the expense of Marian Gaborik and the Rangers.

There's lots more information in there, but this post is already about six times longer than anything the average reader will get through, so I'm going to stop it here, give you all a link to the data, and invite you to find your own interesting observations.

This item was written by a member of this community and is not necessarily endorsed by <em>Broad Street Hockey</em>.

Trending Discussions

forgot?

As part of the new SB Nation launch, prior users will need to choose a permanent username, along with a new password.

I already have a Vox Media account!

Verify Vox Media account

As part of the new SB Nation launch, prior MT authors will need to choose a new username and password.

We'll email you a reset link.

Try another email?

Almost done,

By becoming a registered user, you are also agreeing to our Terms and confirming that you have read our Privacy Policy.

You must be a member of Broad Street Hockey to participate.

We have our own Community Guidelines at Broad Street Hockey. You should read them.