The Specialest Teams in the League

Now that Kari’s got his groove back, there is no more common topic of conversation for Stars fans than how our Special Teams are playing. It’s long been known that Merrin & I are fans of the Stars’ Penalty Kill, but that our Power Play leaves something to be desired.

Recently, Josh Liles at Defending Big D ran a short article on “Special Teams Efficiency”. His conclusion was that we have average special teams because of where we fall on the combination of the two. He’s right, to an extent. As of 3/21 we are 18th in PP% and 23rd in PK% (Power Play % = Power Play Goals/Power Play Chances, PK% = Penalty Kills with No Goals Against/Times on the Kill), which puts us overall 22nd in the league. It’s not good, but it’s not bottom of the barrel. (Note: Since then, our PP% has improved, but to keep the integrity of all the data presented here, that’s what we’re going with.)

But really, that doesn’t help us understand overall what’s going on with our Special Teams and how to break them down, especially on a player by player basis.

First, we have to make sure that PP%+PK% is actually going to tell us something useful, like whether or not it affects winning. Conventional wisdom says yes, but you know what they say about “assuming” things.

3.21 ROW vs PKPP% - 3 season

This shows the correlation between the PP%+PK% calculation and ROW, games won in regulation or OT. It’s a better measure than using straight wins because special teams have literally no effect on a shootout. The R² value here is 0.28. Remember, the higher this number is, the better a correlation it is (a perfect correlation equals 1). For example, if you compare last season’s 5v5 Score Adjusted Fenwick For % to ROW, even everyone’s “best predictor of wins” only had an R² of 0.398, so this 0.28 is quite good.

But again, this doesn’t help us pinpoint issues with the team, as individual players aren’t measured in PP% & PK%. They are, however measured in goals, and more importantly, Scoring Chances.

Goals For % on the Power Play is basically the same thing as PP% (a successful power play results in a goal). Goals For% on the Penalty Kill is a roundabout way of saying “we didn’t get scored on!” because if we’re putting the puck in the net, the opposing team isn’t, so it’s a rough approximation of PK%.

Unsurprisingly, Goals For % Efficiency (PP GF% + PK GF%, henceforth abbreviated as GF% Eff), correlates even better with ROW than PP%+PK%, with an R² of 0.32. Still, goals, especially shorthanded goals, are rare events, which is why we don’t use them to measure possession, and instead use shots or even better, Scoring Chances.

3.21 ROW vs SCF% Eff - 3 season

This shows 3 seasons worth of ROW vs 3 seasons worth of Scoring Chances For% Efficiency (same as GF% Eff, SCF% Eff  = PP SCF% + PK SCF%). Yes, the R² drops to 0.18 so it is a looser correlation, but is one nonetheless.

What’s particularly interesting to me is that when you make this same graph, but just for this season (14/15), the R² for SCF% Eff vs ROW climbs to 0.33, while the PP%+PK% vs ROW R² drops to 0.27. In fact, doesn’t this chart:

 

3.21 Sort by SCF% Eff

 

Make more sense than this chart?

 

3.21 Sort by PKPP%

 

When sorted by PP%+PK%, you have three teams bound for the playoffs in the bottom six. By SCF% Efficiency, you have teams with no chance at the playoffs at the bottom (Dallas included), and President’s Trophy contenders at the top.  It also probably explains some of the reason that Calgary has had such a great season thus far – they are 2nd in GF% Eff, 9th in SCF% Eff, but 17th in PP+PK%. Traditional methods of analysis aren’t always going to cut it.

All this is to say that SCF% Efficiency is a decent proxy to measure individual players contributions on special teams, especially this season.

A quick glossary for you:

Special Teams TOI: Total Minutes spent on Power Play and Penalty Kill

% TOI on PP: Power Play Minutes/Special Teams TOI

SCF% Efficiency: PP Scoring Chance For% + PK Scoring Chance For%

SCF% Efficiency Rel: A player’s SCF% Eff – Team’s SCF% Eff when that player is not on the ice

3.21 Dallas SCF% Eff Chart

3.21 Chicago SCF% Eff Chart

So here are two charts that look at all players with 50+ minutes of Special Teams ice time. Chicago is one of the best teams in the league at SCF% Eff (and not coincidentally, PP%+PK%). Dallas is one of the worst. There are a couple things that stand out to me when comparing the two like this.

First, the Blackhawks have three out of four of their main Special Teams skaters pulling up that SCF% Eff number (Toews, Hossa, Seabrook). The only one with a SCF% Eff below Chicago’s average is Keith, and he’s still above the league average of 100. Of Dallas’ six biggest TOI skaters, only Spezza, Jamie Benn, and Seguin are helping the team on a relative basis, and only Tyler is above league average.

Now, it’s easy to think that SCF% Efficiency on a player basis is linked to how much time is spent on the power play, and if we were discussing GF% Eff, you would be correct (R² = 0.33). However, looking at 108 skaters in the 14/15 season (that’s 4 teams’ worth), there is an R² value of only 0.03. Which…is nothing. There is no correlation between SCF% Efficiency and being used on the power play. (For Dallas, the R² is even lower, actually).

The takeaway here is that Dallas is having trouble because our heaviest minute players aren’t putting enough pressure on opposing teams, whether we’re on the Power Play or if we’re on the Penalty Kill. And the skaters at the bottom of the pack – Hemsky and Fiddler especially – probably need to make better decisions when they have the puck.

So go ahead and yell “SHOOT IT!” whenever you feel like. Because you know what? You’re right. They should.

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s