I am having a lot of emotions right now, and as a Midwesterner, I’m very bad at handling them. So instead of feeling things, I’ve decided to dive into special teams and try to determine who was the very best in the NHL. Forwards are separated from Defenders, for obvious reasons, and each group has a cut off for minutes over the last two seasons. So while Rookie So & So might have a ton of skill on the PP, if he only logged 50 minutes last year, he won’t be on this list.
The Penalty Kill
The data for this set is all Forwards and Defenders who spent over 150 minutes in the last two season on the penalty kill – 142 fwds and 146 dmen total. This is a weighted ranking, giving 45% weight to Fenwick For % (unblocked shots for the PK/total unblocked shots – FF%), 45% weight to Fenwick Against per 60 (a rate of unblocked shots against the PK – FA/60), and 10% weight to Time on Ice per Game spent on the PK. All data is from war on ice and score adjusted.
Why FF% & FA/60?
Well, there are really two schools of thought on the PK – first, there’s the Stars method: when they get the puck, they try to keep the puck and break out of the Defensive Zone. This method can create dangerous turnovers, but also eats valuable seconds off the clock. Or there’s the old standard – suppress shots and “ice” the puck. This puts less pressure on the power play unit, but can prevent dangerous chances and is less likely to create a costly turnover. FF% relates best to the first style of penalty killing, FA/60 the second. By weighting them equally, we’re not penalizing the players for the style of play. And by adding in TOI per game, we’re rewarding players that the coach trusts to get the job done.
Why not just use the same method that the NHL does for their “PK%” stat? Well, to be honest, that stat is fairly useless for skaters, as it only correlates highly (R^2 = 0.43 for the nerd crowd) to PK Sv%, something skaters have little influence over.
Now to the good part. Here are your Top 20 Penalty Killers: Forwards & Defenders!
What’s interesting to me is that when you look at the list of who is “least effective” for Forwards it ends up being the guys you’d expect (ie Tanner Glass #138, Steve Ott #131, Marcel Goc #130) but for Defenders it’s almost all guys who spend a ton of time on the power play, suggesting the skills don’t often translate (OEL #131, John Carlson #136, Kris Letang #127).
The Power Play
While the Power Play is easier to rank on an eye test basis (who’s the best? Ovi!) when I started looking at it from a fancy stats perspective, a lot of issues popped up. For instance, the purpose of a power play is to get a goal, but do we consider P60 the best measure of that? Because it’s hard to get a goal on the PP without a good assist. And is that really the best measure, because shooting % on the Power Play often comes down to a combination of skill and luck – we know high Shooting %s often fall back down over time and low ones (for skilled players like Patrick Sharp) often rise. So that can’t be the only factor in play. But then to base success only on possession doesn’t reward the players on bad teams who do manage to put up points, regardless of the help they’re surrounded by. So it has to balance out.
In the end I tried to balance actual success (25% weight of Goals/60 & 25% weight of Assists/60) with individual effort (20% individual PP Scoring Chances) and repeatability (20% PP Scoring Chances For/60), with a dash of “coach’s trust” (10% Time on Ice per Game) thrown in.
Why not use the NHL Power Play%? Well, similar to PK% it only correlates to one metric – Shooting% (nerd alert! R^2 = 0.325). And as already discussed, Shooting% is not a reliable metric over time.
For Forwards, I set the cutoff at 250 minutes over the last two years (166 players total) and for Defenders I kept it at 150 minutes (103 players total). All data is still score adjusted.
The moment you’ve been waiting for: The Top 20 Power Play Forwards & Defenders!
For me, the most interesting part of this exercise was looking at the team breakdowns. It was obvious to see how Washington structured their power play (pass it to Ovi, he scores), and there has been some interesting analysis on how Philly has tried to mimic that set up to great success. Toronto’s favorite punching bag, Tyler Bozak, scored much higher than I expected (#54) in large part because he ranked higher than any of his teammates in Goals/60, and was a respectable 30th in individual chances. Again for Defenders, the least effective Power Play guys are great penalty killers – Girardi & Bieksa came in at #97 & #98.
Now here are some rankings I suggest you take with a grain of salt (as opposed to the other rankings which are 100% accurate). These are just straight averages of the players on each team over the last two years. FYI, the rank of Defenders will almost always be better because there are just less Defenders overall. Again, these ranks are just of skaters so if you have, say Carey Price and not Kari Lehtonen, it could make a big difference to the overall success of your Penalty Kill.
As for next season, a lot of these top guys have been moved around. Timonen, who was a big part of the Flyers PK is retiring, and Kevin Bieksa was traded from Vancouver to Anaheim, where he’ll assuredly be used on the PK there. Meanwhile, Ryan Garbutt, Dallas’s best PK Forward, was moved to Chicago, and while Patrick Sharp is assured to improve the Stars’ lackluster power play, we might see some slippage in our ability to kill penalties.
Swedish Elf Prince John Klingberg ranked #30 on the power play while Aaron Ekblad ranked #58, though I’m mentioning that for no specific reason at all. With the addition of Dougie “Loner” Hamilton, Calgary has 4 Defenders in the top 40 on the power play, which should be more than a little scary for their western conference rivals.
Special players Patrice Bergeron, Erik Karlsson, and Mark Giordano managed to make the Top 20 in both lists, which is beyond ridiculous. In addition to being quite good on the 5v5, UFA Cody Franson is also #29 on the power play list.
Anyway, here are the full ranks as a google doc for you to peruse and draw conclusions yourself.