Penalty-Killing Plugs


Michael Langlois made an interesting comment on the Leafs season preview podcast that was published on the Globe’s website yesterday. He pointed out that, to his memory, most of the best penalty-killing forwards (going back for decades) were the best all-around players, and not those we have come to think of as a "PK specialists".

I think that Langlois’ basic premise holds water, but it also seems to me that anytime a team acquires a player with ham hands (and here, I’m glaring in the direction of David Steckel), the player is praised for their sense of defensive awareness and/or their work on the penalty kill. While it seems reasonable that this is the case some of the time, it strikes me as a platitude.

Certainly, it makes sense to expect that a team’s "bottom six" forwards can keep the puck out of their own net. At least if they can’t score, they should prevent their own team from being scored on while the real players rest up, right? This just seems like a cost-effective (or, perhaps more accurate, a cap-effective) way of filling out a roster.

(Side note: In a way, this strategy of finding pluggers seems backwards. Wouldn’t it make sense to play your best forwards in the highest risk situations, and let your "grinders" skate in the lower-risk situations? Perhaps this is a debate for another post.)

The problem is that it just doesn’t work that way in real life. A defender simply isn’t a "defensive" defenseman just because he can’t score. Just look at Mike Komisarek, for crying out loud. He’s not an "anything" defenceman, let alone another title like "rushing", "puck-moving", or "defensive". Similarly, when dealing with forwards, not every replacement-level player is worthy of defensive responsibility.

To me, this brings up two questions:

1)    How common is it for coaches to use bottom-six forwards to kill penalties instead of top-six players?

2)    Are these so-called "bottom-six" forwards any worse or better on the PK, and how many justify their "specialist" labels?

Because this is exactly the kind of situation where Danny Gray would caution us about observer biases, it’s critical that we measure as much as we can. Of course, defensive prowess is a nebulous concept, and we can only approximate the defensive value of a given player, but that doesn’t mean it isn’t worth a try.

Now, the designations of "top-six" or "bottom-six" forwards are semantic debates unto themselves, so I’m going to avoid getting hung up on these terms by looking at players based on how many minutes they receive at 5-on-5 as a more gradual (read: more accurate) indicator of a player’s worth to a club.

At this point, it’s almost a given, but all the following data is from

In the first graph, we’re looking at the 5-on-5 ice time for every forward with a minimum of 20 games played in 2010-11, and checking to see if there is some kind of discernable relationship with 4-on-5 penalty-killing playing time:

Here, we can see a few things happening. First, it’s fairly easy to recognize that players that see more 5v5 ice time tend to see more 4v5 ice time. It’s not a huge difference, but it’s clearly there. I’ve highlighted in yellow an area that encompasses low 5v5 minutes, but relatively high 4v5 minutes. There aren’t that many, but these should, theoretically, be the "PK specialists" that we always hear about. Here are some of the names inside that yellow rectangle that have high PK minutes (here, listed as 4v5 minutes/60), given their EV minutes:

Blair Betts: 3.45
Todd Marchant: 3.45
Darroll Powe: 3.18
Craig Adams: 2.87
Patrick Eaves: 2.69
Colin Fraser: 2.53
Andrew Miller: 2.22
Tom Pyatt: 2.17
Jarkko Ruutu: 1.84
Tim Brent: 1.83
Trevor Lewis: 1.76
Scott Nichol: 1.75
Tom Kostopoulos: 1.71
Chris Drury: 1.68
Dana Tyrell: 1.68
Fredrick Sjostrom: 1.65
Derek McKenzie: 1.64
Philippe Dupuis: 1.59
Mike Brown: 1.58
David Steckel: 1.68

It’s interesting that four of the five bottom rungs in this list are now Toronto Maple Leafs, but that Brian Burke elected to pass on the recently waiver-available Blair Betts, who tops the list for this range of 5v5 ice time. The Montreal Canadiens appear to have added a very useful PK veteran to their lineup for nothing.

Of course, just because a player sees more ice time on the penalty kill, doesn’t make them efficient. So let’s take a look at the relationship between 5v5 playing time and 4v5 Goals Against per 60 minutes, for players that have played a minimum of 20 games, and play a minimum average of 0.25 per 60 minutes on the PK per game. I’ll admit that this is a somewhat arbitrary cutoff, but if a player plays less than 15 seconds per game on the PK on average, they’re probably not specialists.

Here, we see that any player that can keep their opponents to less than 10 GA/60 is doing well enough on the 4v5 PK. That seems to be the benchmark that many of the more talented players (that is, the ones getting more 5v5 mintues) have set. It seems that we have a slightly larger crowd than in the first graph that is just as defensively responsible as those forwards that are entrusted with much more ice time at 5-on-5. Players that appear on both lists have been bolded. Just look at Blair Betts go:

Cody McLeod: 1.70
Patrick Kaleta: 1.85
Kyle Wilson: 2.44
Jon Sim: 3.07
Daniel Paille: 3.32
Matt Martin: 3.42
David Desharnais: 4.59
Andrew Miller: 4.84
Blair Betts: 4.87
Andrew Murray: 4.88
Tom Pyatt: 5.00
Ryan Johnson: 5.04
Darroll Powe: 5.13
Trevor Lewis: 5.22
Alexandre Bolduc: 5.37
Scott Nichol: 5.50
Craig Adams: 5.50
Jamal Mayers: 5.59
Kris Draper: 5.99
Patrick Eaves: 6.01
Colin Fraser: 6.37

Ryan Carter: 6.37
Darren Helm: 6.49
Greg Mauldin: 6.60
Tim Sestito: 6.66
Jesse Winchester: 6.74
Matt Hendricks: 6.76
John McCarthy: 6.92
Adam Mair: 7.04
Philippe Dupuis: 7.15
Todd Marchant: 7.26
Troy Bodie: 7.34
Chris Drury: 7.43
Jim Slater: 7.44
Derek Mackenzie: 7.54
Jarkko Ruutu: 7.67
Toby Peterson: 7.63
Tanner Glass: 8.66
Tim Brent: 8.70


OK, so we’re not blowing anyone away with statistical revelations, but I think most people would agree that we can see a couple general trends emerging.

In response to the first question I posed earlier in this article, it would appear that coaches rely more frequently on high-minute players to do most of the heavy lifting, even on the penalty kill. That said, players that can play defense, but somehow are incapable of offense, do at least appear to exist.

It’s clear from these graphs that there are players that performed very well – sometimes exceptionally well – on the penalty kill, despite not being main cogs in their respective teams’ offensive strategy. The snag here is that these types of players are few and far between, and "PK specialist" is a term reserved for a small handful of players. What I would surmise from this is that most players playing in a "bottom six" capacity are mostly just… well, not as good at hockey.

  • The '67 Sound

    How hard would it be for you to run the second graph as 5v5 TOI vs. 4v5 corsi? Sample sizes here too small for GA/60 to be meaningful. The chart suggests there’s no real relationship between overall talent and PK talent. Maybe that’s true (at least for Fs) and it’s more a matter of the D/G and luck.