October 13 2011 09:14AM
When I last wrote about this issue, fellow LeafsNation contributor The '67 Sound rightly pointed to two issues in the study: 1) Sample size is something of a problem, and 2) Corsi numbers would probably be more accurate, as the stat represents an increase in sample size in and of itself, since it's counting shots instead of goals. I'll admit it: I was getting lazy.
The only remaining thing to do is do it again, this time using three years' worth of data instead of one, and focusing on a relationship between 5v5 playing time per 60 minutes and 4v5 Rel Corsi, instead of GA/60. Once again, I've put a minimum of 20 games played in any given season.
From this angle, the numbers seem to bear out Michael Langlois' assertion that the best penalty killers are usually the best offensive contributors as well.
I split all the players' individual seasons into four buckets, each roughly representing a first, second, third, or fourth "line". Now, I should point out that I set a minimum amount of time on the PK for this, since any player with an average of less than 15 seconds of PK time per game isn't really a PK player. So any player with less than 0.25 minutes per 60 at 4v5 is out, which precludes players of all kinds, from Ilya Kovalchuk to Colton Orr, but we're looking for penalty-killing specialists here, so I'm not worried about losing a few, actual "first" or "fourth" liners.
So, aside from the graphical correlation that you might guess is there, check out what happens in four, equal groups of players, sorted by 5v5 ice time:
There is a marked improvement in the second and first buckets or "lines", so I would argue that using more offensively talented players does in fact result in better penalty killing. Sure, it's far from conclusive evidence, but it's enough for me at this point.
Now, it's also obvious that, while there appears to be a connection between 5v5 ice time (read: offensive talent) and shots against while on the PK, it's not a very strong one, and there are plenty of low-minute players with good PK performances, or "specialists". Let's focus on our fourth and third buckets to see if we can suss out any players that show up in more than one year to see that it wasn't a fluke.
First, a closer look at those fourth and third buckets:
Now, mathematically, this is a somewhat arbitrary cut-off, but if the first bucket/line kills penalties with an average Rel Corsi of 8.21, I think it's reasonable (at least, for the purposes of this study) to call any player with a Rel Corsi over 10 a "specialist".
Below is a list of players from the fourth and third buckets who managed to do that in more than one season. First, the fourth liners:
Brian Sutherby, DAL (2x)
Jesse Winchester, OTT (2x)
Kirk Maltby, DET (2x)
Mike Brown, ANA, TOR (3x)
And the third liners:
Cal Clutterbuck, MIN (2x)
David Moss, CGY (2x)
Huh. So that's it. Similar to my last article, I'd conclude that "specialists" are a lot more rare than we generally assume. I'd also like to point out, that, in an analogous way that SH% affects goal totals, consistent, year-to-year, effective penalty killing is a function of luck as well. Too few players repeat their good performances over a three-year span for it not to be.
The only player that made it on to lists on both articles? Mike Brown.