Alan Ryder, who has done some great work in terms of statistical modeling for hockey, makes a case for the Lady Byng trophy to get a little more respect than it currently earns in most circles:
In hockey there is a clear positive correlation between penalty taking and losing. A penalty has two proximate effects on scoring rates when compared to even handed play (the even handed scoring rate for this season is about 2.52 goals per 60 even-handed minutes).
The first effect is that the rate of goals against increases. I calculate the short-handed goals-against average across the NHL this season at about 5.79 goals (per 60 minutes of short-handed play). This means that, while short-handed, teams give up an average of 0.097 goals per minute versus 0.042 while even-handed (an increase of 0.055 goals per minute).
The second effect is that goal scoring goes down while short-handed. League-average goal scoring while short-handed this season is about 0.76 goals per 60 minutes, or about 0.013 per minute. This is a reduction of about 0.029 goals per minute. This effect is often overlooked but is about half the size of the first effect.
For his part, Juice has only garnered 13 minor penalties all season (notice Iginla's only had 15 himself. That's incredibly impressive considering he typically draws a lot of ice-time against top players).
The next step would be to figure out how many penalties Huselius draws against how he many he takes. I suspect there's some method of doing this (PIM taken by the Flames vs. PIM taken by opposition when Huselius is one the ice), but I don't think I'm up to gathering and managing all that data. Considering his low PIM total and the degree to which I've seen the opposition sent off for hooking and holding Kristian recently, I would hazard an educated guess that Huselius (and, no doubt, Iginla) generates secondary benefits like a net increase of PP time each game.
PS - For those of you mathematically inclined, check out Ryder's Hockey Analytics website. Interesting Bill Jamesian stuff there assuming you're able to decipher stats-speak to any reasonable degree.
hat tip to the Forechecker for the link.
