Wednesday, December 05, 2007


It could be, perhaps, that his shoes were too tight

I try not to get bent out of shape when a newspaper columnist abuses statistics to buttress his personal opinion -- if I let it bother me much, I couldn't read a newspaper at all -- but I expect better from Bill Houston at the Globe & Mail. I've generally found him to be a guy who looks at the data without too many preconceptions, and then draws conclusions from there, rather than the other way around. Which I why I found his piece on NHL scoring yesterday to be so disappointing: Contrary to what Hockey Night preaches, scoring is up in NHL. Problems include:
  1. Arguing with a straw man. "If you watch Hockey Night in Canada, you may assume there has been no increase in NHL goal scoring, despite the crackdown on restraining fouls." I don't believe Ron Maclean or anyone else has claimed that scoring has not increased, and Houston doesn't offer a quote.
  2. Silly black/white characterizations of past and present. "..the 'old' NHL, when obstruction fouls weren't being called" and " 2005-06 with zero tolerance" are both exaggerations, despite suiting the point he wishes to make.
  3. Using figures that are incompatible and/or incorrect. He runs down the past few years' figures for how many PPs the refs are calling, adding the note "(defined as a one-man advantage)", but then when he runs down PP scoring, he uses figures that include 5-on-3s (certainly the 631 he cites from before last night's games include 5-on-3s and 6-on-4s).
  4. An awful lot of imprecision. He uses "at this point in the season" three times without ever telling us if he means "on this date" or "after the same number of games" -- particularly when you're inserting 2005/06 into the discussion, this makes a big difference: the sked was compressed early and late then because of the Olympic break.
And fifth, a whole bunch of statements of opinion, which would be just fine if the point of his column wasn't "proving" HNIC wrong. Like:
If you read too closely, Houston seems to be saying that eliminating obstruction increases PP scoring, while reducing the goalie equipment is good for EV scoring.
But the important point is this: 118 more even-strength goals have been scored this season than in 2003-04. That's progress.

Suddenly I'm not even sure what Houston's beef is with HNIC, as apparently by the end he has given up on demonstrating cause, and is satisfied with simply pointing out that there is more EV scoring now than then. Which again, is not in actual dispute, and could be caused in whole or in part by many other factors, one of which (goalie equipment) Houston specifically mentions as relevant. Nor does he show that Maclean/HNIC agree that increased EV scoring one way or another is progress. Or, that they should. All in a day's work, I guess.

My own take on this is that I'm very glad that HNIC has deigned to critically examine the aims, claims, and results of rule/enforcement changes, because no one else is! (intarweb company excepted). The possibility that it's motivated by Ron Maclean's neanderthal biases or whatever is of no consequence to me.

Also, here's a link to the scoring changes in 2005/06 relative to 03/04, and a discussion of some of the hows and whys.


Whoa. Synergy.

Post a Comment

<< Home

This page is powered by Blogger. Isn't yours?