Twofer

Ross McKeon has a weird name, and is also the hed honcho of hockey over at Yahoo! Over the past little while, he's had two interesting things to say:


"Before the current collective bargaining agreement with its salary cap was put in place, an anonymous league GM predicted the new NHL would be a paradise of parity. Boy, was he right. Take the results from the last week for example. Of the top 15 teams in the rankings last week, 11 produced losing records, including each of the first nine (7-15-2). The top 15 teams went 14-26-5. Conversely, the bottom 15 went 34-17 and only three sported losing records. "

No, this does not signify 'parity.' What it signifies is that within small sample sizes there will be statistical anomalies. This is why we play a full season: to see who is the best in that largest sample size possible of 82 games... and sometimes, there are still flukes. Also, considering the league can only go .500 against itself, if half the teams lose, then the other half have to win. It isn't weird that top teams losing corresponds with bottom teams winning. They have to lose to someone.

Never mind that this doesn't allow for the idea that his rankings might be wrong.

McKeon also commented in his weekly Power Rankings that he was 'reluctant to jump on the Islanders bandwagon' or something of the like, but that he was doing it now. I e-mailed him about the wrongness of this, and I can't find the comment now. I wonder.

Anyway, the reason that is so wrong is because the Islanders have a negative goal differential. They are being outscored, but still winning. This means that they will almost definitely come back down to earth very soon. At the time McKeon published the Rankings, the Islanders were playing .660, while according to their goals scored/allowed, they should only be playing to .450. There will be a huge, monstrous falloff.