IN HOCKEY YOU DON’T ALWAYS GET WHAT YOU PAY FOR

Last spring I presented a chart showing NHL performance versus salaries. As it turned out, New Jersey ($52M salary) eked out Anaheim ($39M) in the Stanley Cup final. This year’s projections (the 2004 Stanley Cup is still half a year away) suggest that a playoff position (about 84 points by the end of the year, the yellow line on the chart) will cost $40M in team salary, about $2M less than last year. The regression line is shown in red. The projections suggest Philadelphia, St Louis, New Jersey and Colorado will be fighting it out for the Stanley Cup, with Philadelphia and New Jersey the greatest overachievers. But 5 teams with below average salaries — Nashville, Atlanta, Tampa Bay, San Jose and Calgary, should make the playoffs, and anything can happen in the playoffs. Nashville, Atlanta and Tampa (none of them really hockey towns) look particularly shrewd, paying only $250-300,000 per point, one fourth of what the Rangers, Dallas, and Washington, this season’s worst under-performers, are paying per point for teams not expected to make the playoffs despite their big salaries. Last year’s biggest bargain teams have either become much more expensive this year (Vancouver, Ottawa) or stopped performing (Minnesota).

My prediction for the Stanley Cup? A New Jersey repeat, with Colorado the bridesmaids. As an underdog lover, I’ll be cheering for Vancouver and Tampa, teams with heart, talent and self-discipline.

3 Responses to IN HOCKEY YOU DON’T ALWAYS GET WHAT YOU PAY FOR

Susan says:

December 14, 2003 at 12:39

If only that were true for baseball…
Stu Savory says:

December 22, 2003 at 11:00

You know Dave, to be really honest you should have pointed out that the correlation factor R of the factors is only about 0.36 which means that only R*R = (about) 13% of the variance of the points is explained by the salaries. You should never draw a regression line without stating the correlation, we teach that in statistics 101.My estimates might be slightly off as I had to reconstruct the data from your graph, but the principle remains the same.”There are lies, damned lies, and statistics” (Winston Churchill, AFAIK).Stu
Dave Pollard says:

December 22, 2003 at 11:16

You’re absolutely right (I took stats too). But wouldn’t it be boring if the correlation was very high (as it appears to be in baseball)? At least in hockey the small-budget teams have a chance.