Stat of the Week



Since 1969 ...

Why 1969? Because I'm a Michigan fan, and 1969 is very important to Michigan fans. Anyway, since 1969, the programs with the least variance in year to year results (as measured by winning %age) are

1. Michigan (var = 1.1%, mean = 77.6%)
2. Ohio State (var = 1.7%, mean = 76.6%)
3. Nebraska (var = 1.8%, mean = 80.0%)

(for the statistics novices, variance is a measure of how close individual points in a dataset are to the mean of the dataset - so a basketball player who scores 20 points each and every game has zero variance, while a streak shooter who scores single digits one night and 30+ the next has a very high variance)

Not surprising results; you've got three programs that have been consistent winners over that time frame, and consistency = low variance. When you go 39 years without a losing season, only one .500 season and only one perfect season, you're posting the most consistent results in football.

But what do you make of ...
4. Michigan State (var = 2.0%, mean = 51.3%)
Consistently mediocre? Perennially average?

For the record, the highest variance belongs to K-State (var = 7.0%, mean = 46.7%). The lowest variance for a sub .500 team, call it "the most consistent loser in 1-A", belongs to Vanderbilt (var = 2.5%, mean = 30.9%). Keep in mind, re: Kansas State in particular, this is not variance against a trend line. Kansas State may not be "high variance" as much as it may be changing its profile, from a perennial cellar dwellar to a respectable program.


Posted: Tuesday - June 03, 2008 at 09:48 PM