EAST LANSING, MI - OCTOBER 22: B.J. Cunningham #3 and Keith Nichol #7 of the Michigan State Spartans celebrate Nichol catching a game-winning 44 yard touchdown pass from Kirk Cousins #8 as time ran out to defeat the Wisconsin Badgers 37-31 at Spartan Stadium on October 22, 2011 in East Lansing, Michigan. (Photo by Mark Cunningham/Getty Images)
In today's Varsity Numbers, we revisited some of the measures we discussed in Fun Stat Nerd Tidbits throughout the summer: standard deviation, momentum, and covariance.
This week, we'll take a look at what we can do with this type of measure. Below are some of the tools we can use once we have a nice, weekly, opponent-adjusted evaluation at our fingertips. But first, a primer:
- Standard Deviation: This is simply the standard deviation of each team's week-to-week Adj. Scoring Margin. The higher the standard deviation (and rank), the more a team's performance varies from week to week. This allows us to differentiate between one good/bad team and another. No. 9 Michigan State is all over the place, while No. 4 Oregon is steely and consistent. No. 72 Northern Illinois has shown some high upside and low downside; No. 81 Ole Miss has been consistently below average.
- Off. Standard Deviation and Def. Standard Deviation: This is the same thing, but we just look at the performance of each individual unit.
- Wtd. Margin: This is an attempted look at momentum. If we weight each progressive game as ten percent more important than the one before it, some teams' Adj. Scoring Margins change. We can compare Wtd. Margin to the full 2011 Margin to get the DIFF column, which signifies the difference between the two numbers. Teams with a positive "DIFF" have shown improvement as the season has progressed. Teams with negative numbers have not. No. 30 TCU is coming on strong after a rocky start, while No. 23 Georgia Tech, No. 33 Illinois and No. 40 Florida are dropping quickly.
- COVAR: This stands for Covariance, a look at which teams play better against good teams and which teams play better against worse. The higher the COVAR figure, the more likely you are to rack up great numbers against lesser teams but perform at a lower level against good ones. I played with this a bit this summer, but I'm still figuring out what it could mean in terms of predictive value. Teams like No. 26 Ohio State and No. 43 Missouri are on the high side (meaning they dominate when they have an advantage but don't play up to a higher level of competition), while No. 22 Miami and No. 42 Texas Tech have tended to play better against better teams. This doesn't mean the "playing up to competition" teams are better -- it could mean you just end up with competitive losses against good teams and competitive wins/losses against lesser ones -- but it is interesting.