Wednesday, December 28, 2011

According to Massey Ratings, Calvin Underperformed Early This Season, Have Overperformed Recently

I won't pretend to know all of the intricacies of Kenneth Massey's "Massey Ratings" (one of the computer rankings used in college football's BCS rankings), but I do know most the basics. For each game (college basketball, in our case), the score, venue, and date are used as inputs into the computer system. The computer then does a massive 'best fit' calculation of sorts for every game played in the college basketball universe, and comes up with a rating for each team.

For a real explanation, visit Massey's FAQ and Theory pages.

One of the more interesting aspects of his ratings website is perhaps the predictions that his ratings allow. For instance, we see that Calvin is projected to have only a 16% chance of beating Whitworth tomorrow, and that the 'most likely' score is 78-68 in favor of the Pirates. That's cool and all, but what I'm going to look at are his retro-active predictions for the Knights.

As the season goes on and more data points are added to the system, the ratings for each team are adjusted, and so are the predictions for the games that have already happened. For instance, just before Calvin and Finlandia met on the eve of Thanksgiving, Massey said that Calvin was expected to win by 25 points. Now, a month later, the prediction page says that Calvin should win by 14.

I went back through the nine games that the Knights have played to see how the actual score lines up with the current "retro" prediction:

DateOpponentHome/AwayExp. MarginAct. MarginOverperform/Underperform
Nov 15AndersonAway-4-13Underperformed by 9 points
Nov 18Grace BibleHome14Overperformed by 3 points
Nov 19WillametteHome95Underperformed by 4 points
Nov 23FinlandiaHome148Underperformed by 6 points
Nov 25CornerstoneAway-10-17Underperformed by 7 points
Nov 26AquinasNeutral-6-6As Expected
Dec 2CarthageHome-23Overperformed by 5 points
Dec 13WheatonHome-9-5Overperformed by 4 points
Dec 10Trinity ChristianHome1220Overperformed by 8 points

As you can see, the Knights have out-performed their projection four times, hit it exactly once, and underperformed four times. The encouraging news is that the underperformances came early, while the overperformances have mostly been in the recent games*.

*Note that the only early-season overperformance was versus Grace, a game that they "should have lost" twice and needed two overtimes to finally win.

Here's a line graphs that shows us mostly the same thing as the chart above.

In the first five games, they performed (on average) 4.6 points worse than their current rating predicts. In their last four games, they've performed 4.3 points better than their rating predicts. Compare this graph to Hope's graph, which shows some 'first half' to 'second half' (of the season to date) improvement (with respect to current Massey rating, not with respect to beginning or current team talent level), but in a much less obvious fashion than Calvin's does.

I'm hoping this is a sign that our young Calvin Knights are indeed getting better as they gain experience, and that the team they'll be next month is vastly different than the team they are right now, or that they were a month ago.

Want to stay up to date on the happenings at FFTMAG? Follow me on Twitter, “like” us on Facebook, and grab our RSS Feed. Need a ride to a game? Check out the Ride Board and post your needs.

No comments:

Post a Comment