So this post from the Sports Economist pops up in my feed reader linking to a paper just put out by Ohio State's econ department by one Trevon D. Logan, Assistant Professor of Economics. And hey, it's titled...
WHOA, NELLIE! EMPIRICAL TESTS OF COLLEGE FOOTBALL'S CONVENTIONAL WISDOM
- it's better to lose early than lose late,
- teams are rewarded for beating strong opponents, and
- "style points" count.
Contrary to conventional wisdom, I find that (1) it is better to lose later in the season than earlier, (2) AP voters do not pay attention to the strength of a defeated opponent, and (3) the benefit of winning by a large margin is negligible.Wow! It's time to check this paper out.
[Checking] [Jeopardy Music] [Etc.]
Oh, Lord, people from Ohio State are... well, they're not good at doing things. First we must establish what people mean by these old saws. Logan does a fine job of this:
The conventional wisdom of college football dictates that teams who lose early in the season stand a better chance of being highly ranked at the end of the season than teams who lose later. The logic is that teams who lose early have a greater opportunity to climb up in the polls after a loss, and also a greater chance of leapfrogging teams that lose at later points in time. Also, since ranking in the polls reflects recent performance, it is better to avoid losses late in the season. Similarly, the wisdom holds that voters view late losses unfavorably as they are a signal of low team quality.End of the season. End of the season. End of the season. So how does Logan test this? He gets a database of various AP rankings over approximately the last twenty-five years. The alarming bit: he proudly notes "this is the first study that looks at weekly data for a large number of teams."
Wait... weekly data? When we're testing the theory that losing early is better for your end of season rankings, which come out, you know, once a year?
I test for the conventional wisdom by looking at the relationship between game characteristics and changes in AP point-totals. Since teams play one game only between rankings, this strategy will capture the relationship between game characteristics and AP point changes. In particular, I test the conventional wisdom outlined above withLet's say Michigan loses to Appalachian State in week one. To determine how badly this impacts your ranking in January, this paper checks to see what your ranking is in week two.(2) E(P(t) − P(t-1)) = Γβwhere I regress the change in AP points from week t-1 to t on the characteristics of the game played between t-1 and t.
Ohio State, ladies and gentlemen!