Archive for January, 2017

JP College Football Rankings: How did we do?

by   |  01.11.17  |  Academics, COBA Faculty, Research

by guest bloggers Dr. Ryan Jessup and Dr. Don Pope

Well, the dust has settled on another bowl season and it is time to evaluate the success (or, lack thereof) of the JP ranking system.

You might remember from last season that we concluded that a blind squirrel would be about as equally effective as our ranking system.  Well, this season that blind squirrel (technically, he is blindfolded) made us look silly.

 

Figure 1. How well did we do? If last season we did about as well as a blind squirrel flipping a coin then this season that blind squirrel took our money.

 

You see, out of 42 bowl games we correctly predicted the winner approximately 54% of the time, and, in the 6 games in which we predicted a different winner compared to the college football playoff (CFP) selection committee, we were correct half the time and they were correct the other half.  However, we correctly predicted against Vegas 48% of the time.  So, this year the squirrel beat us.

What went wrong?  This is always a useful question.  One issue is that the Big 10 was a Big Letdown, finishing a miserable 3-7 in their bowl games, performing overwhelmingly worse than expected.  Likewise, Clemson outperformed expectations as they upset both Ohio State and Alabama, two teams that both we – and Vegas – thought would win.  Why these things occurred is rather difficult to determine: did the Big 10 perform well against non-conference games – which usually take place early in the season – and then fall off later on?  It is hard to know for sure.

Regarding Vegas, last bowl season the Vegas favorite covered the line approximately 60% of the time.  Interestingly, they only covered the line a mere 35% of the time this season meaning even they had a hard time predicting the outcomes.  But in the end, you can’t fight city hall, and, really, you probably shouldn’t mess with Vegas either as the house is truly playing with a stacked deck.

One thing that does give us hope is that, even though hundreds of millions of dollars are on the line for the college football playoff and a prestigious 13 member selection committee generates the rankings, our simple ranking system fares about as well, getting the same number of correct predictions.  So, maybe next year the CFP should drop their committee-based ranking system and just hire that squirrel.