clock menu more-arrow no yes mobile

Filed under:

Better Know a Computer Poll: Billingsley

If I were to say, "Hey, let's look at a fun way of rating college football teams with a computer!"  You would probably walk away, but since this particular computer determines who plays in big time bowl games, let's take a look.

The first, and perhaps only, subject of our Better Know a Computer Poll segment is the Billingsley Report.  For those of you who know the computers as "Sagarin and those other ones," you can distinguish the Billingsley Report in that it's the only computer that ranks Texas in the Top 10.  Let's take a closer look at the inner workings of the Billingsley Report, which can be found at Mr. Billingsley's website.

Unlike Sagarin, who is a mathematical genius, and Wes Colley of Colley's "Bias Free" Matrix who is a Princeton math professor, Mr. Billingsley claims merely to be "a devout College Football Fan."  It's almost fitting then that his computer rankings function less like a computer poll in the usual sense and more like a human poll.

His ratings, unlike most, are not a percentage, but instead a running total of points, which teams can gain or lose.  Teams gain points for winning and lose points for losing, the amount of points being determined by the strength of the team played at the time of play.  Teams see diminishing returns on their wins for every loss on their record, so Ohio State would gain more points for beating the #10 team than Texas would and Texas gains more points for beating them than LSU would.  Billingsley starts the teams off at the positions at which they were at the end of the previous season awarding a certain number of points to each position, so Texas had a 1 point head start over USC coming into the season.  In the first four games of the season, teams can gain more points than in games the rest of the season to compensate for the unequal start, so from here on in teams in the Billingsley rankings have more limited mobility than they do in the other computer polls, like in the human polls.

Other features of note

For SOS he considers the opponents rank both at the time of the game and currently slightly favoring opponents rank at the time the game was played, thus he gives Louisville (#10 in his rankings) more credit for beating Miami than most computers do.

The rankings are built to reward consistency, so teams pulling big upsets can only hope to rise so far with just one game

Three factors have a tiny effect on the rating and serve as tie breakers.  These are, in order of importance:  W/L record, Away games, and holding opponents under 6 points.

The reason Texas is so high in this poll is because of the limited mobility and the head start.  We did have our only loss during the period of heightened mobility and have that loss detracting from our point total for the rest of the season, but it's good enough to put us above all other 1-loss teams except Auburn and Cal (just barely) and above Louisville, though below West Virginia.

I've tried to translate his explanation into a more understandable form and relate it to Texas, but, as you can see, this is all very confusing.