I agree with those of you that there should be some sort of computer element to the CFP, as there was in the BCS days. For those of us who feel like that, you’ll enjoy reading this interview/oral history with the folks who were responsible for the BCS computer rankings.
This is my favorite bit:
The men were awkward, sure, but they were also outspoken, especially when the BCS changed the rules it used to govern the computers and the rankings in general. In 2002, it decided to drop margin of victory from the computer polls after Washington and Oregon had seen 11–1 campaigns marked by close, hard-fought wins come up short in the eyes of the rankings in consecutive seasons. Mike Tranghese, at the time the commissioner of the Big East, summed up the move this way: “A computer can’t get at the nuances of a score.”
Sagarin: “Somebody should have raised their hand and said, ‘Well just what are the nuances of a 59–3 score?’”
Sadly, I can see that being something the selection committee actually tries to fumble around and answer.
Get The Picture 
I don’t know. It was jarring to see the algorithms’ fetish for Ohio State after two got-pushed-around-and-just-whipped losses.
LikeLike
That was a really good article. Thanks for linking to it.
I liked the computers sort of like the judges in Olympic sports. Throw out the highest and the lowest rankings and give them a weighting in the ranking system (let them have margin of victory in their formula if so designed). Take the human polls and give them a larger weighting. Throw in a variable for strength of schedule with a small weighting (say 3/5, 2/5 and 1/5). Top 4 are in.
The WWL can still have their weekly taking heads show to unveil the rankings, and 13 people with day jobs don’t have control of the entire sport.
With the law of large numbers, you generally get to the best answer.
LikeLike
Math is hard … let’s just say 60%, 30% and 10%.
LikeLike
Boom approves. He thinks a ‘nuance’ is a low country chicken pot pie.
LikeLike
The BCS rankings were terrible at choosing 2 teams, but more than likely ideal for selecting a playoff.
LikeLike
How about this let the Selection Committee hire a really good computer firm to write a program that the Selection Committee designs to take into account all the “nuance” they want. It beats the hell out of the politics that selects the top four now. Also they are already looking at volumes of data if they are to be believed so why not computerize the whole data mess.
LikeLike
The last thing I want to do is side with the UCF whiners, but I think there needs to be some form of screen to prevent a non-divisional winner from essentially getting a bye when the conference championship amounts to a play-in game.
In other words, not a total block…….let’s assume Auburn won the division with 1 conference loss and 2 non-conference losses (3 total), thereby essentially limiting them to the role of spoiler for Georgia. In that case, I think it’s perfectly reasonable for 1 loss Alabama to get the nod. But with the divisional winner Auburn controlling their own destiny and then only losing said destiny BECAUSE they played an extra game (by virtue of actually beating Alabama), it does stink a little bit.
LikeLike
Create an 8 team playoff made up of 8 conference champs and settle it on the field. Do away with the beauty pageant aspect of the committee or polls or computers. 8 12 or 14 team conferences would encompass every team that has any realistic shot at winning a championship. Let the leftovers drop down to FCS where they probably belong anyway.
LikeLike
This is a great idea in concept but is completely impractical in today’s game. Today’s mid-major conferences (sorry, UCF) have no business taking up a slot in an 8-team playoff when there is a more deserving non-conference champion who has played a P5 schedule. Also, Notre Dame isn’t joining a conference in football without its NBC contract coming along with it. To make this work, you would have to radically realign conferences to get some level of competitive balance. Given the fact that the conferences act as independent entities, there’s no way to do this.
LikeLike