Saturday, January 18, 2020

My Plan for a Much Better College Football

Now that a season is behind us, let's dream of a better world. In this case I am limiting the thought experiment to the competitive structure of the college football game leaving aside the ever-important and (thankfully) increasingly popular discussions of allowing players to more fairly benefit financially from their contributions.

The three key features I would envision for the highest level of college football are:

  1. Create a true Power 5 - an elite 60-team top division of football comprised of FIVE 12-team conferences
  2. Continue the 12-game regular season - standardize 8 conference games and a maximum of 2 non-conference opponents outside of Power 5
  3. Crown a champion using an 8-team playoff  - conference title game winners get automatic playoff berth; 3 at large playoff teams determined by CFP committee (see below)
The CFP committee would be comprised of 25 members chosen by the conferences (each conference gets 5 members). The committee members' individual votes would be secret but the discussions in committee would be eventually fully released to the public after the season and playoff concluded. There would be no abstentions or exiting of the room when a member's affiliated school was discussed or any other charades to pretend there isn't bias. Rather, biases would be out in the open for all to examine and take into account. 

In regards to the establishment of a Power 5, we have to consider who is in and who is out. There are currently 130 FBS football teams (the highest level in college football). It is clear this is too many if we mean to group teams competitively. Just about any early-September match up reveals this outside of the handful of games designed to be competitive. Much of conference play itself reflects Have vs. Have Not. 

I would suggest a system where by teams bid to own a seat in the Power 5. The process might be something like the highest 60 bidders would pay the 60th highest bid amount plus 10% of their own bid amount into the pool. The bottom 70 teams would receive out of the pool 50% of their bid amount. Essentially, we would want to elicit from teams good information about how much they value football and how valuable football is to them. The remainder of the pool would be returned to the winning bidders in proportion to their bid size.

For example, say Texas is the top bidder at $200 million, Washington is the 60th bidder at $60 million, Kansas is the 61st bidder at one dollar less than $60 million, and Akron is the 130th bidder at $15 million. In that case Texas pays in $80 million ($60mil plus 10% of $200mil), Washington pays in $66 million, Kansas receives out $30 million (50% of ~$60mil bid), and Akron receives out $7.5 million.

Teams would be able to sell their seat with 50% of the sale price going to the selling team and the remaining 50% equally distributed to the other 59 seat-owning teams. Again, we want to know who should be in the Power 5 on an on-going basis and have a good incentive system to do so. Perhaps this is too capitalistic for American sports, which have always had a strong socialistic tendency as opposed to European sports where relegation rules soccer and revenue sharing and spending limits are less prevalent. But were dreamin' here . . .

A comment on the playoff to determine a champion every season: There exists a tension between a judgment-based approach (AP poll, BCS, CFP committee) and a merit-based approach (computer polls, conference champions feeding playoff). We can have a certain process of crowning a champion or a certain outcome of who is crowned. These two goals cannot both be realized in all cases. Usually there is a tradeoff between them. If we don't clearly define it ex ante, the ultimate criteria people want to use is fluid and subject to biases and inconsistencies. If we do clearly define the criteria out front, we tie our hands. Either approach has a degree of feature and bug. Remember before you condemn any outcome (actual or hypothetical) that we are dealing with sample sizes of 13 just to get into the playoff if we keep a 12-game season plus conference title game. For only the two teams in the championship game of an 8-team playoff do we even get to a 16-iteration test. We NEVER know as much about how good teams actually are as we think we do. In fact one could argue that in today's college football we know less than that implies given that so many games are nonsense gimmes against gravely inferior opponents. Because of this the only system I think we can justify is a structured conference feeder to a multi-team playoff. We can't possibly know who the best team is or even who the best teams are. The best we can do is define the process of earning a title and hope (with confidence) it matches up for who the best teams are most of the time.

No comments:

Post a Comment