We are now officially through the regular season, and the College Football Playoff selection committee has tabbed the four teams that will play for the national championship and the eight other teams that will complete the field playing in the New Year’s Six games this year.
Just as has been the case over the past few weeks, the BCS rankings would have chosen the exact same top four in the same order as the CFP committee decided to produce:
Comparing the first dozen spots in the CFP rankings to the BCS projections, we find that no team is ranked more than two spots apart between the two lists. The furthest gap for an individual team between the two sets of standings is four spots, occurring first with Oklahoma State (16th in the CFP, 12th in the BCS) and occurring twice more with Tennessee and Temple.
More interesting is seeing how these numbers correlate to the in-house Pigskin Rating System numbers. Seven of the teams featured in the CFP top 25 are ranked outside the top 25 by the PRS numbers; the biggest disparity in this realm is with Northwestern, who is ranked 13th in both the CFP and BCS projections but is found all the way down in 58th in the PRS calculations.
Teams like Michigan State, Iowa, Northwestern, and Oregon are undervalued by the PRS relative to the CFP and BCS numbers. Conversely, teams like Ole Miss, Michigan, USC, and Baylor are perceived by the PRS algorithm to be far better than either the human voters of the CFP or the various processes used to calculate the BCS have given credit throughout the year.
Ultimately, the reason I have continued to calculate the BCS numbers is not merely to show how it compares to these other systems, but to show that there is no logical reason why we should have completely scrapped the attempt at objective analysis in favor of a dozen selectors sitting in a room crafting their opaque poll.
No matter how long that Jeff Long sits on ESPN explaining the picks each Tuesday, the methodology used to explain the selections is always based on a moving target. Ostensibly based solely on the games that have already been played up to the date of the poll release, the words that come out of Long’s mouth each week have proven otherwise.
For some teams (specifically teams representing blueblood schools with respected names and extensive championship pedigrees), the benefit of the doubt allows the committee to argue that their best is yet to come. For other programs (specifically nouveau riche schools, traditional walkovers, and mid-majors galore), the full weight of analytical scrutiny is thrown at them in order to justify deflating their standing.
Sure, the semifinals and the plus-one championship game are far more desirable than the one-versus-two final that was the sole stated purpose for the BCS. But as the BCS rankings for this season show, the system would have done essentially the exact same job as the committee in figuring out how to select and seed the top four teams.
When 12 people are held entirely responsible for determining the endgame of the college football season, with proprietary data and a complete lack of objective transparency, the gains that a bracket provide are offset by the fact that there is no discernible way to double-check the committee’s logic. With the BCS we could calculate ourselves whether the numbers were correct.
Now there are no numbers, just the vacuous decisions of a group that has shown the exact same sort of recency bias and proclivity to pick powerhouses as the historic polls. The only difference is that the AP and coaches polls provide a far larger sample size of national opinion.
In our search for a definitive national champion, we have expanded the field of candidates while making the selection process far less objectively definitive. With that in mind, here are what the numbers and the pecking order would look like if the BCS still reigned supreme as the means of determining the most deserving playoff teams of the year.
Leave a Reply