On Sunday afternoon, the College Football Playoff Selection Committee awarded the sport’s fourth and final playoff spot to Alabama over Ohio State. It’s clearly the most controversial pick of the committee’s four-year tenure. One immediate question is this: What four-team playoff field would the old BCS Standings have produced? The answer is: It’s complicated.
It’s unlikely that the BCS standings ever would have been used to select the playoff field without any regard for conference championships. If, for example, the top-4 in the BCS standings included multiple teams from a single conference, and included only one or two conference champions, the conference commissioners and most fans presumably wouldn’t have thought it acceptable to simply take the four highest-ranked teams. It is therefore likely that, by rule, the fourth and final selection would have been required to be a conference champion, or at least a non-champion whose conference wasn’t already represented in the top three slots.
This would have been a sensible arrangement for two reasons: First, it would have helped ensure an intersectional playoff field and hence greater national interest (not to mention that it would have protected against any short-selling of a particular region or conference). Second, for conference championships to mean anything (at least in this context), non-champions should presumably have to clear a slightly higher bar.
These criteria would have been fulfilled by requiring non-champions to make the top three, not merely the top four—at least if another team from their conference had already made the field. With all of that said, what would the BCS standings have looked like this season?
Approximating the BCS Standings requires using the coaches poll, the AP poll (the best stand-in for the Harris poll, which was created by and for the BCS, but no longer exists), and four former BCS computer rankings: Anderson & Hester (which I co-created), Billingsley, Colley, and Wolfe. The two other BCS computer rankings, Sagarin and Massey, no longer publish the version of their rankings that met the BCS’s requirement that they not be based on margin of victory.
The BCS dropped the high and low computer rankings and averaged the four in the middle. Because the highest and lowest of the four existing BCS computer rankings might have been dropped, or might instead all have been among the four that were averaged (with Sagarin and Massey being dropped), I tallied the four computer rankings each way—once with the high and low rankings dropped, once without dropping any of the four—and then averaged those two tallies. This seems like the best approximation of what the BCS computers would have yielded.
Here’s how the top 12 spots in the BCS standings would have looked at the end of this weekend, with each team’s point-value listed:
Approximate BCS Standings (out of 1.000):
#1) Clemson (12-1): .984 (#1 in the polls, #1 in the computers)
#2) Georgia (12-1): .939 (#3 in the polls, #2 in the computers)
#3) Oklahoma (12-1): .927 (#2 in the polls, #4 in the computers)
#4) Alabama (11-1): .846 (#4 in the polls, tied for #6 in the computers)
#5) Ohio State (11-2): .841 (#5 in the polls, tied for #6 in the computers)
#6) Wisconsin (12-1): .785 (#6 in the polls, #5 in the computers)
#7) UCF (12-0): .732 (#10 in the polls, #3 in the computers)
#8) USC (11-2): .730 (#7 in the polls, #8 in the computers)
#9) Auburn (10-3): .689 (#8 in the polls, #10 in the computers)
#10) Penn State (10-2): .649 (#9 in the polls, #9 in the computers)
#11) Miami (10-2): .615 (#11 in the polls, #12 in the computers)
#12) Washington (10-2): .528 (#12 in the polls, #14 in the computers)
So, if a rule had been in place that the fourth and final selection to the playoff field had to be a conference champion, or at least a non-champion whose conference wasn’t already represented in the top three, then the final pick for the four-team field would have been Ohio State. If the top four had simply been taken as is, then the final pick would have been Alabama—by .005.
The biggest difference between the BCS’s rankings and the committee’s rankings involves UCF, the nation’s sole undefeated team. The Knights, who amassed a 12-0 record versus a perfectly average schedule (.500), would have been ranked #7 by the BCS. The committee stingily ranked them #12.
Some college football seasons are conducive to a two-team playoff (a championship game), while others are conducive to a four-team playoff. This season was better suited to a four-team playoff—but not because there are clearly four teams that are a cut above the rest. Rather, it’s because there are three one-loss teams—Clemson, Georgia, and Oklahoma—that aren’t separated by a whole lot, and in order to keep all three of them in contention for the championship, a marginally deserving fourth team had to be included, too. It would have been nice if that final selection had provided some representation from the nation’s best conference this season, or from somewhere not in the lower-right quadrant of the United States.
As it is, one has to wonder what the upside of the playoff system is for the Big Ten and Pac-12. Their champions (Ohio State and USC), not invited to the playoff, could instead have played each other in the Rose Bowl—which would have been a satisfying outcome.
But that bowl—the best of bowls—was stolen by the playoff.
