Metrics can help the CFP committee

The 13-member College Football Playoff selection committee met in Washington, D.C., this week, and it was announced that it will reveal four interim rankings before the inaugural selection weekend in 2014.

Of course, a little light bulb immediately fizzled to life -- bizzz-fizzzz! -- over the heads of many in the ESPN offices in lovely Bristol, Conn. Interim rankings? Why wait until 2014? Let's do it now. As in right now.

That's what everyone wants, right? The playoff to start this year, not next. Or so you think.

The committee will be made up of 13 human beings, ready-made with their own ideas about evaluating and ranking college football teams. Many of them will be old-school, laboriously reviewing hours of game tape and giving teams their fearsome "sight test." Some will embrace the sabermetric revolution that has unceremoniously stomped on cherished sporting ideas such as clutch play, momentum and locker room chemistry.

For our purposes, the ESPN.com 2013 College Football Playoff Selection Committee will be made up of the guys who made A's in calculus in 10th grade and then moved on to the Jacobian conjecture: ESPN Stats & Information geniuses.

Those folks have devised an all-encompassing metric -- Championship Drive Ratings -- that measures the résumés of college football teams, putting a heavy emphasis on strength of schedule, as the CFP committee is expected to do. It measures how difficult it is for an average FBS team to achieve the team in question's results.

So, the season ends today, and based on Championship Drive Ratings -- drumroll, please -- the four-team playoff would go: 1. Alabama; 2. Stanford; 3. Florida State; 4. Ohio State.

That's right, unbeaten Baylor, you are eclipsed by Stanford, which lost to now-4-5 Utah on Oct. 12.

Got a feeling ESPN's Stats & Info's Q-rating just dropped precipitously in Waco.

The good news, of course, is the season doesn't end today. We were just making that up so Oregon fans could go, "Why couldn't you have done that a week ago?" Baylor's schedule is significantly backloaded, with a visit to No. 12 Oklahoma State on Nov. 23 looming large. CDR, although interesting to follow on a weekly basis, is ultimately about measuring the totality of the season because nothing is won in college football on Nov. 13.

Yet it is important, even at this juncture, to ask why an objective judgment of merit, such as CDR provides, ranks Stanford No. 2 and Baylor out of luck because it's the sort of distinction the CFP committee is going to make, and thereby make one or two or five fan bases livid.

Essentially, it was more difficult for Stanford to get to 8-1 against its schedule than for Baylor to get to 8-0. Or for Florida State and Ohio State to reach 9-0.

Stanford has three wins over teams in the CDR top 10. Only Ohio State has even one. The Cardinal are 5-1 versus the top 50. Baylor and Ohio State are just 2-0.

Further, the Utah loss is not as bad as it looks at first glance. The Utes have played the nation's toughest schedule: The combined FBS record of Utah's first nine opponents is 45-17 (.726). They have quality wins over BYU and Stanford. And they lost those five games by a combined 38 points.

Based on CDR, Utah is better than all but one team Baylor and Ohio State have played (Oklahoma and Wisconsin, respectively). The Utes are the fourth-best team Stanford played.

Ah, but we can hear you: eye test. What about how a team looks and how it has been consistently dominant against whatever schedule it played?

Although it's important that the committee emphasize strength of schedule, to measure merit and to push teams to schedule more aggressively, it also will be critical that the committee not get too bogged down in a tyranny of merit.

What that means is that often strength of schedule is unpredictable. The overall strength of a conference goes up and down, as does the quality of a nonconference matchup scheduled years in advance. Ohio State and Oregon will play a home-and-home series in 2020-21. That sounds like an A-list showdown from our present perspective, but it's also possible one or the other or both could go all Tennessee on us, which could mean folks lambasting the Buckeyes' strength of schedule again in 2021.

A tyranny of merit would mean automatically rejecting a team with an inferior résumé even if a majority of educated observers believed that team to be among the nation's four best and most talented teams.

So, let's introduce another ESPN Stats & Info metric: Football Power Index, which is a predictive measure of team strength that uses the expected points added breakdowns that correlate most with future results. It includes efficiency rankings that are the point contributions for offense, defense and special teams to a team’s opponent-adjusted scoring margin per game (overall EPA).

To be less math-y, it's about ranking how powerful teams are.

In this metric, the top four are: 1. Florida State; 2. Oregon; 3. Baylor; 4. Alabama.

Now the Stanford folks are mad.

This measure is all about dominance, and Florida State and Baylor certainly have dominated on both sides of the ball. Oregon was doing the same until the ill-fated visit to Stanford. And Alabama has dominated of late after a middling start.

Although it would seem reasonable to use CDR and FPI side by side, FPI at season's end might serve as a better metric for a responsibility of the CFP committee that is not presently getting as much attention as perhaps it should: seeding.

The Stats & Info folks, in fact, believe the CDR is best for selecting the four teams for the playoff -- merit! -- and the FPI could then be used to seed those four teams in terms of their accumulated and demonstrated "power" during the season.

In that scenario, the four-team playoff would go like this: No. 1 Florida State vs. No. 4 Ohio State and No. 2 Alabama vs. No. 3 Stanford.

Chew on that one for a bit. A first reaction might be a celebration in Tallahassee and Columbus that they won't face Nick Saban in the semifinal.

It also is worth noting a second time that these metrics, although interesting to follow through the weeks of the regular season, only reach their most informative state when all the games have been played. These metrics could end up looking significantly different on Dec. 7.

Although metrics assuredly won't be the only evaluative tools the committee will use -- it has yet, in fact, to define the evaluative process in any specific terms -- what it aspires to do is remove the emotions from the process.

So. Anyone out there have some emotions they'd like to share?