Can the CFP selection committee be fixed ... or replaced? Let's dive into the data

Eighteen months from the start of the 2026 College Football Playoff, the event has no format, and the people in charge of putting one together are hitting the reset button on the topic.

Thats a reset , not quite a restart .

Everybody involved seems keen on the idea of expanding the 12-team model that will be used this season, which itself features a tweak on the inaugural 12-team Playoffs seeding structure .

Sixteen teams is the most likely number, because bigger is the easiest thing to agree upon .

Though it should be noted: Everything is subject to change right now.

Advertisement How to pick teams and fill the bracket is the real sticking point of the moment.

I think what we need to do, from our perspective, is we need to take a step back, Ohio State athletic director Ross Bjork said during a news conference with media last week in Columbus.

Theres so much work that has to be done.

We believe in what we believe in.

Were willing to listen, and I think thats a position that we have to take, but playing it out in public and staking claim, what good does that do? A Big Ten-backed plan to populate the bracket with mostly automatic qualifiers from the power conferences, determined by league standings and play-in games, has been discussed for more than a year, but now there is renewed interest within the SEC in a format heavy on at-large bids.

The SECs recent pivot away from AQs at its annual spring meetings caught some in the Big Ten off guard.

One week it was 4-4-2-2-1, Bjork said, referencing the distribution of automatic bids among the Power 4 leagues in the Big Tens preferred expansion model .

That had a lot of traction that was out publicly, and then the next week, it was 5+11, but have we accomplished anything yet? No.

So lets play it out where it should be played out in these conversations, where the authority lies.

The authority lies with the 10 FBS commissioners and Notre Dames athletic director, who make up the CFP management committee.

More precisely, the power lies with the SEC and Big Ten .

Regardless, the larger group will gather in Asheville, N.C., this week for its fourth in-person meeting since the end of last season to talk about the not-so-distant future of the College Football Playoff.

The focus will be on the selection process, which everybody seems to take issue with in some form or another, while politely acknowledging that selection committee members are making a good-faith effort to do a difficult job the best they can.

Advertisement I think anything we can do to make the postseason more objective and less subjective is going to be better, Florida athletic director Scott Stricklin said at SEC meetings in Destin, Fla.

Whether you could do that some other way than automatics? I dont know, but I think thats the goal.

Not surprisingly, every conference commissioner looked at the selection committee rankings last year and thought their leagues teams should be ranked higher.

In the SEC in particular , leaders have asked questions about whether the committee is properly weighing strength of schedule.

We learned something the first time through (the 12-team playoff), SEC commissioner Greg Sankey said.

And that raises the need for deeper analysis and understanding.

If were gonna just incentivize wins, playing fewer winning teams can get you to more wins.

I dont think thats great for football.

So how to fix the selection process? The possibilities include: Running the numbers The SEC concluded its meetings by handing out a seven-page document, filled with charts and graphs to visualize various popular advanced metrics used to evaluate and rank college football teams.

The conclusion, according to the SEC, was that the depth of quality in the conference has been unmatched over the past decade: In general, SEC teams were facing the toughest competition.

Among the handful of rating systems cited was SP+, a forward-looking measurement that evaluates teams on a play-by-play basis, sprinkling in a programs recent history and recruiting rankings.

The irony of the SEC using average SP+ rating over the last 10 years to indirectly make the argument that its teams were given short shrift by the selection committee last season is that the ratings creator disagrees with the conferences premise.

It was definitely conflicting hearing the SEC refer to me as the one of the reasons why it should get more Playoff teams when I say the opposite, said Bill Connelly, who created SP+ 17 years ago and started working for ESPN in 2019.

Advertisement Connelly has a derivative of SP+ that he says more accurately equates to selecting Playoff teams.

His resume ranking includes raw results of games i.e., wins and losses which SP+ does not.

While the final SP+ ratings of 2024 placed Mississippi at No.

2 and Alabama at No.

4 neither made the CFP field the SP+ resume rankings on selection Sunday were not quite so bullish.

Alabama was ninth, good enough to squeeze into the field over SMU using just those numbers; the committee chose the Mustangs over the Crimson Tide after SMU lost the ACC Championship Game to Clemson on a last-second field goal.

But even using the resume ratings, Ole Miss would have missed the 12-team field at No.

11, squeezed out by conference champions Clemson and Arizona State of the Big 12.

CFP selection committee members do not lack statistics to help them evaluate teams.

The 13 committee members receive a large binder packed with information on each FBS team, provided by SportSource Analytics.

Offensive numbers.

Defensive numbers.

Strength of schedule.

Its a lot.

Maybe too much? Connelly, like many media members, has taken part in the annual mock selection sessions run by the CFP to provide a window into its process .

He said the side-by-side comparisons of teams in the CFP binders, with green shading showing which team has the advantage in each category and red shading highlighting deficiencies, is probably not the best use of visualization.

Not all components of performance are created equal.

It can be a useful exercise, but its definitely worse than using good advanced stats, Connelly said.

Consolidation could help.

But Connelly said even he wouldnt recommend using just one rating system but rather a combination of several.

Brian Fremeau, whose FEI rating system has been ranking college football teams since 2006, echoed Connellys sentiment.

Even he doesnt take his own numbers as the last word.

Advertisement Theres definitely an assortment of (rating systems) that Id bring to the table, not a single one, said Fremeau, whose ratings use possession-level data to evaluate teams.

I believe theres a single data source that the committee uses, and I dont say this as distrust of them or that kind of source, but I actually think a variety of sources would be just a better approach.

Because one source, and Ill say this about my own stuff, doesnt tell the whole story.

The NCAA mens college basketball committee has leaned into a more metrics-based approach in recent years, with the NET rating fueling a quad system to evaluate wins and losses, and committee members regularly citing the popular KenPom.com efficiency ratings.

Connelly cautioned that this type of a system wouldnt translate well to football, where teams play only 12 regular-season games, compared to 30-something in basketball.

Not enough data points.

The CFPs staff has been evaluating the data provided to the selection committee and expects to present some ideas to the managers this week.

Ultimately, though, the commissioners need to decide what they want to emphasize.

Once upon a time, the Bowl Championship Series used a compilation of computer ratings and human polls to determine which two teams would play for the national title.

The formula underwent near-constant tweaks, eventually to a point where human polls were heavily weighted and the computer ratings acted almost as a tiebreaker.

The reason why everybody thought the committee was a good idea is because everybody hated the BCS formula, Connelly said.

The reason everybody hated the BCS formula is because it could only pick two teams.

If youre picking 16 now, or 11 at-large, whatever we end up getting here, that formula, that exact kind of formula approach where you mix subjective and objective, would work fine.

Last year, a simulated BCS formula would have produced a very similar set of rankings to the selection committee, with the exact same teams ranked in the top 16 and few differences in the order.

Here it is, what weve been waiting for the season-ending simulated #BCS rankings.

How will the @CFBPlayoff rankings compare? pic.twitter.com/tHvmrDOEVK BCSKnowHow.com (@BCSKnowHow) December 8, 2024 Fremeau said he wouldnt want the CFP selection process to be devoid of human input.

The margin between teams is often so slim it requires a well-informed panel to make a decision.

I wouldnt want people who dont understand the data at all, he said.

I dont think they should be bound by: The number said this, and therefore its the only thing.

(Photo: Todd Kirkland / Getty Images).

This article has been shared from the original article on theathleticuk, here is the link to the original article.