An inside look at how the College Football Playoff creates its rankings

GRAPEVINE, Texas

In about two weeks, the College Football Playoff’s 13-member selection committee will convene at the Gaylord Texan Resort & Convention Center in Grapevine, Texas, to craft its first of six weekly rankings on Nov. 3. The final of those rankings on Dec. 6 will determine the 12-team field that will participate in the playoff for the right to be national champions.

Before that happens, though, the CFP invited a handful of media members to go through a mock selection last week to get an inside look at exactly what goes on inside the room where the voting happens. I was part of that group.

The committee’s process has (understandably) been questioned over the years in terms. From the outside, it seems as if what gets factored into the decisions changes regularly — and honestly, it’s a fair assessment. But I’ll also say this: It’s a tiresome process, one that involves hours upon hours of discussion and doesn’t always end with unanimous agreement.

And those internal discussions are only going to get prolonged moving forward, with the field all but formally set to double in size starting next season.

I’ll try my best to sum up the experience here, but be warned: There’s a lot that goes into it.

The format

We were tasked with creating a 12-team playoff field utilizing new requirements for this year based on the results from the 2025 season.

Those new requirements, for those who need a refresher:

▪ All four power conference champions, plus the highest-ranked Group of 6 team (not necessarily a conference champion) are guaranteed spots into the field. Last year, the five autobids simply went to the five highest-ranked conference champions. So for this exercise, Duke was automatically going to be added to the field as the ACC champion; the SEC’s Georgia, Big Ten’s Indiana and Big 12’s Texas Tech are the other autobids plus whoever we ranked as the top G6 team (American Conference champion Tulane and Sun Belt champion James Madison both made the field last year, with Duke left out).

▪ Notre Dame automatically gets into the field if it is ranked in the top 12, a clause that was not in effect last season. The Fighting Irish were ranked No. 11 last year but didn’t get in under the old rule because two autobids ranked below them had to be in the field. If the new rules applied, they would have taken Miami’s at-large spot last year.

▪ The remaining six or seven spots (depending on where Notre Dame factors into all of this) are given to the highest-ranked teams that don’t have an automatic bid.

The voting process

Before formal voting begins, each committee member is tasked with privately submitting a list of 30 teams the person believes should be considered for the ranking. Every team that gets votes from at least three members is included in the overall pool of contenders eligible to be ranked. We had 41 teams in our pool.

From there, teams are compared in groups of six and the rankings are done in chunks.

It starts with each member picking top six teams from the pool of eligible candidates. The six teams with the most votes are named and the discussion begins. The discussion at this point is only on these six teams; no one else in the pool can be ranked at this time.

After all six teams are discussed and compared, the pros and cons of each resume debated, committee members privately rank who they think the top four teams are from this subset of six teams. The top four teams after the blind vote are now Teams 1, 2, 3 and 4 in the rankings.

The two teams that did not get selected stay in the pool and the process repeats. Members pick their next six teams in the pool, with the four highest-voted teams added to the two leftovers from the first round. The six are discussed among each other, another ranking vote takes place and the next wave of teams get ranked.

For example, in our exercise, we had Georgia, Indiana, Ohio State, Ole Miss, Oregon and Texas Tech in our initial pool. Our top four was the same as the overall committee: Indiana No. 1, Ohio State No. 2, Georgia No. 3 and Texas Tech No. 4, although there was serious debate about the order of Ohio State and Georgia.

After that first vote, Oregon and Ole Miss remained in the pool to be discussed for the 5-8 spots, along with four new teams.

This process went on four more rounds to determine Teams 9-12, then 13-16, then 17-19, then 20-22 and finally 23-25.

This voting format is why the committee was able to say it couldn’t directly compare Miami and Notre Dame head-to-head early in the process despite the Hurricanes beating the Fighting Irish. Miami and Notre Dame were never in the same six-team pod during at least the first four rankings. The direct discussion between the teams never materialized until the end, when Miami finally overtook Notre Dame.

Now, after the top 25 are picked, conversations can be had about whether the committee wants to re-rank any cluster of teams next to each other. They call this the scrubbing process. If this does happen, the same process from above repeats but only with the cluster of teams selected. That cluster can cross pods, so if the committee wants to look at, say, teams 8, 9 and 10 even though the No. 8 spot was voted in from a different pool of candidates, that’s allowed.

After the ranking is finalized, teams are placed into their spots on the bracket. If there are teams with autobids outside the top 12, they get placed starting at No. 12 and work their way up. So based on last year’s final rankings with the new rules in effect, unranked ACC champion Duke would have been No. 12, Tulane (the highest-ranked G6 team both by the committee and in our round of exercise) No. 11 and Notre Dame (No. 11 in the committee’s final ranking) No. 10 — although in our version, the Fighting Irish replaced Alabama, not Miami.

If no G6 team cracked the top 25, the committee would debate the G6 teams available in the pool as it did the other ranking pods and vote for the one that will make the field.

The metrics used

As for how committee members come to their decisions and what shapes those decisions?

The committee’s directive is to “select the best teams, rank the teams for inclusion in the playoff, and then assign the teams to the playoff bracket.”

The phrasing “best team, not necessarily the most deserving team” was used multiple times throughout the exercise (even though the best teams probably should be the most deserving teams, but I digress). That wording was how they justified keeping an undefeated Florida State team out of the four-team field in the 2023 season and likely played a heavy role in a three-loss Alabama team getting into the field last season.

Committee members are given a slew of data points and metrics (some publicly available, others not so much) to help guide their decisions, from correlative stats (which break down a team’s efficiency) to schedule strength and record strength (determining which teams had tougher roads and how they handled them compared to other teams). They also have access to video of every game played — both full games and condensed versions — and, if all else fails, they’re allowed to use their gut and the all-famous “eye test” to help guide their decisions along with metrics.

That said, all the data the CFP provides is results-based and is updated each week. The committee doesn’t utilize predictive models like FPI or SP+. They are supposed to solely judge what happened on the field, not what might happen moving forward.

All this begs the question: Which metrics or guidelines take priority?

Does head-to-head determine everything (short answer, no but it is key when teams are in the same pool — see Miami and Notre Dame)? Is schedule strength valued more than the metrics of how a team won? Does a team losing to weaker teams outweigh wins over bigger opponents and a tougher strength of schedule?

Honestly, it’s all subjective to each individual committee member, which can make the process frustrating for teams since they don’t know year to year (or sometimes even week to week) what is prioritized when each ranking is made.

One quick final takeaway: At a certain point (for us, it started with the fourth round of voting, in the back half of the 13-16 pod), there is a noticeable drop-off in teams being considered. Team resumes at this point all have notable flaws, being it three-plus losses, a weak strength of schedule or poor performative metrics.

More changes coming

And the irony of doing this selection when it’s becoming widely apparent that the field and the format are about to change again was not lost on those in attendance. ESPN reported Thursday, hours before we were debriefed on the process that was about to unfold, that the network is nearing a deal with the CFP to expand to 24 teams starting next season.

Members of the CFP wouldn’t discuss an expanded playoff in absolutes during the meeting since nothing is finalized, but it was front and center on ESPN as we had breakfast before doing the exercise Friday morning.

That will cause changes for the committee moving forward, most directly in the form of likely having to expand the field it ranks.

But regardless of how many teams they rank, the process for ranking them remains the same.

It all leads to plenty of internal constructive conversations and plenty of outside scrutiny.

Those conversations, and the scrutiny that follows, formally begin again in two weeks.