How the field is picked
Twelve teams. Five get in automatically: the champions of the ACC, Big Ten, Big 12 and SEC, wherever they're ranked, and the highest-ranked team from the American, Conference USA, the MAC, the Mountain West, the Pac-12 and the Sun Belt, whether or not it won its league. Notre Dame is also in if the committee ranks it in the top 12. The other six or seven spots go to the next highest-ranked teams.
The bracket follows the committee's final ranking straight down. The top four seeds get first-round byes, and an automatic qualifier ranked outside the top 12 is placed at the bottom of the seeding. Seeds 5 to 8 host first-round games on campus on Dec. 18 and 19. There's no reseeding and no moving teams to avoid rematches.
The No. 1 seed picks its quarterfinal bowl and its semifinal site, then No. 2 and No. 3 pick their quarterfinal bowls and No. 4 gets the one that's left. The quarterfinals are the Fiesta Bowl (Dec. 30) and the Cotton, Peach and Rose bowls (Jan. 1); the semifinals are the Orange Bowl (Jan. 14) and the Sugar Bowl (Jan. 15). The title game is Monday, Jan. 25, at Allegiant Stadium in Las Vegas, the Raiders' home.
What changed from 2025: the automatic bids used to go to the five highest-ranked conference champions, which let two Group of Six champions in ahead of ACC champion Duke, and Notre Dame had no guarantee. The 13-member committee, chaired by Arkansas athletic director Hunter Yurachek, ranks the top 25 on Nov. 3, 10, 17 and 24 and Dec. 1, and sets the field on Sunday, Dec. 6, between noon and 3 p.m. ET. When teams are close, it's told to weigh championships won, strength of schedule, head-to-head results and results against common opponents.
Method
We simulate the rest of the season 20,000 times. A game's expected margin is the gap between the two teams' ratings in ESPN's Football Power Index plus 2.5 points for the home team (none at a neutral site), with a standard deviation of 14 points, and each simulation nudges every rating up or down a little (standard deviation 3 points) because the ratings are estimates too. FCS opponents are rated at -20.
Conference title games come from the simulated standings: the top two by conference winning percentage (the Sun Belt's two division winners), with ties broken by head-to-head and then by rating, a shortcut for each league's longer tiebreaker list. Power 4 and MAC title games are at neutral sites; the others are at the top seed's stadium.
To pick a field in each simulation we need a stand-in for the committee. Ours ranks teams by wins above what a playoff-level team (the 12th-best FPI rating) would expect against the same schedule at the same sites, plus a small credit for team strength and half a win for a Power 4 title, and it counts a title-game loss as half a loss. Then the format's rules are applied. It isn't a model of the committee and won't match its weekly rankings; it's a way of counting how often each team ends up in.
The field on this page is our call, built from those counts: each Power 4 league's most likely champion, the most likely Group of Six team, Notre Dame, then the at-large teams that make the field most often. The seeds are our read of how the committee would order those teams. They follow the simulations' average ranks, except for Notre Dame, which our stand-in underrates because it has no 13th game to play.
One adjustment this week: Ohio State's rating is four points lower than FPI's because Julian Sayin is out indefinitely. Four points is about the gap between FPI's number for Saturday's game against Oregon and the market's.