2025 CFB Playoff Preview

Written By Ravi Betzig

Methodology

Longtime readers of my blog will know that my favorite picks I make all year are my preseason college football bets. Over the last month, I have posted lots of bets on win totals, odds to win the conference, etc. 

Last week the major sportsbooks posted odds to make the CFP for the first time this offseason. I was super excited for this, as these were my most profitable bets last year. If you were following my picks last year I’m sure you remember when I picked Indiana to make the playoff at 30/1 odds in September. The 12-team CFP is so new that I think a lot of people struggle to properly model it, leaving some nice opportunities.

To prepare for this, I spent some time this offseason developing some logic to simulate the playoff selection process. It’s hard to exactly replicate the logic that a human committee does, but I made my best guess. I won’t bore everyone with the details of how it works here, but if you’re interested in learning some details, feel free to shoot me a DM on Twitter. I’d be interested to discuss this with other people who have tried their hand at this problem.

Later this week I’ll post the bets I like for odds to make/miss the CFP. Today I’ll go through some big picture thoughts on how the playoff will shape up this year.



An Example Simulation

To start, I simulate the regular season 10,000 times. At the end of these simulations, I then guess who I think the committee would pick in that scenario. Here’s an example simulation to give you an idea of how it works.

In this simulation, the auto-bids go to major conference champions Clemson, Texas, Penn State and Kansas State as well as James Madison, who is the best G5 team. The first few at-large spots are quite clear: South Carolina and Michigan went 12-0 in the regular season before losing their conference title game, and Texas A&M went 11-1 but missed the SEC title game.

For the last at-large spots, the model selected the teams that lost the Big 12 and ACC title games (Arizona State and Georgia Tech, respectively) who were each 11-1 prior to that loss. It went with 10-2 Auburn and Tennessee for the last spots, leaving out 10-2 Illinois, SMU, Miami and Louisville. This was on the basis of strength of schedule- when comparing power conference teams with the same record, it goes with the one that played the hardest schedule.

Keep in mind that this is just one simulation, and in different simulations, different teams outperform and underperform expectations. In this particular simulation, South Carolina was one of the best teams in the country (not an impossible scenario given LaNorris Sellers’ upside) and Georgia stumbled to 8-4. In the simulation after this one, Georgia steamrolled everyone and went 13-0. This is why I simulate the season 10,000 times- to capture a wide array of possible outcomes. 

Note that I built this logic before auto-byes for conference champions were scrapped. Since I’m just trying to predict selection in the playoff and not seeding, this is fine for now, but I’ll fix it later.


Expected Playoff Bids by Conference

There’s been a lot of discussion as to how many teams each conference can expect in the playoff. Last year’s breakdown was as follows:

Big Ten: 4 (Ohio State, Penn State, Oregon, Indiana)

SEC: 3 (Texas, Georgia, Tennessee)

ACC: 2 (SMU, Clemson)

Big 12: 1 (Arizona State)

Mountain West: 1 (Boise State)

Independent: 1 (Notre Dame)

My gut feeling last year was that the Big Ten was not going to get 4 teams in most years, and they got lucky there was such a gap between the top four teams in the league and everyone else. Similarly, I think the SEC will probably get more than three teams in a lot of the time.

Based on 10,000 simulations of the playoff, here is the expected number of bids for each conference this year:

SEC: 3.45

Big Ten: 2.98

ACC: 2.01

Big 12: 1.84

Independent (inc. Notre Dame): 0.70

G5 Conferences: 1.02

I think most people would expect the Big 12 and ACC to be lower and the Big Ten and SEC to be higher. We’ll see how it plays out in the future, but I think last year’s ACC and Big 12 had an unusual amount of parity.


Expected Playoff Bids by Record

Alabama nearly got into the playoff with a 9-3 record last year. Do we expect 9-3 SEC teams to make the playoff in the future? Could we ever see an 8-4 team in the playoff?

We can answer these sorts of questions with my 10,000 simulations of the playoff. I have divided teams into different “categories” of resume. Not all teams within a category are created equal (10-2 Ohio State is a shoo-in for the playoff since they play Texas, 10-2 Indiana is unlikely to make it with a weak non-conference schedule), but it serves as a general guide.

After looking at a lot of individual simulations, here are a few general rules I believe in:

  1. Any power conference team with 0 or 1 losses will make the field. Last year’s 11-1 Indiana is about as weak a 1 loss team as you’ll ever see, and they ended up making it in with some room to spare.

  2. 11-2 teams in the at-large pool (i.e. teams that go 11-1 in the regular season and lose their conference title game) are a good bet. SMU is an example of such a team from last year. The absolute weakest 11-2 teams might miss the field. For example, a hypothetical 11-2 BYU team that lost the Big 12 title game last year might have been left out.

  3. In most simulations, the last spot is a battle between 10-2 power conference teams. 10-2 teams with stronger schedules (e.g. most SEC teams, a Clemson team that plays LSU and South Carolina out of conference, etc.) are in good shape. 10-2 teams with weaker schedules (e.g. Illinois, NC State) are in bad shape.

  4. About half the simulations feature a 9-3 team in the playoff. Only 9-3 teams that play one of the absolute toughest schedules in the country have any real shot. 

  5. 8-4 teams have no shot at the playoff.

In the median simulation, we have 14 power conference teams with double digit wins at the end of the regular season. In both the 2023 and 2024 seasons, we had 14 power conference teams with double digit wins at the end of the regular season. This was one of the key metrics that I looked at to tune the simulation and I think is a good sign that it matches reality pretty well.

Simulating the playoff selection process is an interesting problem. I used the results above to inform my bets on teams to make/miss the playoff and I’ll be back later in the week with some of my favorite bets.

Ravi Betzig