2025 College Football Analysis, Part Five: The Group of Five

So far our math-informed exploration the 2025 college football season has taken us back in time to 2024, explored the schedule of the Michigan State Spartans, and delved into the Big Ten, SEC, Big 12, and ACC races. Today's stop takes us to the remainder of Football Bowl Subdivision (FBS) conferences, with a brief detour looking at the likely accuracy of all these picks.

In the 12-team college playoff, the highest ranked champion from the Mountain West, American Athletic Conference, Conference USA, the Sun Belt, or the Mid-American Conference will also earn a spot on the bracket. 

But which team will it be? Let's pay a quick visit to each conference and once again make some math-driven predictions.

Mountain West

The analysis of each conference will take the same form as the analysis of the Power Four conferences in parts four and five of this series.  Table 1 below shows the detailed results of my preseason simulation of the Mountain West race, including the consensus preseason rank of each team, two sets of projected final records, strength of schedule data, and the odds for each team to make and win the conference championship game and the odds to make and win the 12-team playoff.

As in the last two parts of this series, the most likely scenario is the one where the favorite teams win all games. The disruptive scenario modifies the most likely scenario to inject a historically accurate number of road upsets into the overall schedule.

Table 1: Summary of the preseason projections for the Mountain West conference, based on the consensus preseason rankings and a 50,000 cycle Monte Carlo simulation of the full college football season.

Boise State (67%) have the best overall odds to qualify for the Mountain West Championship Game. This is largely due to the fact that the Broncos are almost 30 slots ahead of the second-highest ranked Mountain West team, the No. 63 UNLV Running Rebels. The Rebels (33%) lead the pack of potential Mountain West spoilers which includes No. 78 San Jose State (30%), No 85 Fresno State (16%), No. 89 Colorado State (14%), and No. 91 Air Force (12%).

UNLV has the second best raw odds to make the Mountain West final, but the Rebels have the misfortune of drawing Boise State (-11.4) on the road in conference play. Due to this game on the schedule, UNLV is picked in both the most likely and disruptive scenarios to finish in third place behind Boise State (8-0) and the San Jose State Spartans (8-0).

The Spartans are blessed with the easiest Mountain West conference schedule that includes no conference road games against any team ranked in the top 100. Boise State has the second easiest conference schedule and would be a 11.5-point favorite over the San Jose State on a neutral field. Boise State is therefore the clear favorite to win the Mountain West.

American Athletic Conference

Table 2 below provides the same set of data for the America Athletic Conference (AAC) as Table 1.

Table 2: Summary of the preseason projections for the America Athletic Conference, based on the consensus preseason rankings and a 50,000 cycle Monte Carlo simulation of the full college football season.

If the Mountain West Champion does not claim a spot in the college football playoffs, the champion of the AAC is the next most likely option. The top of the league projects to be very competitive with six teams with between a 21% and 39% chance to make the AAC Championship Game.

No. 47 Tulane (39%) hold the best overall rank and odds with No. 61 Navy (38%), and No. 59 Memphis (29%) rounding out the top three. It is also notable that those three teams have three of the four easiest schedules in the preseason.

No. 68 Texas San Antonio (23%), No. 72 Army (22%), and No. 67 South Florida (21%) makes up the tier of secondary dark horse contenders. This group are both considered slightly less talented than the teams in the lead pack and they also possess more challenging schedules.

In both the most likely and disruptive scenarios, Memphis (8-0) is able to narrowly defeat both Tulane (+0.5) and Navy (+4.1) in the final month of the season en route to a perfect conference record, leaving Tulane (7-1) and Navy (7-1) in a tie for second place.

Tulane (-3.1) projects to be favored to beat Memphis on the neutral field. Therefore, the Tulane Green Wave are my pick to win the American Athletic Conference.

That said, The Green Wave will have to avoid an upset on Nov. 1 at Texas San Antonio (+1.7). If Tulane drops this contest, Memphis would become the favorite to win the conference.

Sun Belt

The summary of the simulation result for the Sun Belt Conference is shown below in Table 3.

Table 3: Summary of the preseason projections for the Sun Belt Conference, based on the consensus preseason rankings and a 50,000 cycle Monte Carlo simulation of the full college football season.

The Sun Belt remains the only FBS conference that maintains a two-division structure. In the East Division, the No. 62 James Madison Dukes (38%) have the strongest preseason ranking and thus the best overall odds to win the division.

James Madison also has the advantage of playing the projected three other best teams in the East Division at home: No. 82 Georgia Southern (21%), No. 97 Old Dominion (11%), and No. 100 Appalachian State (10%). This translates to the Dukes being the clear favorite to win the East Division.

In the West Division, there is less separation between the teams. No. 81 Louisiana (22%) has the highest preseason rankings and also the best odds to win the Division. However, No. 87 South Alabama (22%), No. 88 Texas State (18%), No. 95 Troy (19%), and No. 98 Arkansas State (12%) are also expected to contend.

In the most likely scenario, Louisiana (6-2) loses road games at James Madison (-8.3) and at South Alabama (-2.0) while South Alabama (6-2) loses road games at Texas State (-3.4) and Troy (-1.2). In this case, South Alabama wins the head-to-head tiebreaker with Louisiana and would advance to the Sun Belt Championship Game.

In the disruptive scenario, there is no change to the record of South Alabama (6-2), Louisiana suffers two additional close road losses at Troy (+0.4) and at Arkansas State (+1.4) to finish conference play at 4-4.

In both cases, James Madison and South Alabama (+6.3) advance to the Sun Belt Champions where the Dukes would be favored. Thus, James Madison is my pick to win the Sun Belt.

Mid-American Conference

The summary of the simulation result for the MAC is shown below in Table 4.

Table 4: Summary of the preseason projections for the MAC Conference, based on the consensus preseason rankings and a 50,000 cycle Monte Carlo simulation of the full college football season.

Five total MAC teams have between a 10% and 52% chance to make the MAC Championship game. No. 69 Toledo (52%) leads the way in odds, followed by No. 83 Ohio (40%), No. 94 Buffalo (34%), No. 90 Miami of Ohio (27%) and No. 104 Northern Illinois (16%).

The three teams with the highest odds own three of the four easiest conference schedules, while Miami of Ohio and Northern Illinois drew two of the three toughest conference slates.

As a result, in both the most likely and disruptive scenarios, Toledo (8-0) and Buffalo (8-0) to run the table in conference play and face each other in the MAC Title Game. Ohio projects to lose a close road game at Buffalo (-0.5) to close out the regular season. 

My model projects Toledo as the MAC Champion, as I project the Rockets to be a 6.7-favorite over Buffalo on a neutral field.

Conference USA

Table 5 below provides the same set of data for Conference USA.

Table 5: Summary of the preseason projections for Conference USA, based on the consensus preseason rankings and a 50,000 cycle Monte Carlo simulation of the full college football season.

Similar to the MAC, Conference USA has five teams that all have between a 19% and 65% chance to advance to the conference championship game. Also like the MAC, Conference USA has two clear favorites with relatively easy conference schedules.

No. 74 Liberty (65%) and No. 96 Western Kentucky (43%) lead the way with No. 105 Sam Houston State (29%) projected with the third best odds and the most likely spoiler. No. 111 Jacksonville State (20%) and No. 114 Louisiana Tech (18%) are the most likely dark horse contenders.

In both the disruptive and most likely scenarios, Liberty (8-0) runs the table in conference play. Western Kentucky projects to be favored in every conference game, but in the disruptive scenario the Hill Toppers get tripped up in two tough conference road games at Louisiana Tech (+3.4) and at Jacksonville State (+21).

Western Kentucky (6-2) finishes in a second-place tie with Sam Houston State (6-2) in the disruptive scenario, but wins the head-to-head tiebreaker. Either way, I project Liberty to win the Conference USA Title, as the Flames project to be a 6.1-point favorite over Western Kentucky as a neutral field.

Group of Five Overview

Now that we have reviewed and made selections for each Group of Five conference, Table 6 below summarizes the 25 teams from the Group of Five which are most likely to earn the fifth automatic playoff bid. Table 6 also gives the most significant non-conference for each team. An upset win in the non-conference would certainly bolster the playoff resume of any of the teams listed below.

Table 6: Overview of the Playoff Odds and Marquee Non-Conference Games for the 25 most likely Group of Five Teams to make the playoffs.

Based on the picks above, the Selection Committee would need to pick between

• Boise State (12-1) with a loss at No. 6 Notre Dame
• Tulane (11-2) with losses at No. 59 Memphis and at No. 16 Ole Miss
• James Madison (11-2) with losses at No. 74 Liberty and No. 27 Louisville
• Toledo (12-1) with a loss at No. 46 Kentucky
• Liberty (13-0) with the best win over No. 62 James Madison and the worst strength of schedule in the FBS

On paper, Boise State is the strongest team on this list and would likely end the season as the highest ranked Group of Five team, even with Liberty being undefeated. That said, there are a few interesting other scenarios.

For example, if Tulane could survive the road test at Memphis, if Toledo were to upset Kentucky on the road to open the season, or if James Madison were to run the table, those three teams might get a look from the committee.

Memphis is not projected to win the AAC, but the Tigers will have a realistic chance to upset Arkansas on Sept. 9 which could provide some momentum going into conference play. Similarly, UNLV might even be favored to beat UCLA at home on Sept. 6 which could similarly provide a boost.

Last Year's Picks in Review

Now that all we have projected the outcome for all ten FBS conferences in 2025, let's take a quick look back at how accurate the picks were last year using the same methodology.

Table 7 below shows my picks for the 2024 conference title game participants using both the most likely and disruptive scenarios. The correct picks are shaded in yellow and listed on the right side of the table. The odds for both the projected and actual winners, based on the 2024 preseason simulation, are also shown.

Table 7: Summary of last years projections for the conference title game participants using the both the most likely and disruptive scenarios. The odds preseason odds for both the projected and actual winners (shaded yellow) are also provided.

The same data for the winners of each conference are given in Table 8.

Table 8: Summary of last years projections for the conference winners using the both the most likely and disruptive scenarios. The odds preseason odds for both the projected and actual winners (shaded yellow) are also provided.

In regards to conference title game participants, my predictions using the odds (i.e. the most likely scenario) got seven of the possible 18 picks correct. The analysis using the disruptive scenario made six correct picks. The disruptive scenario correctly picked Penn State to make the Big Ten Championship Game and UNLV to make the Mountain West Championship Game. However, this scenario mistakenly picked Louisville, Bowling Green, and Ole Miss to make the respective conference title games.

On the conference level, both scenarios got just two picks correct out of nine champions. Both scenarios correctly picked Boise State out of the Mountain West. The disruptive analysis got the Oregon pick correct while the most likely scenario correctly picked Georgia in the SEC.

While this success rate may seem low, the sum of the odds in each table gives the preseason expected value for the predictions, and those numbers are well within expectations. My analysis only expected to get 7.1 picks correct on the division level (and it got six or seven correct). At the conference level, 2.6 picks was the expectation, and each model made two correct predictions.

As a final note, Table 8 also shows that while the odds-based lens suggests that teams with at least a 30% chance to make their respective conference championship game is most likely, the reality is a bit different.

Of the 18 conference title game participants, seven of the teams had preseason odds no better than 20%. Arizona State (0.9%), Army (9%), Marshall (11%), and Iowa State (14%) all started the season with odds below 15%. This year will likely play out similarly. College football, as always, is full of surprises.