My college football model comes from my background (Ph.D. in Chemical Engineering from Stanford) and uses play-by-play and market data to make spread predictions. From 2022 through 2024, the model went 52.6% against the opening market (907-816 with 29 pushes in games in which prediction differed from the market by a point or more).
My process involves using these numbers along with subjective factors that the model doesn’t consider to make long-term profitable bets. I do not recommend betting a market as difficult as college football spreads based on my numbers alone.
Here are five predictions for Week 3 of the college football season to help with your handicapping:
Odds courtesy of BetMGM.
No. 6 Georgia (-3.5, 49.5) at No. 15 Tennessee
The Power Rank prediction: Pick-em
Tennessee started the season with a 19-point win over Syracuse as a 13.5-point favorite. QB Joey Aguilar, a transfer from Appalachian State, threw for 8.82 yards per pass attempt against the Power 4 opponent (6.18 is the college football average).
In contrast, Georgia hasn’t impressed in two games against inferior competition, with a 45-7 win over Marshall before a dud of a 28-6 victory over FCS Austin Peay. QB Gunner Stockton has thrown for 7.19 yards per pass attempt, but it’s hard to rate this pass offense without adjusting for opposition (a calculation I can start next week).
In the preseason, my model would have made Georgia about a 6-point favorite. On Sunday morning, Georgia -6.5 was available, which seems to suggest that neither team has changed in two weeks.
Since then, the market has moved towards Tennessee. This probably reflects a downgrade for Georgia, as Tennessee has the exact same rating now as it had in the preseason in my model.
My model is probably making too large an adjustment for Georgia based on these two games against inferior competition. In addition, Tennessee will be without top CBs Jermod McCoy and Rickey Gibson III, two players who starred last season but have yet to play in 2025.
Bet: Pass
No. 16 Texas A&M at No. 8 Notre Dame (-7, 49.5)
The Power Rank prediction: Notre Dame by 10.6
QB CJ Carr started his first game for Notre Dame against Miami (FL) in the opener. He did fine with 5.97 yards per pass attempt against — probably — a solid defense, but he also added 40 rushing yards on eight scrambles.
Marcel Reed took over at QB for Texas A&M in the middle of last season, and he was great for a true freshman. After adjustments for opposing defenses, my metrics predict Reed to throw for 7.01 yards per pass attempt against an average FBS pass defense.
Texas A&M most likely has the edge at QB, but Notre Dame should have the better defense. My prediction includes an extra 4.2 points for both home-field advantage and rest advantage (Notre Dame didn’t play last week but Texas A&M did).
Bet: Lean Notre Dame -7
Kansas State (-1.5, 54.5) at Arizona
The Power Rank prediction: Kansas State by 3.8
It has been a rough start to the season for Kansas State. After a loss in Dublin to Big 12 rival Iowa State, the Wildcats lost a home game to Army last week as a 17-point favorite.
There is no doubt that Kansas State should be downgraded, especially since they only beat FCS North Dakota by 3 in their other game — the only question is by how much.
My college football model still thinks the Wildcats are good enough to win and cover at Arizona. In addition, there is a possibility they get RB Dylan Edwards back. He rushed for 7.4 yards per carry last season before getting hurt on the first play in Dublin.
Bet: Kansas State -1.5
Wisconsin at No. 19 Alabama (-20.5, 46.5)
The Power Rank prediction: Alabama by 19.3
After a dismal showing to open the season at Florida State, Alabama rebounded with a 70-3 pounding of Louisiana Monroe. Top RB Jam Miller has yet to play this season and probably misses this game.
To my surprise, Wisconsin has declined the past two seasons under Luke Fickell, a coach who worked wonders at Cincinnati. A rebound this season seems murky as QB Billy Edwards Jr, a transfer from Maryland, got hurt in Week 1 and is questionable to play in this game.
Bet: Pass
Florida at No. 3 LSU (-8.5, 48.5)
The Power Rank prediction. LSU by 7.5
Florida closed as an 18-point favorite at home against South Florida. However, the Gators couldn’t avoid the upset in a game in which both teams had similar total yards and yards per play.
After a solid opening-week road win at Clemson, LSU also struggled last week but managed to beat Louisiana Tech by 16 points (closed as a 36-point favorite). My model downgraded LSU more than Florida.
The market opened LSU -9.5 but this line has moved towards Florida.
Bet: Pass
Ed Feng runs The Power Rank, a football analytics and betting information site.
