Can data really predict who’ll take home the NCAA trophy before the season’s even over?
Spoiler: it can—and it does.
In State College, Pennsylvania, the home of Minitab headquarters and Penn State University, fall means one thing: football.
As the air cools and the playoff race heats up, it’s the perfect time to crunch some numbers and see which team’s performance won’t “turn over” when it matters most.
To predict the 2025 national champion, I pulled data from sports-reference.com, a treasure trove of stats going all the way back to 1869 (yes, really!) when Princeton ruled the gridiron. But college football has changed—drastically. So instead of digging through dusty history, I focused on the College Football Playoff (CFP) era, which began in 2014.
Fun fact: Since 2014, every national champion has been ranked no lower than #5 in the final November AP Poll.
Translation: by November, the true contenders are already in view.
That insight led me to focus only on Top 10 teams, trimming the noise from the non-contenders.
My mission: predict the probability that each Top 10 team will win it all. Using classification models like Logistic Regression, Random Forests, and TreeNet, I trained Minitab to think like a football coach—with a spreadsheet.
Here’s what went into the model:
In short: everything that separates champions from teams that just look good in September.
I exported stats from 2014–2024 and sent them through the Minitab Data Center for cleaning and prep. With Minitab AI allowing much of the clean-up to be done through natural language plus the ability to save and automate repetitive steps, what used to take hours now runs like a well-oiled offense.
Because let’s be honest—no one wants to clean up 10 years of football data by hand.
Once the data was ready, I stacked it in Minitab Statistical Software and used: Automated Machine Learning > Discover Best Model (Binary Response)
That’s Minitab’s Predictive Analytics solution for quickly testing multiple models and identifying the one that scores best.
The winner? Random Forest. It delivered the most accurate predictions for “Winner” vs. “Loser.”
The Random Forest model uncovered the four biggest predictors of championship success:
Takeaway: Defense still wins championships—confirmed by data, not just coaches.
Drum roll, please…
Using 2025 midseason stats for the Top 10 teams, I calculated each team’s probability of winning the title and standardized the results so they total to 100%.
As of November 12th, 2025, the model predicts:
Ohio State and Indiana have the best shot at the national title. Things are looking good for the Big Ten. Sorry, Georgia—this might not be your year.
To make the results easier to follow, I built an interactive Minitab Dashboard.
With Minitab AI, it’s as simple as connecting your data source—charts, tables, and model results automatically combine into one dynamic view. It’s like your own version of ESPN’s College GameDay—minus the mascot heads.
Because the dashboard updates automatically as new data comes in, you can instantly see which teams are trending up or falling behind. It’s a real-time, data-driven way to follow the championship race.
In short: stop guessing. Start modeling.
This analysis proves what football fans already know: championships aren’t decided by hype or headlines.
Defense, discipline, and data win championships—on the field and in business.