Factor Analytics was born in March 2024 on the UConn campus, right in the middle of their back-to-back national championship run. Everyone around us was filling out bracket pools, picking teams based on gut feelings and mascot preferences. We figured there had to be something more systematic underneath all of this.
One search for “Four Factors basketball” later and we were deep in efficiency stats, pace adjustments, and schedule strength corrections. What started as an attempt to win a bracket pool turned into a full projection model, then a second independent model to cross-check the first, then a backtesting framework to see which disagreements between the models were actually predictive.
The framework has since expanded beyond college basketball into the WNBA, and NFL data is being gathered and tested. Today the system runs fully automated, scraping scores and lines every morning and overnight, updating every team’s ratings, and generating fresh projections for every upcoming game on the slate.