Recent legislation has changed the landscape of college sports, a multi-billion dollar enterprise with deep roots in American sports culture. With the recent legalization of sports betting in many states and the SCOTUS O’Bannon ruling that allows athletes to be paid through so-called “Name-Image-Likeness (NIL)” deals, evaluating talent and projecting results in college sports is an increasingly interesting problem. By considering both talent accumulation and recent on-field results, our models aim to predict relevant results for sports betting/team construction. In this iteration of the project, our targets are regular season win percentage (using a season level model that we’ll call Model 1) and individual game results (with a game by game model we’ll call Model 2) in the regular season. Our datasets come from a variety of sources including On3, ESPN, 24/7 Sports, The College Football Database, and SportsReference.com.