UCLA

Sports betting can be lucrative for some while unfavorable for others, but what if you could tilt the odds in your favor? This thesis helps uncover whether or not machine learning can accurately predict who will cover the spread in NFL games. I investigated which combi- nation of game statistics, box scores, power ranks, and Elo ratings would yield the best results. Additionally, I tested three different machine learning models with varying levels of interpretability and predictability. In the end, I found that I can correctly predict who will cover the spread enough times to slightly tilt the odds in my favor.