Our CEO, Greg Godbout, published an article to describes the his journey in creating an AI predictive model to beat NFL point spreads and over/under bets. He emphasizes the iterative nature of AI development, highlighting the importance of understanding domain data, analyzing trends, and learning from failures. Despite setbacks with initial theories, he pivoted to a simpler approach using Lasso and Polynomial regressions. The model focuses on games with higher AI confidence, aiming for gradual improvement throughout the NFL season. The author plans to refine the model based on live game results. This is part 1 of a multi-part series.
For more details, you can visit the original article here.
