Back to the basics

I've been using this model to pick my men's and women's March Madness moneyline bets. In the men's tournament, it has gone 38-14. In the women's tournament, it has gone 38-6 thus far. The model is a neural network that is trained on the team stat differentials from all March Madness tournament games from 2010 - 2025.

You can train and run it yourself in less than 5 minutes following the instructions in the linked repo.

These are my picks for today's NCAAW tournament. In general I took the best moneyline price I could get across all of my outs unless it was shorter odds than about -300. In that case, I took the favorite against the spread.