# Logistic regression and game winning probabilities

I'm researching how machine learning can be applied to predicting outcomes of games such as football and basketball with logistic regression. I was wondering if anyone had a brief overview of the steps involved from a theoretical perspective and any associated literature. Specifically how one constructs the input vector for two teams. And how one determines the classification categories. I'm guessing the classifications are 1) team A wins 2) team B wins. Please forgive my ignorance like I said just started researching this may be way off base.

• This question might get better answers at stats (dot) stackexchange (dot) com. $\qquad$ – Michael Hardy Feb 13 '17 at 23:19