MODELING NFL FOOTBALL OUTCOMES

The purpose of this research is to develop models for NFL football games which help identify the important factors in determining outcome game. A variety different methods will be used developing these models: ordinary least squares regression, logistic and proportional odds. Models explain an they also predict based on differences certain in-game statistics.
 Values from thirty-five statistics both teams playing each game were collected 752 over three seasons between years 2011 2014. Data first two time period data last season was test models. found websites, Scores Pro Football [1] [2].