ABSTRACT

For NFL football fans, comparing team abilities and speculating on the outcomes of team matchups is one of the most enjoyable aspects of following the game. The comparison of teams can be a source of intense argument, with discussions commonly peppered with anecdotes and cherry-picked information to support the contention of one team being better than another. Various web sites promote a semblance of scientific rigor by computing power rankings, numerical team ratings, and other related quantities that are ostensibly reliable measures of team strength. But if interest centers on a serious study of NFL team strength, how can we appeal to a solid statistical foundation to measure ability and predict game outcomes?