Online Program Home
  My Program

Abstract Details

Activity Number: 447 - Essential Uses of Statistics in Football
Type: Invited
Date/Time: Wednesday, August 2, 2017 : 8:30 AM to 10:20 AM
Sponsor: Section on Statistics in Sports
Abstract #322291
Title: Consistency in Professional Sports: The NFL's Parity Paradox
Author(s): Michael J. Lopez* and Gregory Matthews and Benjamin Baumer
Companies: Skidmore College and Loyola University Chicago and Smith College
Keywords: Football ; Bayesian modeling ; state space ; competitive balance ; simulation ; NFL
Abstract:

The National Football League (NFL) prides itself on parity; "every fan thinks that their football team has a chance" said Commissioner Roger Goodell a few years back. But is it parity? Or a function of the league's unbalanced schedule and small sample size of games?  Using a Bayesian state-space model within each of the four North American sports organizations, we propose a set of novel competitive balance measures, contrasting each metric to better understand the relative equivalence of NFL franchises. Next, using time-varying estimates of team strength, we estimate the impact of the NFL's unbalanced schedule on league standings by simulating season results under different fixed schedule designs. Altogether, results suggest that the NFL is less competitively balanced than has previously been suggested, with the league's scheduling format, in which teams play like-placed finishers from the prior year, accounting for upwards of half a win per year. These findings inform fans of sports in general with respect to the underlying randomness of game outcomes, and in particular supporters of football, where season-outcomes lie in part at the behest of inequitable organizational policies.


Authors who are presenting talks have a * after their name.

Back to the full JSM 2017 program

 
 
Copyright © American Statistical Association