Football

Abstract

NFL modeling is simplified by excluding S(i,t-τ), the vector of lagged statistical shocks, from the exploratory relation in (3.2.1: Part II):

$$D(i,t){\text{ }} = {\text{ }}{f_{it}}[L(i,t),G(i,t - \tau )',x'(i,t)]$$

(1.1.1)

. This simplification is necessary since lagged effects of unknown statistical shocks cannot be reliably estimated with such outcomes per team. Effects of S(i,t-τ) are examined in basketball and baseball where per team models are based larger sample sizes.