This study investigates the efficiency of National Football League betting markets from 2002 to 2015 utilizing three rest patterns (Thursday Night Football, post-bye week, and regularly scheduled games). The logistic model indicates that inefficiencies due to the rest patterns occurred in Thursday Night Football and post-bye week games, with bettors expecting high rates of return. The forecast errors also provide evidence that point spreads did not correctly forecast future outcomes. Nevertheless, this deviation from efficiency was not sufficient to make positive returns from Thursday Night Football when teams of record were tested separately. A bias from post-bye week games remained profitable across the studied years, suggesting previously confirmed biases were not temporary. Although further analyses with a larger sample size are needed to confirm whether Thursday Night Football creates a bias consistently, this study suggests that the betting market misses important information about rest intervals and team performance.
JEL classification: z23