Most sports teams play as either the favorite or the underdog and either at home or away. The failure to recognize the symmetric and interdependent relations between these characteristics has led previous researchers to use potentially biased methods to test for rationality and efficiency in football betting markets and thus to reach inappropriate conclusions. We develop a more general specification, which also incorporates ‘pick-em’ games and games played on neutral sites, and find little or no evidence against market efficiency in the NFL and college betting markets for regular season games. We do, however, uncover evidence of biased betting lines for Superbowls.