Market Efficiency and a Profitable Betting Rule

Abstract

This article presents empirical tests of efficient markets on the professional football market for totals from 1979 through 2000. The forecast errors in the totals market were found to be skewed, so a test of a fair bet on National Football League totals was performed using a log likelihood ratio test as suggested by Evan and Noble. The under on the total is a winning proposition more than 50% of the time, but a fair bet cannot be rejected. A betting rule that looks at totals in the high end of the totals distribution and takes a contrarian strategy of betting unders for games that are 7,6, or 5 points above the mean is shown to violate a fair bet. Rejection of no profitability, when incorporating the take of the sportsbook, is also rejected for the highest of the totals.