Affiliation:
1. Kogod School of Business, American University, Washington, District of Columbia 20016
Abstract
This paper tests the efficiency of a set of sports betting markets using detailed betting line movement from opening until closing for four different sportsbooks for each of 3,681 Major League Baseball games. The reliability of the markets’ forecasts are assessed at several lead times. The forecasts are mostly reliable, but there are simple betting strategies that would have yielded significant profit. In addition, forecasts do not always improve monotonically as the games get closer, despite more information being available; forecasts at weekend day games’ start times are significantly worse than forecasts 90 minutes earlier. Furthermore, analysis of the sequences of forecasts within individual games reveals that these betting markets do not incorporate information optimally. There is sufficient evidence to reject weak form market efficiency; specifically, betting lines tend to overreact, exhibiting significant negatively autocorrelated changes that could be exploited by sophisticated bettors. This paper was accepted by Manel Baucells, behavioral economics and decision analysis. Supplemental Material: The data files are available at https://doi.org/10.1287/mnsc.2022.00456 .
Publisher
Institute for Operations Research and the Management Sciences (INFORMS)