Abstract
Gambling markets for sporting events provide opportunities to examine factors that influence the forecast accuracy of prediction markets in a broader sense. In this study, we use information from the betting market on NCAA men’s basketball contests to analyse the rate at which markets incorporate changing information into pricing. Using data from the 2009 to 2022 seasons, we investigate how turnover in player personnel, coaching, and conference affiliation between seasons impacts the accuracy of betting lines. We find some evidence that coaching changes and, to a lesser extent increased player turnover, lead to decreased accuracy and a greater amount of movement in market lines, but that these effects are generally only evident for the first few weeks of the season.