This paper examines the efficiency of the “in-play” Australian Rules football fixed odds betting market at quarter time, half time, and three quarter time. Tests of semi-strong efficiency are performed on the 2009 AFL season using logistic regression analysis. The results demonstrate that as the match progresses, there is a significant bias against the team that is currently leading—a bias that increases when the team that is leading is also the away-favorite. This bias is shown to yield significant profits utilizing simple betting strategies.