We test for a relation between football match results and the specific national stock index returns during the period 1990–2006 by means of an event study approach. We employ two different econometric frameworks to cross-check our results and prevent them from being solely model driven: the constant mean model and a two-state Markov-switching market model. Both approaches find no significant results. Consequently, in a modified setup, we control for expectations about probable game results by applying a “surprise” variable, which is computed from betting odds and is integrated into a regression analysis. Again, there does not seem to be a connection between a specific national soccer team’s win or loss and stock index prices. In addition, through a few modifications in our empirical setup, we show how easy it would be to “produce” significant results. Our results are contrary to those of Ashton et al. (Appl Econ Lett 10:783–785, 2003) and Edmans et al. (J Finance 62(4):1967–1998, 2007) and support market efficiency.