This paper tests the house money effect, i.e., that prior gains increase subsequent risk-taking behavior, in the field. We use individual-level slot machine gambling records of 5,169 players from a real casino and employ jackpot hits as exogenous house money shocks. Our results show that players wager significantly less money after hitting a jackpot, which implies that players reduce their risk-taking behavior and thus act more cautiously after experiencing gains. We fail to replicate the house money effect in the field and find evidence for a reverse house money effect. Furthermore, while risk-taking behavior is reduced after hitting a jackpot, the jackpot size has no additional effect.