From an employer's perspective a tournament should induce agents to exert productive activities but refrain from destructive ones. We experimentally test the predictive power of a tournament model which suggests that – within a reasonable framework – productive and destructive activities are not influenced neither by the number of agents taking part in the tournament nor by the fraction of the winner prizes. Our results clearly confirm that sabotage in tournaments indeed occurs. While tournament size has virtually no effect on behavior, a balanced fraction of winner and loser prizes seems to particularly enhance productive activities.