Abstract: We investigate experimentally how the amount of information on an opponent’s ‘toughness’ affects the chances that a conflict over scarce resources between two individuals results into a ‘fight’. The design aims at capturing situations, such as those found in prisons, in which there are no enforceable rules of allocation and in which ‘might is right’, and to find out how much order vs violent conflict emerges. We measure toughness by asking subjects to do a wall-sit for as long as they can resist. We ask them to do the exercise twice, once ‘veiled’, when they do not know that a contest will follow and once ‘unveiled’, when they do know. The information on how long they resisted in both exercises is then revealed to the opponent who decides whether to challenge or ignore. If he challenges the other player may yield or resist. If he resists a fight ensues and yields a winner and a loser. We conjecture that when information on fighting prowess grows fighting decreases, exploitation does not increase, and fighting does not decrease among equally ‘tough’ subjects. Some (prison) policy implications are drawn.