Many animals engage in contests with conspecifics for access to resources. Understanding which resources individuals are contesting for, and what influences the outcome is central to our understanding of contest behaviour. We initially observed female bees of the genus Tetralonia (sp. n.) aggressively competing for access to burrows in the ground, without any clear indication of exactly which resources were being contested, or what factors might predict the outcome.We then individually marked bees and assigned ownership of burrows to individuals, before observing over 100 aggressive interactions. After excavating burrows cast with molten wax, we concluded that burrows were nests for provisioning larvae. We found that ownership (as putatively designated by us) had no influence on contest outcome, but rather that the position of the bees in the burrow was decisive; whichever bee was already in the burrowwhenthe contest beganwonthe vast majority of interactions. Furthermore, bees that were designated ownership of a burrow did not engage in longer fights for possession of that specific burrow, indicating either that they were not committing any kind of ‘Concorde fallacy’ by basing decisions on past investment, or that the assignment of ownership was incorrect. Instead, fights were longer later in the day, presumably as the value of the burrow as a refuge from the cold and/or predation increased as night approached. Nest parasitism does not seem to be a common strategy in this species, as owners were not more likely to attempt ejecting an intruder than vice versa. This indicates that contest settlement may not always follow theoretical predictions, but rather that insights and inferences into a species' ecology can be made from observations of dyadic contests.