This study compared virtual screens against a homology model of a G protein-coupled receptor and the experimentally determined crystal structure. Following determination of the crystal structure of the dopamine D3 receptor, but before the information was publicly available, the scientific community was asked to predict the structure of the D3 receptor based on homology models. Remarkably, a docking screen against one such homology model was as effective at prioritizing active D3 ligands — with respect to hit rate, potency and novelty — as a screen against the subsequently released crystal structure.