At present, the drug-discovery process centres on ligands that either block or produce physiological responses. However, there are therapeutic uses for ligands that do neither of these things, but which still affect receptors in other ways. This review discusses the intimate relationship between the affinity of a ligand for its receptor, and the probability that the binding of the ligand will produce some change in the receptor, resulting in efficacy. This, in turn, argues that ligands that have affinity should be tested more broadly, for a wider range of efficacies, to detect hidden therapeutic activities.