For efficient structure-guided drug design, it is important to have an excellent understanding of the quality of interactions between the target receptor and bound ligands. Identification and characterization of poor intermolecular contacts offers the possibility to focus design efforts directly on ligand regions with suboptimal molecular recognition. To enable a more straightforward identification of these in a structural model, we use a suitably enhanced version of our previously introduced statistical ratio of frequencies (RF) approach. This allows us to highlight protein-ligand interactions and geometries that occur much less often in the Protein Data Bank than would be expected from the exposed surface areas of the interacting atoms. We provide a comprehensive overview of such noncompetitive interactions and geometries for a set of common ligand substituents. Through retrospective case studies on congeneric series and single-point mutations for several pharmaceutical targets, we illustrate how knowledge of noncompetitive interactions could be exploited in the drug design process.