Benchmarking sets are important for testing virtual screening algorithms. There are several kinds of benchmarking sets for several separate kinds of tests. Pose Prediction test sets are drawn from crystal structures in the PDB. Here the goal is to have the docking program reproduce the crystal structure. Enrichment test sets test the ability of a virtual screening program to enrich known actives relative to assumed (or known) decoys. Pattern Matching test sets are used for testing substructure based methods of enrichment.