The fragment docking program SEED (Solvation Energy for Exhaustive Docking) is evaluated on 15 different protein targets, with a focus on enrichment and hit rate. It is shown that SEED allows for consistent computational enrichment of fragment libraries, independent of the effective hit rate. Depending on the actual target protein, true positive rates ranging from 5% to 28% are observed at a cutoff value corresponding to the experimental hit rate. The impact of variations in docking protocols and energy filters are discussed in detail. Remaining issues, limitations and use cases of SEED are also discussed. Our results show that fragment library selection or enhancement for a particular target is likely to benefit from docking with SEED, suggesting that SEED is a useful resource for fragment screening campaigns. A workflow is presented for the use of the program in virtual screening, including filtering and post-processing to optimize hit rates.