Homologous recombination DNA repair deficiency (HRD) is a functional defect in homologous recombination DNA repair, arising from germline or somatic mutations in BRCA1/2 or other mechanisms. Cells with HRD are more sensitive to platinum and poly(ADP-ribose) polymerase inhibitors (PARPi). HRD generates permanent changes in the genome with specific, quantifiable patterns (“genomic scars”). Clinical tests for HRD, such as the Myriad genomic instability score and Foundation Medicine loss of heterozygosity test, aim to predict the presence of HRD based on genomic features. Clinical trials of PARPi in ovarian cancer have evaluated genetic mutations and HRD genomic assays as potential biomarkers of response. Patients with HRD due to BRCA1/2 mutations are more likely to respond to PARPi than those with wild-type (WT) BRCA1/2. In some clinical trials, patients with WT BRCA1/2 who were predicted to be HRD by a genomic test exhibited greater clinical benefit from PARPi than patients with WT BRCA1/2 and no evidence of HRD. HRD tests therefore hold promise as predictive biomarkers for PARPi and other DNA-damaging agents. However, HRD tests vary in terms of the specific genomic features they measure, and the methods used to determine thresholds defining patients with HRD. Also, HRD test results and PARPi responses can be discordant: for instance, tumors with reversion mutations that restore HR function still exhibit a “genomic scar” of HRD, and PARPi resistance mechanisms independent of HR can result in lack of PARPi response despite HRD. Emerging methods to predict HRD, including genomic and functional assays, may overcome some of these challenges. Evaluation of HRD in the clinical setting is an important tool that has potential to aid patient selection for PARPi and other DNA-damaging agents in ovarian cancer, but understanding the details of these tests and their limitations is critical to ensure their optimal clinical application.