Patent litigation is widely regarded as one of the most complex types of civil litigation, with costs often totaling millions of dollars and typical cases lasting years. Also, the burdens of patent case complexity land on both sides of the technological divide, as large producers face skyrocketing defense budgets and inventors and startups risk being “priced out” from enforcing their rights. Yet, the complexity of patent cases is poorly understood as an empirical matter. Instead, patent litigation is generally accepted to be a “Pandora’s Box” of incalculable complexity, which, once opened, is only arduously and unpredictably concluded.This study undertakes a comprehensive exploration of patent litigation complexity, first defining robust metrics of complexity and continuing with rigorous analysis of the determinants thereof. We focus our study on the eight years of U.S. District Court litigation leading up to passage of the America Invents Act, and we mine extensive detail of more than 1000 cases during this timeframe. Using these data we ask targeted questions about patent case complexity, including what types of cases are most complex, how defense costs compare to enforcement costs, what factors are associated with particularly high complexity, and how complexity has changed over time. Finally, we conduct a large-scale event study to identify the causal impact of key policy changes on case complexity, specifically the landmark shifts in remedies law over a series of recent Federal Circuit decisions. The analysis herein is of crucial importance to patent policy. As juridical property, patent rights are ultimately enforced and defended against in legal proceedings, and thus the complexity of such proceedings directly impacts the rights afforded by patents and recourse thereunder. Understanding case complexity is therefore a necessary contribution to patent policy discourse. Moreover, the framework developed herein sets the stage for future analysis of the complexity impact of new policy measures.