In an effort to elucidate how the challenges of nonobviousness decisions affect different decision-makers, this article presents an original dataset of nonobviousness decisions throughout the patent decision process. This dataset includes nonobviousness decisions at the Patent and Trademark Office and in the federal courts for the time period subsequent to the effective date of the Leahy-Smith America Invents Act’s (AIA) first-to-file provisions. The results provide strong evidence that nonobviousness decisions are highly indeterminate: similarly situated decision-makers in the same cases reach differing conclusions on nonobviousness at a strikingly high rate. The data does not support the hypothesis that technologically sophisticated decision-makers are better able to make judgements from the perspective of a person of ordinary skill in the art. Finally, the analysis provides some potential support for the possibility that technologically trained individuals may exhibit slightly less of a hindsight bias than untrained decision-makers. Comparing the data reported here with earlier studies of nonobviousness decisionmaking indicates that both the district courts and the Federal Circuit reacted significantly to the Supreme Court’s nonobviousness decision in KSR v. Teleflex in 2007 by establishing a higher threshold for patentees to demonstrate nonobviousness. Subsequent to the AIA of 2011, however, both judicial bodies have reverted to nonobviousness decisionmaking that is more consistent with pre-KSR outcomes.