Experimentation conjures images of laboratories and equipment in biotechnology, chemistry, materials science, and pharmaceuticals. Yet modern day experimentation is not limited to only chemical synthesis, but is increasingly computational. Researchers in the unpredictable arts can experiment upon the functions, properties, reactions, and structures of chemical compounds with highly accurate computational techniques. These computational capabilities challenge the enablement and utility patentability requirements. The patent statute requires that the inventor explain how to make and use the invention without undue experimentation and that the invention have at least substantial and specific utility. These patentability requirements do not align with computational research capabilities, which allow inventors to file earlier patent applications, develop prophetic examples, and provide supporting disclosure in the patent specification without necessarily conducting traditional, laboratory-based experiments. This Article explores the contours and applications of computational capabilities on patentability, proposes reforms to the utility doctrine and to patent examination, responds to potential critiques of the proposed reforms, and analyzes innovation policy in the unpredictable arts. In light of increasing computational experimentation, this Article recommends strengthening the utility requirement in order to prevent a state of patent law in which enablement is subsumed into utility.