We analyze the incentives to disclose intermediate research results during the course of a patent contest. Despite knowledge spillovers, the leading innovator sometimes discloses to signal commitment to the project, and so potentially inducing a rival’s exit. Surprisingly, when development costs are low the leading innovator does not need to disclose to induce the same strategic deterrence effect as that which arises from disclosure. Taking into account wasteful duplication of R&D effort, a patent office can increase welfare by choosing the probability of granting a contested patent and so altering the proportion of rivals that the leading innovator deters.