We develop a model of strategic contractual incompleteness that identifies conditions under which principals might omit even costlessly verifiable terms. We then use experiments to test comparative statics predictions of the model. While it is well known that verifiability imperfections can limit complete contracting, researchers know less about how the degree of imperfection affects endogenous incompleteness, particularly with repeat trading. In our baseline treatment with perfect verifiability, subjects overwhelmingly used complete contracts to conduct trades, achieving nearly first best outcomes. In our partial verifiability treatment with a reduced set of verifiable performance levels, the results reversed and parties relied heavily on incomplete contracts that omitted even costlessly verifiable terms. However, the efficacy of incomplete contracts in outperforming available complete contracts depends critically on the continuation probability of repeat trading. With a small continuation probability, incomplete contracts did no better than complete contracts while exposing parties to considerable strategic uncertainty.