We examine innovative contexts like scientific research or technical R&D where agents must search across many potential projects of varying and uncertain returns. Is it better to possess incomplete but accurate data on the value of some projects, or might there be cases where it is better to...