In my (one point) experience of industry R&D more projects failed than succeeded, but project leaders were not blamed.

How could one predict success or failure in risky project proposals? In both cases you learn something.