Your claim that this advantage exists is based on an assumption that you've accurately evaluated the likelihood that the people who were rejected would not have been able to do the research. I don't think you have reliable data to draw such a conclusion. As such, I don't accept this claim to advantage.

However, I do agree that some plausibility metric relating the funding requested to the difficulty of the proposed research would be valuable to weed out overly ambitious projects. Moonshot programs can then rely on the lottery.