Evidence shows how difficult it can be to rank research projects that, after going through peer review, fall in the fundable category. There is an element of chance in the decision-making process anyway, so once you triage out those who are not fundable and those who should definitely be funded, why not make random allocation a formal part of the decision-making process?