Besides for the extra-ordinary performers I think at some point flipping a coin might be better for this sort of problem, and there's some precedent. It actually be more fair than trying to find differences in the sub-top, actually more often than not introducing bias, when there's a lot of competition. Just put the top 3, top 10% or top 20% of those who fulfil a set of requirements in a raffle.
https://en.wikipedia.org/wiki/Funding_by_lottery