This presumes (a) a highly reliable ability to identify the "best person", which we simply don't have, (b) the idea that "best" is unidimensional and unrelated

This presumes (a) a highly reliable ability to identify the "best person", which we simply don't have, (b) the idea that "best" is unidimensional and unrelated to the context one finds oneself in and (c) the idea that such a "best" person will also be the one who will do most to advance science. Just take a look at a century of Nobel prize winners and you'll see that none of those assumptions hold.