[gumby]

> When actually hiring, you also have benchmarks that aren't purely relative, and past experience to fall back on.

(This problem is framed as hiring the best which I assume means closest to optimal weights on those non-relative benchmarks -- whatever "optimal" means.)

But your comment is interesting, as this problem was originally as "the wife problem": how do "you" -- presuming incorrectly even at the time all mathematicians to be male -- find the optimal wife.

The thing that made your comment funny to me is that in majority of marriage cases (by anyone to anyone) neither party has much, or any past experience to fall back on!

[Retra]

In those cases, even if you select poorly, you can change your weight and preferences afterward.