terminal value of being better able to distinguish between true and false; to "be less wrong"

I don't think that's quite a terminal value.

Regardless, modern science aims to minimize false positives. Strong ability in that is hard to compare to weaker abilities to adeptly manage the ratio of false positive to false negative errors, calculating expected value of information, account for cognitive biases, etc.

Two people, each having one of the ability sets I described above, would perform differently in different scenarios, with variables being environment and values. It may be possible to objectively compute values for each describable skill set. Assuming one can and there are hard limits, as no person will approach either hard limit, and even if they could approach one hard limit the best thing to do would probably be to reach a Pareto optimum far from the limit it could nearly reach, I say this is vague and sounds true but not important.