It is you who does not understand statistics. If the 'top people' is a population with a higher than average IQ, then that is probably a useful thing to measur

It is you who does not understand statistics. If the 'top people' is a population with a higher than average IQ, then that is probably a useful thing to measure when trying to predict if someone is going to be a top person, even if not all high IQ people are top people and not all top people have high IQ. See Bayes. Also, if you were coming to this situation with no prior knowledge that IQ means intelligence (that is, imagine that you didn't know what the term IQ means) then you could suspect that its correlation with 'top people' might not be causative. The trouble is, you live and work in the real world, and so have seen smart and not-so-smart people at work. So you have strong reason to suspect that the arrow of causality points from IQ, not towards it. Intelligence is not the only thing, but it's important and it can be measured.