jes, there's no question but that the *best* guy among 4-6 will be much better, on average, than 7, once you know who's best among 4-6. Absolutely, and that supports the argument that when you're picking at 7, you're selection pool will often be missing a good player. But if you took the *best* of picks 8-10, that also works significantly better than 7, even though the best of the 8-10 guys were available at 7.
Yes, maybe at 7 you'll miss out on a guy you wanted, and who'll turn out great. And certainly your chances to get somebody really good will increase if you have 3 top-ten picks rather than just one, whether those 3 are 4-5-6 or are 8-9-10. But I'm not sure there's strong evidence that success ratio picking one guy, at 4, without the benefit of retrospect, is lots higher than picking one guy at 5, or one guy at 6, or one guy at 7.
If we had perfect prescient scouts who'd know exactly which guy on the board will work out best, it would be very helpful. So would having three top-ten picks. Unfortunately they'll only get one pick, and they won't know how the guy will do.