> If we alter selection for Navy SEALs

But also, I think you're ignoring some things. First, it is much easier to measure physical strength than it is to measure intelligence or someone's ability to succeed in college (two different things. Related, but not the same). There is also much more mental flexibility than there is physical, especially at the ages we're talking about.

Second, is that every one of those aspects is trained. They teach you to be athletic. They teach you tools to help motivate yourself (and provide external motivation). They teach you how to be team oriented. Hell, they teach you how to take power naps. Your genetics aren't the only thing at play here and the genetics defines your limit to things like physical strength, not you're ability to have physical strength. The whole point of military training is to tear you down and build you up to be somebody else. Sure, some people come in with a leg up and have genetic advantages but these aren't necessarily the most important factors.

My point is we're talking about... well... 18 year olds. Just like Navy recruits. The gap between a good high school student and an average high school student isn't that large. Even if it seems so at that time. But with university what matters is where somebody ends up, not where they start. Are those gaps able to be crossed? Usually, yes. That is why measuring high school performance is messy as well as why final college grades are weak measurements too. Everyone knows the whole problem of certain professors being better but giving lower grades, but there's also the fact that you aggregate over 4 years. Messing up Freshman and Sophomore year can fuck up your GPA and this metric won't tell anyone if you became the star student by graduation. I don't think we should throw out GPA but it helps to recognize it is noisy.

  > To be even more explicit, a lot of the reason that these colleges are so strongly correlated with high incomes and high achievement is that they allow attendees to network with people whose parents and family members have already reached the upper echelon.

I think the bigger problem is that conversations around university tend to be about meritocracy. I don't think admissions measure that. Nor do I think GPA does. They can correlate, but there's just a lot of noise. I just wish we would stop pretending. It would make conversations change a lot. Frankly, I think at the end of the day, even after a full undergrad education, the biggest difference from a kid from Standford and a kid from some Cal State is that one of these kids will talk about how they went to Standford. At best, it is a weak signal of skill. But it is a strong signal of one's network. But if we're concerned with merit, let's not confuse the two[1]. As long as we do I don't think we can even know what problem we're trying to solve, let alone solve it.

[0] Could be wrong. But high confidence not more than 2 top universities can meet that. Number of open positions is just very small. Even if I were 10x off, the effect could still occur without much time difference.

[1] Especially true considering how common it is for people to say that what you learned in college doesn't apply to the job. If that's true, then it only makes that merit correlation weaker, though it doesn't change the network one...