I agree with this, and I don't think that actually needs much evidence. It follows straight from assuming that:
- YC has prestige and is recognized as a top accelerator (which is true)
- YC is competent at picking strong applicants (which also appears to be true)
- there is no magical cutoff in the applicant pool where applicants below it are significantly worse than above it, and the cutoff point happens to be exactly at the point where YC stops accepting people (proposing the opposite would require serious evidence)
You have to remember that investing in companies is a lottery game. Most angels/vc's profits are made on that 1 in a million company with a 1000x return.
If YC wants access to truly the best companies, they need to institute a rolling admissions. Otherwise, these scrappy second tier incubators will get lucky a few times when a nervous founder takes a bird in the hand over two in the bush. The policy only hurts YC itself.
IMO, there is still too much luck involved to make this true to the extent it matters in this case. YC has picked several things that didn't pan out, and of course several that did. There have also been multiple follow ups from YC rejects that went on to be much more successful than some companies that WERE picked
If this were true, YC would be the size of Exxon. Picking A+++ companies is a heuristic.
Sounds like a terrible expected ROI ;)
That is thoughtfully provocative. I have literally zero evidence to go from, but I suspect that Y Combinator is like Harvard or Yale: they could get rid of the group that they actually admit and go for the next tier, then achieve similar outcomes. (In other words: there are so many strong applicants that they can't take them all.)
This is just to say that even taking YC's cast-offs could be perfectly acceptable. YC is dominant for now, but it's not hard to imagine a world in which there are a few accelerators which are true YC peers (just like Harvard, Yale, Princeton, and Stanford are variously preferred by different people with real differences in preference).