I don't think that's true. I think even on-the-napkin stage ideas also can use a lot of YC input to drive development in right direction. I believe it is important to at least have a prototype at least a few weeks before demo day in these programs, but that depends, of course.
I think you're looking at companies that would fail without sucking in every last bit of YC juice. Those kind of companies will fail anyway.
Given the same company before and after launch, it's hard to see why waiting until you have traction before doing YC would decrease your chances of being successful. At best you get a lot of interest on demo day and save a month you would have otherwise spent raising money, and at worst you're not really any worse off than before. (This is assuming you can figure out how to make a good product on your own, and that not doing YC earlier won't slow you down at all. Not true for all startups.)
There's no question that earlier companies still get a ton of value out of YC and that it's still worth doing even pre-traction, but that wasn't really my question.
I think you're looking at companies that would fail without sucking in every last bit of YC juice. Those kind of companies will fail anyway.