YC can't unambiguously tell whether or not a particular rejected candidate should have been in the top 50 or not, at least not for a significant number of those rejects.
My point remains: quality or success cannot be foretold in advance. Particularly when much of that success depends on passing prior selection phases.
In evolutionary biology, as a comparative example, there's no question that luck plays a role in evolution. Your ancestors had the bad luck to live in an asteroid strike zone, on a tidal-wave flood plain, or within the affected region of an ancient pandemic? Pity, you're not here now. So the influence on survival exists.
The relevant part to evolution though is that luck is not an inheritable trait. It cannot be selected for. It's part of the random noise.
Similar logic holds true for YC prospects: some are successful, some are not. Passing the bar itself comprises in significant part a random "lucky" event. Yes, that can be influenced, but you cannot eliminate the error.
I never said that they could assess it "perfectly and without error". I said that it means that "YC doesn't think that you're one of the 50 or so best startups".