However, there are certainly things they would be fooled by, especially given a) the relative shortness of YC applications/interviews, and b) that they are all very smart and experienced people and as such are very likely to succumb to cognitive biases.
In fact, given the number of "10 tricks to get into YC" posts we see around YC application season (I wish my mock title were an exaggeration), I would be very surprised if there weren't any traits that did in fact fool PG et al and that could be used advantageously if one's only goal were to get into YC for the sake of it (it's very silly, but I've met many such people during my years in the startup scene).
We know that there are attributes that are good signals to them - they have said so. Examples include being a self taught hacker (preferably programming since high school), being close friends/having worked in the past with the other founders, focusing on things that don't scale, talking to the customers, etc. I would be terribly surprised if there has never been a founder who has filled all these checkboxes, made PG say "whoa this person reminds me of our past founder X, and they were super successful, of course we want them in the batch", only to have them fail miserably during YC.
So what did PG get tricked by in the past? That would be an interesting essay.