I'd be interested to know stats about the impact of Y Combinator on companies with a nice control setup. Revealed preference: people who get in, but choose not to go to Y Combinator, vs. those who do... or something like that.