There is certainly no guarantee that with a "solid" team you'll get in. At the end of the day, it's all a matter of serendipity.
Typewire - disqus for liveblogging
Attend.me - conference management software
Prospector - medical chart review
and finally LightTable. Oddly enough we had/have domain expertise for all of them :D
I think it's way more than serendipity. I think the YC staff is extremely good at identifying where good teams and good products meet.
While I agree that they've gotten very good at identifying potential good team/product pairings, that just means they have relatively few false positives. It makes no claim to how many people they pass over who may in fact have an amazing team and product. As the noise in the process increases, you're bound to end up with more and more people who for completely random reasons didn't end up being what they wanted that day. This isn't a purely objective process (thank goodness).
All other things being equal, anything that relies on the judgement of a few people has an element of serendipity to it.
"All other things being equal, anything that relies on the judgement of a few people has an element of serendipity to it."
Also recall that the judgement is made on the basis of a 10minute interview. Ask any of these startups if they'd ever hire a group of strangers into their company based on a 10 minute interview. (I'm aware that's not really a fair comparison, but it should still make people think).