If "Finding a problem worth solving that hasn't been done" is easy, it wouldn't take three tries for a great team, and failure wouldn't be possible for any companies with great teams.
The experience of countless startups with great teams but failed businesses disagrees with that philosophy in my opinion.
Even in your own example, the team wasn't the problem it was the idea & the lack of market knowledge. You made a product before you knew the market. If you had of done more market research before building, would you have tried the same idea?
I think you misunderstand my argument. I definitely agree with you about the importance of understanding your market. I'm just saying that YC does not choose teams this way. This probably has a lot to do with the fact that its hard to figure out if someone has validated knowledge about anything in a 5 minute interview, especially when you're trying to discuss a market that you likely do not understand personally.
YC's perspective seems to be that "launch early and often" combined with smart analytical teams will lead to successful companies. This seems reasonable enough, but does not always work.
I'm learning that launch early isn't enough. I'd even go so far as to say I think launch early is the wrong approach. You should try to ruthlessly prove that your product/customer hypothesis is wrong before even touching code. I have been amazed lately at how much time & pain you can save by "talking early" instead of "launching early".
As a technologist who has wasted years of time and hundreds of thousands of my own $ on launching early (multiple times) rather than talking early I am speaking from experience here as well :)
The experience of countless startups with great teams but failed businesses disagrees with that philosophy in my opinion.
Even in your own example, the team wasn't the problem it was the idea & the lack of market knowledge. You made a product before you knew the market. If you had of done more market research before building, would you have tried the same idea?