I do think that's possible, and that is one of the confounding factors we spend a lot of time worrying about and trying to control for when we do the data scien

I do think that's possible, and that is one of the confounding factors we spend a lot of time worrying about and trying to control for when we do the data science on this. One of the ways we do this is by licensing data about all the startups in the world and then normalizing our own data against the market overall. Getting a true apples-to-apples comparison is difficult because the world doesn't stay the same. The funding market has changed dramatically since YC started, and the companies we're funding have changed a lot too - they're much more diverse now. We've tried to control for this but you are right that when building a complicated model, you can never be 100% certain and need to stay humble about your ability to correctly control for all these factors. We won't know for sure for many years; for now all I can say is that we've looked at this from many angles and the early indicators look good.