It's not statistically unsound to focus on extreme outliers. In fact, if your data set contains lots of outliers that your statistics can't explain, it's more of the case that your statistics are wrong (I'm not saying the math is wrong, I'm saying you're applying the wrong model to your dataset). This is basically the whole lesson behind books like Black Swan.

I agree with the comment you are replying to. You're commenting on a website of the most successful tech incubator of all time where, at least from a quick glance, all of the top companies were founded by people in their 20s. Given the researches intro sentence of "Many observers, and many investors, believe that young people are especially likely to produce the most successful new firms" I think their conclusion is really just using an invalid viewpoint of what "the most successful new firms" means.