But it is possible to tell which group you're in. Nearly everyone correctly estimates their probability of making $20m in 5 years at a startup at near zero, an

But it is possible to tell which group you're in. Nearly everyone correctly estimates their probability of making $20m in 5 years at a startup at near zero, and responds by not starting one. A few people (e.g. Max Levchin) can with equal ease correctly estimate that the probability is high enough to make it worth doing. There are only a few borderline cases of people who mistakenly believe they are the right sort of person to start a startup when they aren't. They're overrepresented here, but they're a tiny fraction of the population. Even if you're one of those people, you can do way better than falling back an average. You can for example ask experienced investors. It's part of their job to judge the probability that you'll succeed.