[riggins]

The thing these aggregate statistic ignore are the quality of the founders.

Let me put it concretely.

When Sebastian Thrun decided to launch Udacity I suspect the chance of failure was less then 90%. OTOH, I have 2 non-technical friends who managed to raise money to develop 2 more social networking app. I think the chance of failure there is easily 90%.

So maybe another way to put it is 'don't fool yourself'.

If you have Thrun-esque talent, start a company. If you don't, think twice, then think a third time for good measure before starting a company.