The most common pattern I see is doing a startup, failing, but learning enough about how to do it right, and then trying again and having moderate success with #2 or #3.

9/10 of startups failing could be explained by "virtually all first-time startups fail, and 9/10 founders give up after their first failure, and those that stick with it, succeed". That would actually be rather encouraging.

Almost nobody I know working for a big company is working for their first big company, so you could also say "9/10 of all employments fail", but that's hardly an indictment of employment.