What you are getting at is that for a single person, the chance of going bankrupt is high but for a population ensemble, the average wealth increases. This is

What you are getting at is that for a single person, the chance of going bankrupt is high but for a population ensemble, the average wealth increases. This is absolutely true. All it takes is to understand that for the single person, the model isn't ergodic but all expected value based models assume ergodicity. See [section 4 here](https://www.jasoncollins.blog/posts/ergodicity-economics-a-primer#losing-wealth-on-a-positive-value-bet https://www.jasoncollins.blog/posts/ergodicity-economics-a-p...) on losing wealth on a positive value bet.