re: Remember how people used to say "The exception that proves the rule".

I read about this once, it may have been a blog about causality or statistics. It's not merely the observation in play, it's the context.

The rule may be that to be a CEO of a major corporation you have to have an advanced degree and spend 10 years in subordinate positions then 20 years in management, climbing the ladder along the way. This is a rule, not a law. It's a way to state the prevailing circumstances. This rule may make sense because virtually all businesses you can think of are things like Burger King or Walmart, they were established long ago, and built up gradually and require a lot of experience to integrate into the structure and become a CEO.

Then came the tech boom and people suddenly became CEO's of giant businesses without going through the process above. In aggregate, there may be a million businesses, and only 100 of them are tech giants and of those only 50 have CEO's that took unorthodox routes. This would be the exception that proves the general rule of how you get to be a CEO in a big corporation.

When you have a situation that is an exception that proves the rule, you probably need to stratify the field. Otherwise you're merely talking about low probability like the tails on a distribution.