Comment by PeterStuer

The problem with your reasoning is you never tackle your "unknown unknowns". You just assume they are "known unknowns".

Diving through the abstraction reveals some of those.