Yeah, the problem with self-ranking is that humbleness is actively harmful. 4/5 for everything is the only safe bet -- not the best, but one notch below. You would probably have people like Howard Hinnant or Dave Abrahams come to a similar question about C++ and rank themselves as a 3/5, thinking: "There must be so many edge cases that I don't know about!". Robot HR or dumb middle managers: "Ah, they are weak. Next!"