I'm not sure I see your point. I take from your question mark that you don't believe failure is a very weak signal? You're exactly right in that if success req

I'm not sure I see your point. I take from your question mark that you don't believe failure is a very weak signal? You're exactly right in that if success requires both competency and luck, then having succeeded means you were competent, but failing doesn't mean you were not. Depending on how much luck is required, failure can be either a fairly strong signal for incompetence or pretty much negligible. This is very easily proven from the bayesian theorem.