My point, the result of interviewing a good developer is potentially random. There is perhaps less probability of a bad developer having a random outcome as well, but certainly can happen.
Though.. define "good". Some people are incredible only when in the right team and environment.
Company says "looking for an expert on X"
Expert on X walks into the room.
Interviewer: "How good are you at Y?"
Expert on X: "I know a little about Y, but not too much. I mostly use X for my work; I could show you how to solve the problem using X."
Interviewer: "Thank you for your time; goodbye."
When I tried to figure out why this happens, the answers are usually something stupid like: "HR posted the job announcement, they have no idea what we really need", or "yeah, we are still looking for an expert on X, but we need an expert on Y more urgently". (The latter still does not explain why they rejected the expert on X without asking him anything about X, but to the person who said this it seemed to make perfect sense.)
And even then, there is a famous story of the dev of the popular MacOS FOSS package manager, Homebrew, failing his interview at Apple.
He was the creator/designer and main dev of Homebrew, the OSS <<package manager>> almost every dev on MacOS uses.
He took an idea to completion, made it production ready and used by probably millions of users.
I don't care about raw coding ability (Carmack style), if building such a large project doesn't prove at least development abilities, nothing does.
