I agree that lines of code is probably the overall best predictor to first order (didn't mean to imply otherwise; I've edited my original post to clarify). I ju

I agree that lines of code is probably the overall best predictor to first order (didn't mean to imply otherwise; I've edited my original post to clarify). I just meant that x lines of overengineered code will almost always be buggier than x lines of non-overengineered code.