RJC
Post Mortem scientific analysis is one way, but not the only way. Scientific proactive analysis is also a way. I hate to use the term "risk analysis" because these are usually done with a view to an outcome, but done properly they can be good. Its just that at the moment, they are rarely done properly.