Speaking of methodology, it's incorrect to relate a development test result to reliability or risk. Source is my personal experience doing reliability calculations for a NASA rocket component and working with the statisticians incorporate my numbers into their risk model.