Did we say it was or wasnt legitimate to amalgamate results from several runs at an ABX test? I would have thought as A and B are the same files in all test runs it was , but does that still hold if, for example , the equipment used changes between runs?