Experimentation is _ALWAYS_ at the "oh look, something funny happens if i do this" stage, otherwise it isn't experimentation/research but rather research/development. If we know what to expect, why experiment? Why waste resources?