I should have qualified this. Only very small changes are predicted for small numbers of generations. I am unaware of any experiment with fruit flies that has involved say several thousand generations, for example. For small numbers of generations, we would expect only small changes.
Well this is where the "believability" factor comes into play. Unless evos can really predict "what some of these changes might be" and "in what time frame or conditions they might occur" then I think the theory is quite literally impotent.
Just saying that "things change whenever and however they change" is not very insightful to this poster.