Maybe this is evidence that natural selection doesn't work.......at least not in fly populations.
......and that's the problem with so much of Darwinian thinking: Natural selection is believed with certainty to explain all things, so contrarian thinking is not even imaginable to the majority of our evolutionary biologists.
That's a fair question, but it's a stretch to think that an answer would show natural selection doesn't work.
I'll point out that this theory that the stripes deter flies wasn't actually tested. They showed a correlation between populations and inferred that the flies didn't like stripes. That should then be tested. The prediction is these flies would avoid stripes. So you get some flies and set up an experiment to find out. This could be wrong, despite the correlation.