Not really surprising that "evidence confirms futility" in this case. Simple moving average crossovers worked nicely 30 years ago. The world has moved on since then, but beginners still have to start learning somewhere. MA crossovers are to experienced traders what a "hello world" program is to experienced programers. Just like the well-known Turtles breakout system worked nicely years ago but lost its edge and then stopped working when too many other traders started to fade ("Turtle soup") its signals, something similar happened with MA crossovers. Beyond the beginning traders who use MA crossovers as a TF entry, more people now take the opposite side as a short-term mean-reversion trade. Profitable trading nowadays requires something a bit smarter than either. Looks like the academics who wrote this paper are a bit behind the times ....
Cheers, best wishes, Tony M (KL, Malaysia)