> Elliott Waves do not meet the scientific criteria to be considered a real theory.
It can't be considered a real scientific theory. It doesn't try to either (although, in all fairness, it was indeed the initial intention of Elliott, but he was not a scientist). Saying that it's not a theory is quite weird. Do you mean it's fact, then?
> They have no predictive power
Yes, it's not enough by itself, it's not a formula that you can apply and get rich. This is something people expect from EWT, and this expectation is the error. Like any technical analysis tool, people can loosely "use" it to see what they want to see. Their fault.
> It's just more after-the-fact rationalizations and attempts by humans to find meaningful patterns in chaos
While it's true that it's way easier to count correctly waves after they happened, it still provides a framework that allows for more successful trading with it than without it if people don't stop at the first interpretation they think they see and instead build strategies based all the interpretations they can find.
I'm always surprised how people easily reject EWT despite all its success, often based on a single failure or on the fact that it wasn't successful for them when they tried it without applying due rigor. My bet here is that people are scared it might be a clue of some yet to be discovered natural law, because it would have huge implications regarding determinism, at least for big enough social groups. People want to be free, badly.
EDIT : btw, saying that trying to see patterns in chaos is unscientific is quite funny : https://en.wikipedia.org/wiki/Chaos_theory