I don’t know why you think it needs super fast machine learning algorithms. You only need that if you are doing automated high frequency trading, That’s not what I’m talking about. I’m talking about looking at similar situations in the past and trying to ascertain probabilities. You can’t predict the future with it, but you seem to think it’s completely worthless.
I’m not a day trader, but I made the decision to start buying stocks again in late march after the covid crash based on technical analysis (not done by me) that suggested we may have already seen the bottom. Turned out it was correct while many were still on the sidelines predicting further falls