Correct me if I‘m wrong; with technical analysis you‘re analyzing the past to recognize patterns and learn about market behavior. I think I understand what you mean saying one is not directly predicting the future. Instead, one is trying to apply learnings from the past to similar situations in the future.
But this is exactly what many ML algorithms do (e.g. recurrent neural networks and some reinforcement learning algorithms) which evaluate such probabilities. The fact that these models aren‘t profitable in the long-term (at least without insane data and compute requirements) means that technical analysis performed by humans also is not profitable in the long-term. Even for sentiment analysis to work one must have an extremely fast ML model and connection to electronic markets which only professional firms might have - a winner takes all endeavor as Renaissance Technologies has demonstrated.
There is some sound reasoning to be made around the effectiveness of fundamental analysis in the long run, however, its superior returns have also been decreasing since the early 2000s.
The most reasonable risk-adjusted stock investment strategy is arguable to hold strongly diversified assets (which reduces unsystematic risk) and invest in some large-scale trends one believes in. Technical analysis will always have negative expected returns (due to transaction costs) similar to gambling.
I really appreciate Freetrade‘s mission to get everyone investing for the long-term although more educational resources could be made available to discourage day trading and technical analysis. Happy to discuss more, as this is an interesting topics and surely debatable.