You can't prove a negative like that. It'd be like academics looking for monetizable edge in the orderbook after properly simulating latency. They wouldn't be a

You can't prove a negative like that. It'd be like academics looking for monetizable edge in the orderbook after properly simulating latency. They wouldn't be able to do it due to a lack of domain knowledge but they didn't prove a negative through their failure. As for the effectiveness of TA, there's a tonne of dogma on both sides with extremely certain people saying it does or does not work. If we take a broad definition of TA (which is edge existing in operations on a time series of prices), I have conclusive evidence that it does work. I have seen a strategy print money almost daily using only that data as an input. If we take a narrow definition of TA, defined as lines on a chart, well I'm a believer of that too, although the evidence is not as strong. I just suspect that it works due to my observation of how things react in the market according to those levels and lines. Here's one you can look out for yourself. Observe what happens when the price moves through yesterday's close price. You will notice that volatility becomes significantly elevated. That's edge, and it's TA edge. I don't believe it works merely due to the self fulfilling prophecy aspect. It works because other market participants put stop orders, resting limit orders, or algo trading rules (in banks' liquidation or acquisition algos) tied to those levels.