So I guess a thought experiment is in order.... Pour yourself a...

So I guess a thought experiment is in order....
Pour yourself a nice cup of tea and drink up. The pattern of tea leaves might be pretty but is effectively meaningless. Ok let’s say you invest based on your reading of the leaves. Well that’s likely to be as successful as any random result.

Ok but what if everyone is looking at the same cup and the same pattern of meaningless leaves. Hypothetically if everyone has the same interpretation then everyone will invest the same way.... provided of course that everyone thinks it’s a good idea to bet their life savings on tea leaves. Hence a self fulfilling prophecy.

There’s a few problems though:
First not everyone will invest this way, but you don’t really need everyone. Long term investors have a low volume of trade compared to day traders. So if day traders lean towards TA then the fact that long term investors use FA will not affect the results much on the average day.
Second we’re looking at tea leaves - not the business - so it is impossible for it to predict major business change. So when major business events happen you will go down or up along with everyone else and major price movements and volatility often follow major business events.
Third and largest problem is that for it to have any predictive power at all, everyone (or at least a critical mass of people) would need to arrive at exactly the same conclusion all the time i.e. use the same analysis method and have the same interpretation of results and be acting on this information dilligently. This is all pretty unlikely.

Both technical analysis and fundamental analysis are attempts to predict the future based on results in the past. So which is more likely, that you can predict future price movements based on past price movements, or that you can predict the future health and profitability of a company based on a history of health and profitability? The later gives a (fallible) basis for assessing long term business risk, the former does not.