What I'm getting from most of what you're saying here, is that technical analysis is by and large an instance of humans seeing patterns where there are none, that for example when the price movement on a chart paints a flag or a pennant, that there is no possible way to to tell in advance it would paint that picture, but that such patterns are merely assigned to it in hindsight. Is this on par with your view, or am I misrepresenting it?
Yes what I am saying is that what patterns people think they see via TA is false.
There certainly are patterns in the market, otherwise it would be random. I think the efficient market hypothesis is not entirely correct. I think the market tends towards efficiency but it's not efficient at every single datapoint, so that can be exploited.
But the patterns are very subtle, probably impossible to see with naked eye. It would be something like a relationship based on a complicated formula, not like a candlestick or any other nonsense like that.
So visual TA is nonsense as it is, there is no other way to put it. But that doesn't mean that with quantitative analysis you can't make some progress in decyphering the market.