There is no reason why it should because you are right it's just squiggly lines. In a casino it's no more effective than watching the patterns in baccarat and trying to figure out the pattern on the sheet.
From a purely logical point of view. The reason why it works is because enough people believe it works. Unlike the casino example (Where it's just random), if enough traders and algorithms actually trade on these technical signs, then they have the ability to move the stock. Which is probably why technicals work during times of noise (No earnings or news, moves with macro or starts doing it's own thing) and break down when actual news comes out.
That is almost completely correct in my humble opinion.
I will only add that even when TA uses the best stochastic models, by definition the models work only when past patterns are exactly the identical ones to future patterns. (That is the most basic ‘boundary condition’). In ideal conditions those can be ~60% correct. That is better than guessing, if one has no better information.
So, how does a greedy quant or the quants employers make money with such vaguely accurate tools? As you say, when they actually trade on the tools they can make inordinate profits, draw in more
marks investors, and publicize the wonderful gains. The glory of this system is that it is absolutely legal. The even better part is that highly skilled analysts, statisticians and investors are so convinced of their perspicuity that they flatly ignore the fundamental principles.
As many of us have pointed out TSLA is an ideal security on which to base such efforts. It is widely held, has very active retail investors who are predominately well-educated, statistically fluent or nearly so, trust data and use statistical modeling on a regular basis. Above all, nearly all of them are fairly new to securities, having not directly experienced even 2008-2009 markets. Further many of them, as always, report successes, not failures.
That is where we are. We repeat versions of this over and over, occasionally being ridiculed by the speculators because they have superior knowledge. The prototypical highly touted technical tool is Bollinger Bands, developed by a PR maven educated at a for profit art school. His clever promotion and convincing demeanor have given him a wide following and academic honors. The problem is that they really have minimal value. Just as with astrology though, it’s useless to try to convince adherents, so most if us don’t try. This one is easy compared with, say, pointing out that the Black-Scholes model is irrelevant to us options valuation, which is anathema to the many corporations that use it for executive options valuation.
In sum ‘in the land of the blind, the one-eyed is king’ . That is why all this continues decade after decade while the false prophets gain awards, professorships and prizes. When reality intervenes they invariably are well prepared to blame other people.
Sorry for being strident. I’m always sad when people close to me walk into these disasters with unseeing eyes wide open. Yesterday I was asked to bail out a relative who lost all his wealth in these idiocies. He had been regularly telling me I did not understand, things were different now. Armed with an excellent resume, led by luminaries in ‘investment’, he utterly failed. From Tesla Economist, who admitted it, to many others I am very sorry we all could not adequately convince people to be prudent.