[nkrebs13]

It can be and some people treat it like that. But it can also be a tool used to identify trends and trend changes. The chart takes into account everything: financials, narrative, public opinion, future earning potential, etc. All of these inputs are probabilistically weighed against each other in real time. Keep the TA simple and it's an extremely useful tool. Anyone who dismisses TA entirely is leaving money on the table.

[nradov]

There is no reliable evidence that TA generates alpha. In reality you're just fooling yourself with confirmation bias. If TA actually worked then sophisticated institutional traders would quickly arbitrage any advantage away before retail traders could even put in an order.

[ethbr0]

"Alpha" is an odd benchmark for efficacy.

A lot of technical analysis's value, in the sense that I understand it and believe most retail traders use it, is based around predicting intermediate timespan market changes to time trading. It's less about beating the market and more about taking its pulse.

Looking at TA statistics generated from daily, weekly, and monthly points, effectively aren't those momentum? Which collapses the question down to "Do stocks exhibit momentum or not?"

[nradov]

I don't understand your point. If TA can't generate alpha then why bother? Whether stocks exhibit momentum or not is irrelevant if you can't profitably trade off of it.

Look at it this way. TA is fundamentally a matter of pattern recognition. If it actually worked then you could program a computer to recognize the patterns and automatically put in trades. But if everyone does that then any profit opportunity is almost instantly arbitaged away, and thus TA quickly stops working (if it ever worked at all).

[criddell]

Is stock momentum the same thing as a hot or cold streak in professional sports?

[JimtheCoder]

"There is no reliable evidence that TA generates alpha."

To be fair, there is no reliable evidence that anything really generates alpha (efficient markets, random walk, etc), but there are hedge funds who consistently outperform (Renaissance, Two Sigma, DE Shaw, etc).

So, maybe there are things that work in practice, but there is no way to produce reliable evidence that they work besides looking at track records...

[nradov]

Public markets are fairly efficient, although not perfectly so. The hedge funds that consistently outperform aren't using anything so simplistic as TA on a single ticker. Instead they employ an army of experts armed with enormous computing resources to identify correlations with data outside those markets that no one else has noticed (arbitrage opportunities). They also do proprietary research by commissioning researchers to go out and gather real world data that no one else has, so they are working with an information asymmetry.

Of course, there are also persistent rumors of insider trading. Maybe just sour grapes from inferior traders, but who knows?

[JimtheCoder]

Yes, I agree. But, none of this means that TA isn't useful, or that it is comparable to astrology...