The purpose of TA is to predict price movements based on objective metrics such as price (or price movements) trading volumes (and types of trades) or other metrics such as price to earnings ratio or dividend yield.
TA doesn't always have to be right but to be valid it should deliver a positive return (in comparison with relevant market indices) in the long term.
A very simple model might note that prices typically rise after n days of falling and so the investor using that model would keep track of a bunch of stocks and buy those with n days of falls.
Of course if it was that easy, everyone would be doing it so TA models get progressively more complex involving the intersection of various movements (or second or third derivatives thereof) until there's just a mass of lines and someone declaring that there's a clear sign to buy or sell.
Of course TA only works if only a few people are doing it successfully. If too much money ploughs into the market, the model is comprised.
IMO for retail investors it's a myth. Any model simple enough for an individual to construct (and run) will have already been considered, maybe used, and discarded my financial institutions who can afford the biggest brains and the most powerful computers.
There maybe someone somewhere doing TA successfully but they're probably very tight-lipped about it for fear of losing their edge.
Even worse, it could be used as a counter strategy if enough retail investors follow that model.
It occurs to me that the most plausible way to be successful at TA could be just being a step ahead of TA crowdthink by somehow analyzing that rather than spotting general market trends.
A solid gambling principle is that the easiest people to outsmart are the ones thinking they are doing the outsmarting.
We might perhaps be conflating two or three different things here.
The edge offered by any valid strategy will answer to the above, in terms of getting blunted through use. That goes for the most obvious of "fundamental" triggers as well. That's the whole "Random Walk" argument, which basically says that at least at the retail level people are generally best served by investing in index funds, or buying for dividend payouts --- that is, by not trying to outguess the market. That applies ---- the basic logic of it, at any rate, and without going at this time into whether and how far we accept that logic ---- to every valid strategy (that the opportunity any strategy that is followed by the herd will end up drying up through use, and indeed dry up more or less instantly in well-oiled efficient markets).
But does the underlying strategy itself have any worth to it? Entirely apart from whether or not an efficient market offers the retail investor any opportunity to profit off of it?
And even at that level, we might look at two aspects of it, the bottom-line evidence of it, and the reasons driving it. For instance, without a shadow of doubt fundamental analysis passes the second test (if your analysis is complex enough to include not just the underlying business at the organizational level, but also the broader market, as well as the regulatory framework, as well as financial/market inflows and outflows and triggers, all of that). Without a shadow of doubt that works, or should work; but whether it actually does is something that is debatable, and the only way to resolve the debate is through actual evidence.
So does TA work, bottom line, when it comes to stock markets, or currency, or, to tie it to this thread, cryto? That's easily proved, should proof be available, by directly presenting evidence. I don't think it does, because I haven't seen any evidence myself (although I'm willing to be persuaded otherwise).
As far as the in-principle thing, at which level fundamental analysis does work, you guys have both presented interesting arguments in favor or TA,
The Don and
Suddenly.
The Don, you've suggested that traders follow certain profit-taking or whatever strategies regardless of the underlying, so that at least for pure-trading trading transactions --- as I read that --- patterns in past prices might indeed coherently amount to some indication about future trading activity. I agree, that does make sense.
But to derive any actual signal off of that logic, we'd need to isolate the trading transactions from the overall transactions (that include transactions driven by fundamental considerations, and also fiscal- or money-inflows-outflows considerations, because those would obviously not be based off of trader-profit-taking-et-cetera considerations). Can we effectively do that, and might that result in a workable strategy? And finally is there any actual evidence for any of this? Those are interesting questions, absolutely! (And at the end of it all is the question of whether an efficient market leaves place to profit from any of it.)
Suddenly, your comment about sporting bets also raises another interesting way how TA might work. You've suggested that psychology of how punters bet might offer some insight into how past prices/odds might track on to future prices/odds/outcomes. (Frankly, and like I'd said, the details of that I didn't actually understand. I've no idea, at all, about how sports betting works.) But if there's actual evidence that what you're suggesting holds, that is if there is evidence that TA works in sports betting, well then that might offer some basis on which we can assess whether TA might work with stocks, or currencies, or crypto. Whether it actually does or not is a separate question, but whether it
might in principle, basis your argument, that hinges on whether there is evidence that TA works in sprots betting. Is there?
---
Like I said, I myself consider TA as no different than reading tea leaves, and the lines on people's palms, and astrology --- people running off into unwarranted conclusions basis random patterns. But both of you suggest arguments that, if they work, might indicate that TA does work, or at the very least that it
might work in principle. It might be interesting to see if those arguments do hold up, if you guys would like to take these further.
