What exactly is TA, and has merit in TA and what does not?

[David]

TA usually follows very specific rules; whether the price went through a restistance/support levels or if a particular pattern was recognized, the rules are very simple.The trader can program the whole process of identifying pattrens, restistance, and whatever personal assumptions into algorithms and computer codes and executes the trades automatically. However the risk associated with emotions can be overwhelming, particularly when other distractive simultaneous factors enter into the trade (high volume, unexpected events, false gaps, etc.). One possible way to somehow smooth these factors is sitting in a large trading room near arbitrageurs, value traders and other traders in order to collect sensive and live information on the market which may be very valuable. Similarly, if one isolates himself from the sociability of the markets and executes his trades from his grandmother’s basement, he may well end up working on a night shift at the local pizzeria in... Wink

[Scotty]

Random shocks to the market will always introduce mis-pricing.  To make money from these mis-pricings in a way that is statistically significant, you need to do two things:



a) Introduce some level of predictability into the market, which is then exploited immediately through hedging or statistically, through the prices moving more often than not to your prediction.



b) Use tools and techniques that are scarce or unique.  If everyone else is doing the same thing  your profitability will quickly go to zero.



The drawbacks to TA are:



Beecause TA seems to offer no explanatory power, you don't know much about the predictions.  For example, perhaps the reason a particular pattern appears to work is that by taking advantage of it, you are exposing yourself to a rare but disasterous event.  Who knows?



More significantly, as the bookshelves are filled with TA books and everyone inside and outside the financial markets has been led to believe that with a bit of focus they can pick stocks, the barriers to entry for TA are very low.  This means there is a lot of competition.  This means that any potential returns are likely to be short lived.

[FDAXHunter]

I think we need to distinguish between systematic Technical Analysis (i.e. stuff that can be turned into a precise algorithm (rather than a fuzzy one)) and discretionary Technical Analysis, which depends to a large degree on the ability/talent of the analyst.



Systematically, it's pretty easy to justify... find whatever works, if the Risk/Reward is small enough... trade 'em up. Using something something different from everybody else is helpful, but not necessary.



The descretionary aspect is, IMHO, different. As I said in my last post, the key is establishing a framework from which you can work from psychologically. If you are buying 114.00, how do you know what's a good exit point, both on the loss side and the profit side? There are traders who are very intuitive on this and just have the "feel" and they don't need to rely on a transformation of the market information (or rather: they map the transformation in their mind.) 

For others not so talented.... you need something to provide you as a backdrop, because otherwise you're flying at night with no radar.



Bear in mind that most technical traders (and heaven forbid... analysts) suck. But this is no different than in any other profession, where a few are spectacular and 80% just sad.



Hm... more ranting... sorry.

[Scotty]

Not rantings - it is absolutely worth thinking about given the impact it has on the world.



You're right - it is importand to distringuish the two - particularly for external parties.  With systematic traders you can at least back-test and develop some level of confidence around future results.  With intuitive traders, it is difficult to differentiate the black swans from the skilled (if that makes sense...it does to me!).



There does seem to be a greater level of perverse information in finance than other industries - belief that historical results are good estimators for future results, the general hype around the skill of equity anaysts, a general belief by everyone that they can pick stocks, lack of diversification, the fact that so many books on TA fill the bookshelves (ie strategies failed, but perhaps selling books is profitable), etc.  OK, now I'm ranting...

[Johnny]

"The level of perverse information" is caused by the fact that finance is a [i]social science[/i] involving complicated feedback systems between people. So when Person A trades his back-tested trend-trading system, Person B will unleash her "Predatory Manipulation" system designed to sucker Person A into thinking that every move is a change of trend direction. Any statistically significant relationships one person can find, another can find and manipulate to their advantage. The trader that fails to realise this becomes the profit opportunity for the rest of the market.



Smiley

[Energetic]

The problem with the discretionary/intuitive "systems" is that the approach is not falsifiable. If there's no precise algorithm how can we discuss whether it works or not? All we know for a fact is that there're some people who seem to use various flavors of TA as part of their analysis. But nobody with positive P&L will explain exactly what he does.