I was only referring to the 'can be automated' part. If it could be automated to where it's worth a dime, you would think someone would have done it by now, as the rules for drawing a trend line are quite simple. Doesn't look as though the computer can do consistently what it's told to do in the attached examples.
I agree, there were fairly decent methods for trend line drawing 5 years ago when I last looked into this. This is a quantifiable approach and therefore objective and valid. My problem with volpunter's argument was that he used discretionary "line drawers" as a definition for TA.
I've seen better approaches to trendlines in the past, methods that might not detect all the good setups but are far less random. Based on this I would think that there are far superior things out there that aren't publicized.
Technical analysis is any method for generating trading signals, said method being based solely on past and present price data and/or volume data of a security. This includes both quantifiable and unquantifiable methods, including mathematical formulae (RSI and ADX), trendlines, pattern recognition (wedges, triangles, flags, head-and-shoulders, etc.), candlestick interpretations, etc. TA is a broad collection of methods, some logic-based and others purely intuitional. Note well: it is literally impossible that academia has tested all methods that qualify as TA and found them all wanting. Therefore academic studies claiming to "disprove TA" are bunk. All they can do is focus on some narrow aspect(s) of TA. Based on the studies I've seen, their studies are doomed to failure just on the basis of poor risk management (all-in trading is a frequent assumption).
Quant is also TA but the methods are different from the old pattern stuff. For quant examples this is the website Quantocracy
The problem with this argument is that even the charts thrmselves have no consistency among traders. Every one is scalled differently and reflects the past in a different way--- there is no objectivity here. It is undefined and subjective.
>kut2k2 I just started a 'What's your definition of TA?' thread. Could you please repost this there? Thank you
> I just started a 'What's your definition of TA?' thread. Yep, can never have too many of those threads...