Trading is such a free field activity packed with choices and so much data can be inputted that it is not suprising that it can be done badly in an infinite number of ways. So TA as the basic philosophy of trading can be made to look ineffective. But there is no way that the basic truths of TA are incorrect -
1. prices moving in a given direction have a stronger tendency to continue than to reverse
2. price trends are interrupted by minor periods of neutral or counter-trend price action.
i.e., rising prices go up - but not in a straight line.
If you only know those two things, you can work out a profitable trading strategy. Most strategies will fail, but not because these TA truths are incorrect, but because of faulty money management, exacerbated by decisions to rely on technical indicators which are assumed to reduce risk.
Can't remember exactly how it goes, but there's some Mark Twain quote about how history doesn't repeat itself, but rhymes with itself (I'm sure he put it more poetically than that). But the gist is that using TA (ie expecting historical patterns to repeat themselves) isn't going to be an exact science. I reckon TA gets a hard time, much like weathermen do - we have unrealistic expectations of how often they can get it right. Nothing can really predict the future, they're just making predictions based on past data.
Yes and given the price of an instrument, couldn't you make predictions as I have above about where price will or will not go with greater than random chance? Therefore isn't price predictive by your definition. And therefore if TA is any analysis using past price, and price as it is unfolding right now, then of course it is predictive.
I thought perhaps you meant prediction as being a certain thing.
I think part of the problem with TA is that there is so much to5h written about it that people misrepresent it or think it's something that it is not.
There is no doubt in my mind that a solid trading edge can be formed from a strategy with an element of random entry however I also believe it is possible to construct an edge with a historic probability of working out.
Of course we do not know the certain outcome of the next trade but we do have a historic probability. lets say the last 1000 times a,b,c & d happened 90% of the time we get X, 10% of the time we get Y. On the next trade #1001 do you think X or Y is more likely. We cannot predict that X will happen but over the next 10 trades X is likely to happen 9 times and Y once, all on past performance alone, everything else being equal.
From my observations the markets are much more neuro-behavioral than I originally thought. Traders bring all of their own irrationality and biases to the marketplace.
Previous support and resistance and/or round numbers are areas where there is likely to be a battle between buyers and sellers (and liquidity games). So I like to look at a chart and say ok if price comes to this point again there is going to be a battle for control.