Quote from intradaybill:
The problem with TA is that you will maybe get a good signal and make $X but then lose 2 x $X in the next 5 signals. In books people will include the pattern that made $X and "forget" the other 5 that lost $2X. It is a behavior borrowed from everyday life. We remember only the good times, good people, good places, etc. We tend to forget bad situations. But in the case of trading that is very costly. So the conclusion is that most TA book authors adopt selection bias. IMO it is extremely hard to maintain positive expectancy with TA in the longer term. For that you have to go automated and use rigorous math.
TA vs. not TA is becoming like the Creation vs. Evolution debate.
Man, you just don't understand how TA works! All you have to do is add a second critter-ion that works one time out of five and that fixes everything!