I've always been suspicious of TA, because the statistical basis for its projections is largely invalid. That doesn't mean it doesn't work. It just means that there is no real theoretic basis for it. As you say, it's an art, not a science.
My son, who does a lot of programming under contract with the military, designed a program using a type of filter that removes unwanted signals and reacts quickly to those signals (i.e., variables) that could predict the direction being taken by a missile, or a stock. Trying it out on a series of selected stocks, some of which were going up, and other which were going down, he was able to get a return averaging 2 percent a day. Now THAT'S more like it! |
