Brad, I'll check out that bibliography and get back.
As to your other points, I think you are incorrect. "Random walk" is not a theory any more than "gaussian distribution" is a theory. It is a mathematical description, and the salient question is whether it is appropriately applied or not. Unfortunately, I believe that it is essentially the null hypothesis. That is, if you can't explain the pattern in any other way, then it must be a random walk.
You make an interesting and common error when you talk about the number of people who consistently beat the averages. First, if you look at mutual fund performance you will find that the conclusion implicit in you statement is incorrect. Just look at the ranking for the funds over the past 40 - 50 quarters and then draw your own conclusions. But here's a simple test: the probability of a return's beating the S&P should be 50% (in fact it is less due to expenses and salaries). Therefore, the probability of a firm beating the averages four quarters in a row is .0625, the probability of a firm beating the average 3 out of the last four quarters is 18.75%, the probability of a firm beating the averages 2 out of the last four quarters is 37.5%.
Now, this distribution was studied about 20 years ago, and it basically found that there was no correlation between how well a fund did in the past to its next year's performance.
In addition, you have ascribed positions to me with which I do not agree. For example, it is possible to out-perform the market year after year. This in no way implies any validity to TA. Second, I don't believe in the strong, semi-strong, or semi-weak forms of the efficient market hypothesis - there is a significant body of evidence against it. All that I've said is that I've seen no credible evidence that TA works, and that I've seen lots of evidence that it performs no better than random.
Regards,
Paul |
