# A Mathematician Plays The Stock Market

### from the not-everything-is-math dept

I’ve never been a big fan of the so-called “technical analysis” method of investing in the stock market, since it seemed to have a huge fatal flaw: it was all based on past events which might have absolutely no bearing on the future. It’s the nature of independent events. However, many people do believe in it, as well as many other “mathematical” methods for predicting the latest turn in the stock market. The San Jose Mercury News has an interview with a math professor discussing some of these methods and the people who get caught up in them in his new book, A Mathematician Plays the Stock Market. Not that the professor feels high-and-mighty about these things, because he fell for a lot of them just as hard (if not worse!) during the boom and bust. The interview does point out the one thing that I think is the most useful “trick” in trying to predict the stock market. You have to realize that you’re not really trying to predict the values of the underlying companies – but you’re trying to predict what everyone else in the market is going to do. For that matter alone, you shouldn’t necessarily ignore something like “technical analysis” just because it, by itself, is basically useless. The fact that so many investors do use it, means that it’s useful in understanding how they’re going to invest – which is useful knowledge.

Doug says:

# Flawed flaw

[technical analysis] seemed to have a huge fatal flaw: it was all based on past events which might have absolutely no bearing on the future. It’s the nature of independent events.

The flaw in the flaw is that stock prices are not at all independent of prior values; they have a very high auto-correlation. And it’s not just the prices themselves that have auto-correlation, it’s pricing trends.

I’m not well-versed in technical analysis, but from what I can tell it mainly is looking for indications that the pricing, and price trends, of a stock are being influenced by something other than auto-correlation.

# technical analysis

i believe technical analysis is important for determining the timing for buying a stock but not the stock itself, for example if i found that its a good opportunity to buy coca cola then i would then check the technical charts to find the right timing

says: