Tue, 10 May 2005

Historical Mathematics

Imagine if mathematicians were taught to understand mathematics in terms of the history of math. In order to discern that 1+2 is 3 and that 2+1 is equally 3, you would have to look at the history of it. Has this relation been true in the past? If so, the historical mathematician thusly concludes that it is a relation that will continue in the future.

Sounds like nonsense, doesn't it? It is. Now imagine a branch of economics that does the same thing, called historical economics. It would be nonsense, and since it actually exists, it is nonsense.

Tradestation is a service which allows individuals to trade stocks on the market by entering orders under the control of a program. The program has access to all the prices of the stock sampled at five minute intervals going back years, and daily intervals before that. Using an appropriate program, you may create a theory about the market. You can test your theory against the historical prices by running your program in test mode, to see how it would have traded.

This, too, is nonsense. Prices on the market are determined by a large number of factors. These factors will not be the same in the future as they were in the past. People's opinions change, their trading method changes, companies change, industries change, and economies change. When you're writing a Tradestation program for the past, that is all that it will reliably succeed at.

This does not mean that Tradestation is useless, or harmful. It can be used to test theories about the market, but those theories must be tested using new data, not past data. Otherwise you're just fitting your theory to the shape of the curve-that-used-to-be.

So how do you trade using Tradestation successfully? You trade on the fundamentals of a stock. You build on what previous traders learned; for example Ben Graham, or Warren Buffett. You look for special situations: a stock that is undervalued by the market.

This is also what the successful economist does: creates a theory about some economic principle based on what other economists have learned; for example Ludwig von Mises, Freidrick Hayek, Ronald Coase, etc. Then she looks to see what people actually do. If she is right, then the theory is a good one. If she is wrong, she goes back to square one with a new theory.

Posted [09:20] [Filed in: ] [permalink] [Google for the title] [digg this]