Wednesday, January 03, 2007

Yes, Gabriel, He Must Be

Over on another blog, Gabriel Mihalache tries to cast doubt on econophysics:
"If Doyne Farmer thinks that there are systematic, exploitable, opportunities for profit... then he must be a very rich man, right?"
I turn to the Web site of the Prediction Company, where I find:
"Founded in 1991 by Doyne Farmer, Norman Packard and Jim McGill, Prediction Company quickly set out to take the financial world by storm. Based on their earlier work in chaos theory and complex systems Drs. Packard and Farmer felt the financial markets were an example of a highly complex system that would be amenable to predictive technology. They assembled a team of world class scientists and engineers to attack the problem.

In 1992 Prediction Company signed an exclusive five year deal to provide predictive signals and automated trading systems to O'Connor and Associates, a highly successful Chicago based derivatives trading firm. In 1994 O'Connor was purchased by Swiss Bank, one of the world's largest banks. Swiss Bank extended the exclusive relationship with Prediction Company for another two years. In 1998 Swiss bank and UBS merged to create the world's third largest financial institution. Prediction Company continues its ground breaking work with UBS AG, and in November, 2005, became a wholly-owned subsidiary of UBS AG."

3 comments:

radek said...

Hmm, somehow this brings McCloskey's "If You're So Smart..." to mind.

James said...

There's a fairly interesting book about the Prediction Company as well. I think it's called "The Predictors" or something like that if you are interested. Farmer is indeed, to my knowledge, profitable. I take that to mean that he has found statistical regularities and has good execution on the implementation side. Whether he correctly identifies the causal mechanisms which generate the statistical regularities he profits from is another matter. Simulations of random order placement may be able to generate time series similar to market tick data, but I think that says more about the problems of testing economic models via simulation than anything else.

Gabriel Mihalache said...

Well, LTCM was another success story, right?

People will try. Either in researching for methods (which may or may not come through, I can't say) either via financing such research.

It makes sense for UBS to invest in this kind and many other kinds of research.

But there's a difference between making money from researching alchemy and making money with alchemy.