Computer Models Won’t Outperform Market Any Time Soon

Although the quest to beat the market with computer models has “absorbed the talents of some of the brightest graduates of math and computer science programs,” a recent article in Bloomberg argues that it isn’t likely that a computerized stock picker will outsmart the market.

The article outlines some of the “devilish problems financial engineers are trying to crack:”

  • The data keeps changing: In the financial markets, the article says, “‘data can change dramatically and in unprecedented ways,” citing the examples of negative interest rates across Europe and Japan in 2013 and the 1998 change in the U.S. stock pricing format from fractions to decimals that “flustered” some traders.
  • There’s more noise than signal: Most market movements are referred to by economists as “noise trading” and not based on anything meaningful. “What’s more,” it adds, “as data sets go, the history of stock prices is relatively thin.”
  • Any edge you’re looking for is small: Most investors are unable to get an “edge” because many signals are barely detectable. If a signal is obvious, it adds, “it’s going to be quickly discovered and traded away by others.”

“To build a truly autonomous investing system—one in which the computer itself is thinking about signals and strategies to try—researchers will likely need to crack the problem of causality. That means not only noticing that, for instance, a rise in a particular stock is often accompanied by a bump in interest rates, but also being able to come up with a reason for it.”