A smarter computer to pick stocks

But some are aware that a former Microsoft executive and chairman of the Nasdaq stock market, Michael W. Brown, is an investor in Kurzweil’s new hedge fund, FatKat, and that Bill Gates once described him as “the best person I know at predicting the future of artificial intelligence.”

More important, many of them have seen Kurzweil’s ideas used by stock speculators. So, they want to learn more about his brave, new world.

“These ideas are the future,” said David Atkinson, a private investor who attended another lecture later that day by Kurzweil. “I’m not really sure I understand them, but they’re making some folks rich.”

Complicated stock picking methods are nothing new. For decades, Wall Street firms and hedge funds like D. E. Shaw have snapped up math and engineering PhDs and assigned them to find hidden market patterns. When these analysts discover subtle relationships, like similarities in the price movements of Microsoft and IBM, investors seek profits by buying one stock and selling the other when their prices diverge, betting historical patterns will eventually push them back into synchronicity.

Today, such methods have achieved a widespread use unimaginable just five years ago. The Internet has put almost every data source within easy reach. New software programs, like the Apama Algorithmic Trading Platform, have made it possible for day traders to build complicated trading algorithms almost as easily as they drag an icon across a digital desktop.

“Five years ago it would have taken $500,000 and 12 people to do what today takes a few computers and co-workers,” said Louis Morgan, managing director of HG Trading, a three-person hedge fund in Wisconsin. “I’m executing 1,500 to 2,000 trades a day and monitoring 1,500 pairs of stocks. My software can automatically execute a trade within 20 milliseconds – five times faster than it would take for my finger to hit the buy button.”

Studies estimate that a third of all stock trades in the US were driven by automatic algorithms last year, contributing to an explosion in stock market activity. Between 1995 and 2005, the average daily volume of shares traded on the New York Stock Exchange increased to 1.6 billion from 346 million.

But as these new techniques proliferate, some worry that promotion is outpacing reality. These techniques may be better for marketing than stock picking.

“Investment firms fall over themselves advertising their latest, most esoteric systems,” said Lo of MIT, who was asked by a $20 billion pension fund to design a neural network. He declined after discovering the investors had no real idea how such networks work.

“There are some pretty substantial misconceptions about what these things can and cannot do,” he said. “As with any black box, if you don’t know why it works, you won’t realize when it’s stopped working. Even a broken watch is right twice a day.”

–CHARLES DUHIGG