How computers read the markets

To most humans, the following text makes little sense: "Symbol ticker = 'MAN' country = 'US' cusip = '56418H100' isin = 'US56418H1005' / symbol changesign = '+' caltype = 'percent' 2.5 / change to price value = '44.52'."

But to a new breed of computers programmed to trade securities automatically on the latest news stories, it could be a means of making a huge sum of money.

Hedge funds and investment bank trading desks are pouring unprecedented sums into such technology to find faster and more inventive ways to outsmart their rivals.

News has always affected market prices. And there already are software programs that track headlines and alert traders if certain market-sensitive terms frequently appear. "Hurricane" could signal a shift to sell insurance stocks. "Drought" could affect wheat prices.

But computers have acquired the ability to analyze years' worth of news stories to see how certain headlines affected market movements. Those patterns can be used to program the machines to trade on the latest developments.

Computers are now being used to generate news stories about company earnings or economic statistics as they are released. This almost instantaneous information forms a direct feed into other computers, which trade on the news.

The result is a boom in demand from news and information providers such as Reuters, Bloomberg and Thomson Financial for "machine readable news," which is written in a computer-friendly language of strings of words and numbers. Computers can trade on such news within milliseconds of receiving it.

"One of the big consumers of news now is a computer," says Matthew Burkley, a senior vice president of strategy at Thomson. "This area has turned out to be broader than we thought. Instead of being limited to a marginal number of our clients, the demand for news which is readable by a computer is very widely spread."

Reuters reports similar demand. "There is real interest in moving the process of interpreting news from the humans to the machines," says Kirsti Suutari, global business manager of algorithmic trading at Reuters. "More of our customers are finding ways to use news content to make money."



The human eye is far from redundant, however. "News events are extremely subjective," says Will Sterling, head of institutional electronic trading at UBS. "Our general approach has been to blend the automation … with a degree of human oversight. It's better to take an extra few seconds to be sure."