Saturday, January 01, 2011

Detecting Author Sentiment, Then Trading on It

The New York Times summarizes some of the services for algorithmic trading that Bloomberg, Thomson Reuters and Dow Jones have been providing for a while. The journalist, Graham Bowley, focuses quite a  bit on the cutting edge stuff, like extracting emoticons and sentiment-bearing words from news reports, press releases, Twitter and other places.

But the real gravy to date has been extracting the more predictable, such as economic indicators, and allowing trading houses to get that data in milliseconds. This allows them to make real money in a flash before humans can react.

No comments: