People have been trying to crack the code of the stock market for as long as the institution has existed, trying to find secret tips and tricks hidden in the charts to predict future performance. A study by the University of California Riverside suggests that there may be a set of magical stock tea leaves out there, and that they're buried somewhere in the social networking site Twitter.
A team of computer engineers and scientists, led by professor Vagelis Hristidis, set out to determine if there was any correlation between the 340,000,000 Twitter posts per day and the performance of the companies at the tips of everyone's Twitter tongue. And it turns out there is: The more a facets on which a company is being discussed, the stronger that company performs in the stock market. If a discussion about company centers around only a small number of news items or discussion topics, its stock underperforms.
After creating a model, Hristidis' team put it to the test. Over a four-month trial, the stocks predicted by Twitter to be top performers lost 2.2%, as compared to an overall 4.2% loss for the Dow Jones Industrial Average. A number of other trading algorithms underperformed the Twitter model, as well.
For his part, even Hristidis is puzzled by his findings, saying, "To be honest, I'm not sure why we found the correlation." It has been hypothesized, however, that the kind of bad news that drives down stock prices tends to focus people's attention, while good news is typically less interesting and prone to creating tangents.
[Image credit: Katrina Tuliao]
More from Tecca: