Wednesday, October 21, 2009

Information Overload

More than 35 years ago, in a now-classic piece of research, Paul Slovic reported on the ability of handicappers to predict the outcome of horse races. At first, Slovic let the handicappers use any five pieces of information they wanted -- the jockey’s weight, for instance, or the horse’s previous race performance. Then he let them use 10 pieces of information. Then 20, and, finally, 40.

Slovic was studying the stress caused by information overload. What he found was fascinating: with more information, the accuracy of the predictions did not improve -- it was no better with 40 pieces of information than it was with 5.

But the confidence in those predictions did improve. This increased substantially, from less than 20% with 5 pieces of information to more than 30% with 40 pieces of information.

I thought about Slovic's findings as I read about the Galleon Group, the hedge fund whose rotund leader, Raj Rajaratnam, was arrested earlier this week on insider trading charges.

The pressure at Galleon was intense. According to published reports, analysts there were browbeaten to come up with new information on companies whose stock the fund was interested in.

But that information, at least in some cases, didn't seem to help -- and may even have hurt. As Alex Berenson reported in today's New York Times, Mr. Rajaratnam "lost millions of dollars from what prosecutors characterize as insider trading."

In one case alone, involving the chip maker Advanced Micro Devices, Galleon reportedly lost $30 million.

What is interesting here is not that some of Galleon's trades went bad; one would expect that. What is interesting is that Galleon apparently didn't learn what Slovic did: More information doesn't necessarily lead to better decisions. But it can lead to costly ones.

Labels: , , , ,


Post a Comment

Subscribe to Post Comments [Atom]

<< Home