When computers run amok on Wall Street, they make millions of dollars’ worth of bad stock trades, causing markets to nose-dive and traders to consume their weight in anti-anxiety medication. That was the case on August 1, when a software bug launched a series of bad orders on the U.S. stock exchanges that cost Knight Capital Group roughly $440 million.
Over the past few years, software that allows traders to buy and sell commodities at lightning-fast speeds has changed how a portion of Wall Street operates. Why let a human being make a big bet on a stock shifting over the course of a day or week, goes the reasoning, when a software platform churning through massive amounts of financial data can earn lots of money off lots of small, microsecond trades? However, those high-speed algorithms have also been blamed for bizarre market activity, including the May 2010 “Flash Crash.”
Given how software can spectacularly malfunction, at least one technology expert is advocating for a system of processes and controls to prevent financial software from accidentally going Skynet.
“What is needed is a change in the way that such critical software is developed and deployed,” Robert Dewar, president and CEO of AdaCore, wrote in an Aug. 8 posting on Dr. Dobb’s Website. “Safety-critical domains such as commercial avionics, where software failure could directly cause or contribute to the loss of human life, have known about this for decades.”
Those critical industries have developed a culture that incorporates hazard and safety analyses into the software-development process. “Until the software has been certified as compliant with the standard, the plane does not fly,” Dewar added. “High-frequency automated trading is not avionics flight control, but the aviation industry has demonstrated that safe, reliable real-time software is possible, practical, and necessary.”
But will it take more market weirdness before Wall Street firms actually introduce such measures into their software?