Editor’s note: One of the most interesting things about the catastrophe at Knight Capital Group—the trading firm that lost $440 million this week—is the speed of the collapse. News reports describe the bulk of the bad trades happening in less than an hour, a computer-driven descent that has the financial community once again asking if its pursuit of profits has led to software agents that are fast yet dumb and out of control. We’re posting this story in advance of its publication in Wired’s September issue because it examines how Wall Street has gotten to the point where flash failures come with increasing frequency, and how much further traders seem willing to go in pursuit of ever-greater speed.
THE 2012 NEW York Battle of the Quants, a two-day conference of algorithmic asset traders, took place in New York City at the end of March, just a few days after a group of researchers admitted they had made a mistake in an experiment that purported to overturn modern physics. The scientists had claimed to observe subatomic particles called neutrinos traveling faster than the speed of light. But they were wrong; about six months later, they retracted their findings. And while “Special Relativity Upheld” is the world’s most predictable headline, the news that neutrinos actually obey the laws of physics as currently understood marked the end of a brief and tantalizing dream for quants—the physicists, engineers, and mathematicians-turned-financiers who generate as much as 55 percent of all US stock trading. In the pursuit of market-beating returns, sending a signal at faster than light speed could provide the ultimate edge: a way to make trades in the past, the financial equivalent of betting on a horse race after it has been run.
“Between the time the first paper came out in September and last week, a guy in my shop had written two papers explaining how it could be true,” a graying former physicist said ruefully, sipping coffee near an oversize Keith Haring canvas that dominated the room at Christie’s auction house where the conference was held. “Of course, you’d need a particle accelerator to make it work.”
Wall Street used to bet on companies that build things. Now it just bets on technologies that make faster and faster trades.
If that were all it took, then by now someone would be building one. One of the major themes of this year’s conference was “the race to the bottom,” the cost-is-no-object competition for the absolute theoretical minimum trade time. This variable, called latency, is rapidly approaching the physical limits of the universe set by quantum mechanics and relativity. But perhaps not even Einstein fully appreciated the degree to which electromagnetic waves bend in the presence of money. Kevin McPartland of the Tabb Group, which compiles information on the financial industry, projected that companies would spend $2.2 billion in 2010 on trading infrastructure—the high-speed servers that process trades and the fiber-optic cables that link them in a globe-spanning network. And that was before projects were launched to connect New York and London by a new transatlantic cable and London and Tokyo by way of the Arctic Ocean, all just to cut a few hundredths of a second off the time it takes to receive data or send an order.
High-frequency traders are a subset of quants, investors who make money the newfangled way: a fraction of a cent at a time, multiplied by hundreds of shares, tens of thousands of times a day. These traders occupy an anomalous position on Wall Street, carrying themselves with a distinctive mixture of diffidence and arrogance that sets them apart from the pure, unmixed arrogance of investment bankers. A pioneering high-frequency trading firm, Tradeworx, has its relatively humble offices two flights up from an Urban Outfitters in a sleepy New Jersey suburb. Twenty people work there, about half of them on the trading floor, monitoring on triple screens the fractions of a penny as they mount up, second by second. Roughly 1.5 percent of the total volume of stocks traded on US exchanges on a given day will pass, however fleetingly, through the hushed, sunlit, brick-walled room.
On the first day of the New York conference, Aaron Brown, a legendary quant and former professional poker player, took the stage in rumpled chinos and a leather jacket to lecture the assembly on game theory. He began his talk by saying, “3.14159,” and then pausing expectantly. From the back of the room came the response: “265358.” Together they made up the first 12 digits of pi—a geek shibboleth. “You won’t see a lot of masters of the universe here,” said Charles Jones, a professor of finance and economics at Columbia Business School. “A lot of these guys, if they’re wearing a tie, it might be the only one they own.”
Faster and faster turn the wheels of finance, increasing the risk that they will spin out of control, that a perturbation somewhere in the system will scale up to a global crisis in a matter of seconds. “For the first time in financial history, machines can execute trades far faster than humans can intervene,” said Andrew Haldane, a regulatory official with the Bank of England, at another recent conference. “That gap is set to widen.”