By SCOTT PATTERSON
A high-tech, high-speed poker game is playing out in the stock market, and billions of dollars are at stake.
The adversaries are high-frequency traders and big investors such as mutual funds. High-speed firms are using computers to detect large "buy" and "sell" orders to adjust their trades, and traditional investors are scrambling to trade undetected.
The showdown has led to an escalating arms race, with players on both sides plowing money into ever-more-powerful technology to trade effectively. It also has led to growing tension between these camps as conventional investors call for greater regulation of their high-frequency counterparts.
The clash came to the fore after the May 6 "flash crash," which underscored the rising use of computers and algorithms to trade stocks. In testimony last week before federal regulators investigating the market plunge, high-speed firms contended they are misunderstood and that critics spread false rumors about their trading methods. Traditional money managers said fast traders have harmed their ability to trade stocks and regulators need to step in.
"Attempts to classify some trading styles as 'desirable' or 'undesirable' are unproductive," said Dave Cummings, chairman of Tradebot Systems Inc., a Kansas City, Mo., high-speed firm, at the June 22 hearing.
Jeff Engelberg, a trader at Southeastern Asset Management Inc., a Memphis, Tenn., value-investing firm with about $35 billion under management, said high-speed traders are jumping ahead of his firm's trades. "Short-term traders are able to get an instantaneous glimpse into the future" through direct feeds to exchange data, he said, turning the market into "something nearer to a casino."
High-frequency trading accounts for about two-thirds of U.S. stock-market volume, according to industry estimates. Advocates of the practice include some heavy hitters, such as mutual-fund giant Vanguard Group Inc., which says it helps lower trading costs by narrowing the spread between what investors pay to buy and sell shares.
Traditional money managers largely agree that some costs have dropped. But some say they find certain trades have become more expensive to carry out, thanks to a practice critics call "gaming." With gaming, if a high-speed firm's computers detect a large buy order for a stock, for instance, the firm will instantly start snapping up the stock, expecting to quickly sell it back at a higher price as the investor keeps buying.
Phillip Krauss, head of stock trading at Chicago's Harris Investment Management, says his firm's orders are getting picked off by high-speed trading firms that use computer programs to detect orders. "They front-run us," said Mr. Krauss, whose firm manages $15 billion in assets.
To protect themselves, traditional investors like Mr. Krauss are stepping up investments in "antigaming" computer power and expertise. Big firms such as Bank of America Corp., acting on behalf of thousands of individual investors who trade through its vast brokerage network, have been rolling out high-tech platforms to help their traders avoid getting gamed.
"Antigaming was a big focus for us last year, and it's going to be a big focus for us this year," says Lee Morakis, head of sales for Bank of America's execution services for the Americas unit.
On Capitol Hill, Sen. Ted Kaufman (D., Del.), an advocate of regulation of high-speed trading, said his office has been fielding complaints from players across Wall Street who say gaming is tilting the field against them.
The Securities and Exchange Commission in January issued a report outlining its concerns about high-frequency trading and related consequences of recent changes in the structure of markets. The agency recently unveiled a plan to give large high-frequency firms unique identification codes to better track their activities.
While gaming isn't blamed for the May 6 crash, that day's downdraft has increased scrutiny of high-speed trading generally. High-speed firms have touted their ability to smooth trading by constantly stepping in to buy and sell stocks. But when trading became chaotic that day, many of these firms—along with other investors—pulled back.
Thirty-eight percent of so-called buy-side money managers—typically long-term investors—say they have a more negative view of high-frequency trading than before the flash crash, according to a recent survey by Tabb Group; 17% expressed positive views.
In many ways, the gaming dynamic isn't new but instead represents an evolution of the market. Wall Street insiders have always tried to use information about competitors' intentions to gain an edge.
Now, information is embedded in streaming market data and computer codes. Algorithms are expected to account for about 60% of stock trading this year, up from 28% in 2005, according to Aite Group, a Boston firm that tracks electronic trading.
A sharp fall in trading size shows the impact of algorithmic trading, which is typically done in 100- or 200-share chunks. The average size of a trade of a stock listed on the New York Stock Exchange was 268 shares through most of 2009, down from 724 in 2005, according to the SEC.
That has made it much more difficult for traders to buy or sell large chunks of stocks. Now, traders must parcel out orders in tiny bits—and some high-speed firms are monitoring the activity for signs of a big trade.
Henri Waelbroeck, research director at New York electronic brokerage Pipeline Financial Group Inc., has designed a trading system that alternates algorithms to keep gamers off the scent of big trades. "You need to understand what the games are and how to defend yourself," Mr. Waelbroeck said.
Write to Scott Patterson at email@example.com