The Trading Test

Brian Eckerly, an electrical engineering student at Ohio State University, booted up his Dell laptop one morning in January and loaded a little program he'd finished coding the night before. Numbers flashed on the screen, and Eckerly scanned them for a minute. Then he went to class.

That day, and in the following weeks, the laptop sat undisturbed at Eckerly's off-campus apartment in Columbus, carrying out his program's instructions. It connected to an online brokerage firm, gathered stock data, crunched some statistics, determined whether certain conditions had been met and, if so, executed trades. Eckerly began with a handsome sum--US $100 000--and in seven weeks his program increased it to $394 190.

If only it were real money!

It was all for a trading contest organized by Interactive Brokers Group, a $2.8 billion securities firm in Greenwich, Conn., that wanted to find tech-savvy engineers and scientists willing to work in the financial industry.

”The old trading world required big, aggressive, street-smart folks--now it's technology skills that count,” says Steven J. Sanders, a senior vice president with Interactive. ”We just don't ever get enough good technologists.”

The contest, which was open to students from the United States, Canada, and Mexico, required participants to elaborate a trading strategy and write software to execute it. Each contestant took a starting stake in virtual money and used it to trade stocks, bonds, options, futures, and currencies during an eight-week period. Although all trades were virtual, taking place on a simulation system created by Interactive, the buying and selling prices were based on real market data.

Real, too, were the cash prizes. Eckerly, who finished first, took home $100 000. Two runners-up each got $50 000, and several other participants got prizes of $10 000 or $1000.

The financial industry has long been known to hire math whizzes as quantitative analysts, or ”quants,” who concoct pricing models, probe new ways to quantify risk, and mine data. Now, as automated trading systems take over ever more of the substantive work on Wall Street, many firms are seeking quants who not only know the math but the nuts and bolts of IT systems, too.

”There's just a huge engineering challenge: How do you get that much data, process it very quickly, and act on that?” says Richard Holowczak, a professor of computer information systems at Baruch College of the City University of New York. ”It's definitely like an arms race.”

Many traditional universities' finance and management schools are introducing courses in quantitative and computational finance. As a result, there are now 40 financial engineering programs on offer throughout the world, about three times as many as five years ago.

Konstantinos Tsahas, a master's degree candidate in Baruch's program, says the best thing about the contest is that it allows students to experiment with real-world trading systems. He compares the competition to ”an eight-week internship from your home.” If so, then his second-place finish made it one of the best-paid internships on record, at $6250 per week.

Interactive's founder and chief executive, Thomas Peterffy, has been a longtime proponent of electronic trading. In 1983, he developed a handheld computer to assist him on the trading floor--much to the resentment of his fellow traders, who promptly tried to get it removed. Peterffy went on to develop ever more complex pricing and risk-management systems.

Now his company devotes entire floors of a modern orange-brick office complex to an army of software engineers and system administrators. Their main job is to keep on improving the firm's vast trading system, which runs at a data center with 300 Linux-based servers and fiber-optic connections to dozens of exchanges around the world.

Sanders, the vice president, acknowledges that his trading contest is not a cheap way to snare talent, for in addition to the $400 000 prize fund, there are marketing costs, including ads on Web sites such as Facebook. But, he says, it's worth it.

”The contest works as a filter that gives us much more than just a résumé to look at,” Sanders says. Last year, after the firm ran the contest for the first time, it hired Bharath Govindarajan and Stan Li, the second- and fifth-place finishers, respectively. This year's contest hadn't led to any new hires by June, but that may change as graduating students enter the job market. The company, which has a dozen technology positions to fill, is already planning next year's contest.

This year's 204 contestants wrote their trading software in C, C++, Java, Visual Basic, and even Excel scripts. To retrieve market data and execute trades, the programs communicated through an application program interface with Interactive's Trader Workstation software, which had to run on the same computer. Professional traders using Interactive's services run their trading algorithms the same way.

Trading strategies varied, but most of them used well-known techniques, such as moving averages, Bollinger bands, and vector analysis. Sanders insists that he's not interested in appropriating the students' strategies. Contestants, he says, don't have to submit their software, only outline their trading plans.

Eckerly, who learned about investing from his grandfather, a stockbroker, won't reveal all the details of his strategy. He says his program analyzes the Standard & Poor's 500 stock index to tell when prices rise or fall to an extreme value, then bets they will move in the opposite direction. It obtained high returns by using a leveraged instrument called a put option, basically a contract from which the holder profits if the stock's price goes down.

Eckerly graduated in March and is off to a job as a data analyst with Capital One Financial Corp. in Dallas. He says he'll invest most of his $100 000 prize, but he doesn't know if he'll use a program to do that. ”When you're dealing with real money,” he says, ”you have to be more serious about your decisions.”