Dallas, Pa., is a long way from Wall Street. It’s a borough of 2,500 people in the hills outside of Scranton. But from his basement home office there, John Joseph is engaged in the kind of super-fast trading that is changing the world’s securities markets. Joseph, 36, considers himself closer to a sophisticated retail trader than he is to the big institutions that dominate the business. Yet in trading he’s holding his own.
“We are beating on a regular basis everyone else in the world on the trade we’re trying to get,” he claims.
High-frequency trading is a lightning-fast, high-volume endeavor that has emerged in the past few years as the hottest thing on Wall Street. It’s dominated by proprietary trading groups, big banks and hedge funds like Getco, Goldman Sachs , Barclays and Citadel. But it’s possible, although risky, for individuals to dip a toe into this world, as Joseph has.
Individuals with a hankering to give it a try should be forewarned: They will be going head-to-head with some of the sharpest minds, and deepest pockets, on Wall Street. Most are likely to lose. That’s because the hyper-competitive business involves writing complex algorithms and using staggering amounts of computing power to capitalize on tiny inefficiencies in equity, futures and options markets.
What makes Joseph think he’s got a fighting chance is the fact that he’s a math whiz with the skills needed to carve out a niche. Joseph was in his fifth year working on a doctorate in mathematics at the University of Massachusetts when he had the somewhat belated realization that he wasn’t cut out for academics. So in 2000, he left school to research technology companies and was hired by a boutique brokerage firm that went bust three weeks after he started. That brief tenure in the securities business was enough to hook Joseph on the notion that automation could improve trading.
Joseph spent the next three years in marketing and spent his free time learning how to build automated trading programs. When he was convinced his programs worked, and that he’d need more time and money to make them really successful, he left his job and used his algorithms to trade futures from home. He now oversees accounts for himself, friends and outside clients worth $10 million that are devoted to the algorithmic trading model he developed.
Joseph’s algorithms essentially turned him into an automated day trader. Initially, he held positions for between two minutes and 2.5 hours. In contrast, the sort of high-frequency trading now taking Wall Street by storm involves buying and selling securities dozens, hundreds or even thousands of times a second. Last year Joseph got into that game too.
He was introduced to it by two former floor traders who’d become interested in electronic trading. Joseph translated their concepts into algorithms, which another partner then coded. They formed a small proprietary trading shop called Rooftop Trading. Joseph’s new algorithms are trading 2,000 times faster than his older ones. He declines to disclose the value of the assets he and his partners have devoted to their high-frequency trades.
With billions of dollars at stake, high-frequency traders are engaged in a high-tech arms race in a bid to beat each other and investors to the most lucrative trades. The rivalry is a game of “co-locating,” which involves hosting computers in data centers where exchanges keep their own computers. Big firms can spend $200,000 a month for such prime real estate. Through his broker, Joseph uses a service called Zen-Fire to lease a server in such a data center. The cost is bundled into his brokerage commissions, which vary depending on how much his firm trades.
Zen-Fire’s founder, Patrick Shaughnessy, says a growing number of retail traders are writing their own algorithms. After signing a confidentiality release, he will tell them if he thinks their work has a fighting chance in the market. He sells the few that seem promising super-fast connections to futures and foreign exchange markets. He plans to offer rapid-fire stock trading next year.
Back in his Dallas, Pa., office, Joseph uses three computers, including standard desktop models and a laptop, to remotely tap into Zen-Fire’s computer at the data center and tweak his high-frequency algorithm. He uses both cable and DSL lines for redundancy, knowing that a strong storm could knock out his Internet connection. If it does, his algorithms will continue to run at the data center without his oversight.
Like big firms that jealously guard their secrets, Joseph won’t talk about his strategy except to say he finds inefficiencies in the market that bigger firms with more capital are unlikely to care about.
Joseph says he’s making money but warns it’s been harder to find opportunities this year because long-term traders and investors are hanging onto their positions.
“I tell my wife every dollar we make could very easily be the last,” he says.
_________________________________________________________________________________________________
AWS with Metatrader, developing algorithmic trading for cloud based large scale computing With support for True Multiprocessor execution.
MetaQuotes should enable Amazon Web services.
