Although many people invest in stock markets to build wealth for retirement, either directly or through mutual funds, precious few understand exactly what happens on the trading floor every day.
That’s the first misconception: There isn’t a trading floor anymore. There was a time not too long ago when wildly gesticulating floor traders represented capitalism itself, trading stocks that financed the birth of companies and the creation of jobs. It wasn’t clear exactly what they were doing, but somehow they made our investments grow, and that was good enough for us.
Today, traders tap computer keys to make their trades, but even that is an increasingly outdated approach. Roughly three-quarters of all trades today are made by high-frequency traders who use powerful computers and sophisticated algorithms. The computers react to changing data and make hundreds of trades instantaneously.
One of the leading companies producing software for high-frequency traders is Cyborg Trading Systems, based in downtown London, the creation of two UWO grads and a third partner.
James McInnes earned his B.Sc. in Genetics in 2002, but went to work trading derivatives and other financial instruments in Montreal. There he met Ben Bittrolff, also a trader. The pair began talking about the prospects of creating their own trading tools – kind of like two engineers at GM discussing how to build a better automobile in their spare time.
When they decided it was something they wanted to pursue, McInnes returned to UWO in 2007 and studied Computer Science for three semesters, upgrading his knowledge to the point where he was ready to become CEO of a company dedicated to revolutionizing high-frequency trading.
The idea of creating their own trading tool quickly evolved into the concept of making a tool they could sell to traders of all kinds, at various levels of sophistication, something good enough to put their customers on an even field with trading behemoths like Goldman Sachs.
McInnes knew Peter Metford through a family connection, and he became the third partner. Metford earned his B.Sc. in Physics at UWO in 1976 and his Ph.D. in Electrical Engineering at McMaster.
Together they created a set of algorithms that direct computers to make trades literally at the speed of light. High-frequency trading takes advantage of the milliseconds after someone offers to buy a share before that desire is broadcast to the market. With that information and brute computing strength, trading desks and hedge funds can outmaneuver other traders and earn a few cents on millions of trades a day.
To make it work, traders need very sophisticated software – known as a black box -- that can execute trades without human approval or intervention.
Cyborg has built a grey box, a set of equally powerful algorithms that are designed to make it possible for traders of varying sophistication to trade with the big boys. It is, as the name suggests, a combination of man and machine. Infinitely customizable, it can help hedge funds do battle with Goldman Sachs or it can be customized for individual investors using a retail web-based broker.
The latest incarnation of the product is the Cyborg Cloud Trader, a product the company will launch this summer, providing a flexible framework for institutional traders and hedge funds to conduct high-frequency trading.
“If you’re simply relying on being the fastest, there’s always someone or something coming along that’s faster,” says McInnes. “But if you can be fast, not the fastest but incredibly fast, and also be flexible so you can change the algorithms quickly to react to changing data, that’s the best combination. That’s our strategy.”
For that kind of flexibility, institutional traders pay Cyborg $10,000 to $40,000 in ongoing licensing fees to use the product.
Cyborg is setting up working relationships with two UWO professors who are interested in different aspects of the company’s work and hope to get their graduate students involved with the company.
Applied Mathematics Professor Matt Davison, a former trader himself, is interested in examining the Cyborg trading data at the equivalent of the molecular level to detect trading patterns.
“A typical approach to trading is to analyze data over a few days, look for trends and develop a strategy,” he says. “But Cyborg has data every millisecond. There may be patterns there, but you can’t see them or you can’t react to them. We want to apply our expertise in data analysis to what Cyborg is collecting.”
Computer Science Professor Stephen Watt studies the way computers compile code to complete tasks. The more efficient the compilation, the more quickly the computer does its thing.
“It’s a question of how to convert information into machine instructions,” he says. “That’s program optimization. If the computer compiles instructions twice as fast as before, you can double its speed without changing the hardware.”
He also studies handwriting recognition and sees an opportunity to apply pattern analysis to the Cyborg algorithms.
“Once you figure out how to make a computer learn, to recognize patterns, you can apply it to handwriting or market analysis. It’s the same concept.”
McInnes is excited at the prospect of collaborating with students and professors. “We love the idea of doing research together,” he says.
Login to view and post comments
Western University, Department of Communication & Public Affairs
Westminster Hall, Suite 360, London, Ontario, Canada, N6A 3K7
Tel: 519-661-2045 • Fax: 519-661-3921
General inquiries: wag.editor@uwo.ca
Website feedback: wag.web@uwo.ca
Accessibility | Privacy | Web Standards | Terms of Use
