Algorithmic Trading
The News Is the News | Fine-Tuning Algos to Fit the Market | Trading on the News | Goldman Emerges as Global Algo Leader | Buy-Side-Driven Trading Takes Hold in Hong Kong |
Fine-Tuning Algos to Fit the Market
April 3, 2007
Over the few years that algorithms have been transforming equities trading, a persistent question has been whether and how soon these automated strategies would replace human traders. Perhaps a better question today is: What would traders do if they were suddenly denied access to algos?
Algorithms, which initially were offshoots or complements of the trading process, have since become instrumental to it. Their importance has grown far beyond the early function of efficiently working a large block order. Developers are working overtime to add features and functionality and stay ahead of the competition as the consumers of algorithms on the buy side become increasingly adept and want to do more with them.
"The first generation was the set and forget' algorithm, where you could only cancel the order if you didn't like what was going on," says Gary Ardell, managing director and head of the Financial Engineering and Advanced Trading Solutions group at agency brokerage BNY ConvergEx, which was spun off by Bank of New York Co. last year. "But it was very hard to know what was happening. Now, people are working very hard to get the algorithms to inform the trader about how the trade is doing and why."
Adds Ardell: "Traders are saying, I don't want to put my order on a bus and find out how it did after it got there. I want to know what's going on, and I want a chance to adjust to it.' That's what we're responding to."
Tony Huck, managing director of New York-based Investment Technology Group (ITG), says, "Clients are looking for more efficiency in the trading process, and if you can automate something so you can deal with those pieces that are high-cost or more difficult, that's where you want to get your focus."
The list-based algorithms that ITG rolled out last June is part of this trend toward increased efficiency. But since list-based algos help clients trade entire portfolios, they allow for far more complex strategies than were available to traders in the past. These advances are also an indication of the pace of change in the algorithmic business.
"The evolution is happening quickly, and the product development cycle is shortening," says Instinet North America president Michael Plunkett, who notes that portfolio-trading algorithms are one of the product categories where demand is hottest among buy-side traders.
Given the competitive forces in the trade execution business, and an agency brokerage business model that stresses best execution, the pressure on agencies has gotten especially intense, says Brad Bailey, senior analyst with Boston-based research firm Aite Group.
Huck says the algorithms that ITG has introduced within the past year allow a trader to, among other things, keep an order dollar-neutral--effectively selling one holding and using the cash proceeds to purchase another stock. That much is fine, but the portfolio algorithms go at least two steps further: They break up a large block and work it piece by piece as they seek liquidity wherever the relevant pools happen to be. Huck says, "That's an area where portfolio and liquidity aggregation algorithms are meeting."
"The biggest advance we've seen is that algorithms are more intelligently seeking out other sources of liquidity," says Brian Carr, CEO of Nyfix Group's Nyfix Millennium alternative trading system. "They are managing not only how much and when to trade an order, but also where to trade it. If they're not finding liquidity at one market center, they will find it elsewhere."
