12:05 PM
Custom Algos Growing in Popularity
Today's buy-side traders wouldn't be caught dead with off-the-rack trading algorithms. Now that algorithmic trading strategies for U.S. equities have become commoditized, custom algos are all the rage on buy-side trading desks. Buy-side traders know exactly what they want, and they aren't shy about demanding it from the sell side and independent algo developers.
Fifty-two percent of head traders on U.S. equity trading desks see no differences in current algorithmic offerings, according to a recent Tabb Group survey. In fact, some traders who participated in the survey pointed to a once-unthinkable thought: that all algos are the same. "Although this may be partially true," the study noted, measuring those differences is difficult and expensive, Tabb Group reports.
"Basic strategies aren't cutting it anymore -- all the brokers offer the same strategies," says Tabb senior analyst Cheyenne Morgan. "It's the level of customization and sophistication that is probably the next step in terms of competition [among brokers]."
AT's Top 11 Stories of 2011
- Top 10 Occupy Wall Street Protest Signs
- The 9 Most Notorious Insider Trading Cases of 2011
- Meet the Team Behind JP Morgan's Electronic Trading Transformation
- Meet Advanced Trading's Gold Book 2011 Honorees
- The Next Rogue Algo
- Custom Algos Growing in Popularity
- Take a Tour of Bloomberg Tradebook's Trading Floor in Manhattan
- Deutsche Bank Aims New Stealth Algo At Buy-Side
- Defining the Future of Prop Trading
- HFTs: Resistance is Futile
- EMS Growing Pains
According to Morgan, long-only asset managers are working with their brokers and algo developers to figure out which strategies work best for their portfolio objectives and build those strategies into their algorithms. Buy-side firms, she says, "are working very closely with their brokers to customize their algos and learn about high-frequency trading and how they could use that to their advantage."
After rolling out its own colocated direct-market-access (DMA) platform in December, Wedbush Securities in Los Angeles assembled a team of programmers and Ph.D.s to work on custom algos. "We can go into an institution and ask, 'How do you want to trade?' " says Kevin Beadles, director of trading at the brokerage and investment management firm.
Then Wedbush developers code and simulate the algo and show the trader the results prior to putting it into effect, he explains, adding that by taking a more scientific and quantitative approach, a trading firm can replay a trading day to see if the custom algo decreased or increased a trader's performance. "We can go back and test it versus the Flash Crash day or versus any other time frame," Beadles indicates.
And in the aftermath of the May 6, 2010, Flash Crash, such back testing of algorithm performance has become an increasing priority for the buy side. Regulators' indication that a plain vanilla algo used by mutual fund company Waddell & Reed sparked the Flash Crash, for example, have alarmed the buy side. While it's commonly thought that the basic strategy performed as it was supposed to, some market participants are concerned that brokers are not testing off-the-shelf algorithms for outlier events.
"Brokers created these algos to capture customer business," says Beadles. But, "I don't think they went through the rigors of testing on a custom basis and said, 'This is how it would have performed.' "
Made to Order
Meanwhile, the big global brokers are diving into the customized algo business. "The buy side definitely wants more customized algorithms," confirms Jose Marques, managing director and global head of electronic trading at Deutsche Bank in an interview in his downtown New York office. Ivy is Editor-at-Large for Advanced Trading and Wall Street & Technology. Ivy is responsible for writing in-depth feature articles, daily blogs and news articles with a focus on automated trading in the capital markets. As an industry expert, Ivy has reported on a myriad ... View Full Bio