October 14, 2013

NASDAQ OMX Group and Thesys Technologies have partnered to offer a trading algorithm test facility, a testing environment that will simulate trading on all major US equity exchanges. It is expected to launch in Q1 2014.

Because a significant portion of trading volume is generated algorithmically, testing is key to maintain industry compliance. The Algo Test Facility will enable users to rigorously examine algorithmic strategies by replaying and interacting with real historical market data, according to the press release. The facility will also provide reports that show the quality of the quote processing, latency, all the way down to the order handling.

Stacie Swanstrom, head of access services at Nasdaq OMX explains the Algo Test Facility is intended to provide traders with valuable insights about risk and performance, including the ability to evaluate the cost of adverse selection for specific trading strategies. The reports from simulations will help firms decipher if things worked the way they expected, and there will be criteria in the reports to help them examine the output.

Economic Value

Thesis Tech, the infrastructure affiliate of Tradeworx, is the power behind the testing facility (also known for powering SEC Midus software). Manoj Narang, CEO of Tradeworx says, "We think there are also valuable economic insights to gain by simulating strategy before launching them."

For example, the people who build strategy also build parameters. With the test facility, users can test multiple parameters to see which ones are best in different environments.

"It costs money to be faster, and you can't possibly make decisions about that tradeoff. We help people quantify that for the first time... This is the first tool we're aware of that allows you to measure economic impact of latency on trading systems. You'll be able to see how much better or worse you would have of done if you were 100 microseconds faster or slower."

[For more on Nasdaq see: Why StockTouch Moved Gamer Mechanics onto Nasdaq's FinQloud]

Essentially, adds Swanstrom, it’s a cost effective solution for broker dealers to manage risk exposures, and more importantly to understand economics around new trading strategies. "When a market participant, whether they be from the sell-side or the buy-side, is gearing up to implement new automated trading strategies, they can now simulate real market situations on a specific day to measure the financial impact. This tool enables market participants to make necessary adjustments before going live."

Risk & Compliance

Broker dealers are struggling to get ahead of regulatory mandates, and Nasdaq admits there have been challenges over last couple years. "Regulatory bodies are looking closely at operational risk management in a complex environment where technical difficulties are all part of the market structure weather,"" says Swanstrom. "So the financial services community is trying to be proactive and concentrate on alpha generation… There is a business incentive to perform tests for the protection of your firm and for the protection of the larger market."

"We are pretty sure it's impossible to build testing systems that can guarantee stability of trading systems, the best you can do is minimize errors, and that's a worthwhile goal," observes Narang. "Glitches in the marketplace occur for various reasons. Regulation covers certain critical systems that exchanges, clearinghouses and some broker-dealers operate, but there's little in the way of actual substance that comes to testing of algorithms, which have been somewhat destabilizing in the past." Knight Capital and the flash crash were both caused by rogue algorithms. "While I think it's impossible to completely reduce risk, you can ask market participants to do their best to cut the risk. That's what this is intended to do."

Infrastructure

According to Narang, the principal technology challenge was adapting the systems to speak FIX protocol and native protocol of all exchanges. Nasdaq's was the first functionality adopted.

NASDAQ is allowing market participants’ trading systems already co-located at its data center in Carteret, New Jersey to simulate their algorithms in the real market environment, interacting with historical data from the entire marketplace.

The Algo Test Facility is a stand alone product, not packaging with other NASDAQ services or test services. According to the press release, the first phase of the Algo Test Facility will include simulations for the NASDAQ Stock Market, and over time the test system plans to support all major U.S. equities exchanges along with their order types and protocols so user activity may interact with the historical orders of other market participants.

ABOUT THE AUTHOR
Before joining UBM Becca served as Editor-In-Chief at Kapitall Wire, the sister site to an online brokerage firm that combines the challenges of the stock market with the thrill of ...