Trading Systems Pass Aug. 24 Test
When U.S. market participants woke up on Aug. 24 to news that China shares fell almost 9% overnight, the trading day promised to be a wild ride.
It did not disappoint. The Dow Jones Industrial Average sank 7% at the open before recovering almost half of that by day’s end. Market volatility as measured by the CBOE Volatility Index touched a six-year high. Equity trading volume totaled 13.9 billion shares, more than double average levels seen in recent years and the most since 2011.
It wasn’t a record, but given the ongoing, rapid evolution in market structure and technology, so much is different from four years ago that Aug. 24 served as a stress test for current-day trading systems, spanning message traffic, order flow and execution.
Market participants, observers and technology providers say the test was passed, as it was on July 8 when the New York Stock Exchange was down for several hours, leaving other exchanges and trading venues to match the orders that would have been matched at NYSE.
Lime Brokerage, a provider of systems for quantitative, high-volume traders, saw its message traffic spike by as much as five times on Aug. 24 from its early-2015 average of 150 million to 200 million messages per day.
“We were able to process that intensity with no issues,” said Tony Huck, Lime’s president and chief operating officer. “The bottom line is that our systems did what they were supposed to do and held up very nicely. I think it probably held fairly well across The Street.”
The infrastructure of financial markets indeed held well on Aug. 24, as other than issues with the pricing of some exchange-traded funds and a less than orderly open, financial headlines were about what happened in markets, rather than what happened with markets.
Craig Viani, vice president of market structure and technology at consultancy Greenwich Associates, wrote in an Aug. 25 blog post that all market participants deserve “a nice round of applause” for their performance on the previous day. Whereas the ‘flash crash’ of May 6, 2010, was a synthetic event caused by a breakdown in the markets, the gyrations of Aug. 24, 2015 were driven by fundamentals.
“It’s amazing how only five years on from that seminal event, the speed, capacity and robustness of market centers have grown and matured to expertly handle such volume and volatility,” Viani wrote. Further, high-frequency trading firms “and (the more sophisticated) broker-dealers have built mature, dynamic routing configurations which regulate order flow based on market conditions.”
Trading technology can be likened to an offensive line in football: it’s not fully appreciated when the defense sends only three rushers (in quiet markets), but amid a fierce blitz (volatility spike), protecting the quarterback (trading alpha) is critical. System downtime is the equivalent of a sack.
“When there’s not a lot of volatility or much going on, the technology doesn’t shine as much,” said Huck. “There’s not the need for all the bandwidth we have built in to our systems. But on the very active days, customers need it, and they take notice.”
New York-Lime has built scalability into its systems, following a 10x rule of thumb, noted Suresh Thesayi, chief technology officer. The 3x-5x upsurge on Aug. 24 was thus handled easily, he said.
Prompted by business and regulatory considerations, trading-technology providers continue to raise the bar on the reliability own systems in all seasons, effectively making hay when the sun is shining. Lime executives note that systems and networks have been highly reliable, and significant resources are applied to maintain that stability and reliability.
“Upgrading our trading system is typically a yearly effort,” Thesayi said. “We also constantly look to see if there are better systems or better hardware available. If so, we proactively reach out to those vendors to get the equipment to test.”
Featured image by Chalabala/Dollar Photo Club
Algorithms have become more prevalent in the spot FX market.
QB’s Algo Suite for futures market trade execution is also being co-located to HKEX.
Breaking data silos is key to deploying automation beyond 'nuisance' orders.
They can be used on quantum hardware expected to be available in 5 to 10 years.
Streaming blocks change the basis of matching and price discovery so institutions can find new liquidity.