Big Data For Intelligent Trading?

It's no secret that the equities markets have seen better days. Volumes continue to sag from the record-setting days of 2008 and 2009 and few think the volumes will ever return. Along with the depressed volume come decreased profits, which have changed the way brokerages, high-frequency trading (HFT) firms and exchanges operate.

In fact, some HFT shops have seen profits drop as much as 75% in recent years, forcing many to change their strategy or exit the market completely, said Derek Keene, president at iSys Capital Technologies in an event focusing on the intersection of low latency technology and big data in the capital markets. iSys is a capital markets technology infrastructure consultancy that specializes in front office and trading technologies.

The topic of big data in the trading business has often been met with jeers or snickers, since HFT players rely on microsecond latency and utilizing big data usually meant increasing processing time outside of acceptable metrics. But, as Jia Chen, a systems engineer for Arista Networks who spends most of his time with capital markets clients, said, "The race to zero is over. There is no more room to compete just on speed." Instead, analytics technology has advanced sufficiently to allow traders to manipulate data to find gain other advantages that could differentiate their strategy (instead of simply competing on low latency).

The need to find newer ways to compete has been evident for a while, says Davor Frank, Senior Solutions Architect at Solarflare. Prior to joining Solarflare, Frank spent a number of years at a HFT firm developing technology to compete in the low-latency arms race. Even five years ago, Frank says, pure speed would only get a firm so far.

According to other speakers at the iSys event, data analytics technologies have advanced sufficiently to provide millisecond latency on large data sets, if configured correctly. While HFT players demand microsecond or even nanosecond latency, traders can create strategies that include other types of data and information that may give them a competitive advantage.

For instance, by using some big data technologies, coupled with flash memory to improve performance, a firm could develop a strategy that includes weather data, social data or geolocation in real time. In fact, some financial services firms are already starting to experiment with advanced analytics, coupled with low latency technology, to develop smarter or intelligent trading decisions, according to the panel at the event.

Speakers on the panel also included Brian Freed, VP of Strategy at SGI; Vamsi Chemitiganti, Chief Architect for Red Hat's Financial Services Practice; and John Zamites, IBM Flash System Specialist.

But with competition based on pure speed being "yesterday's news," how long will it really be until extreme data analytics can produce a usable trading strategy? It may be here sooner than you think, as many firms have already started to look at unstructured data in new ways to help develop intelligent trading strategies. Stay tuned. Greg MacSweeney is editorial director of InformationWeek Financial Services, whose brands include Wall Street & Technology, Bank Systems & Technology, Advanced Trading, and Insurance & Technology. View Full Bio