The CEP Conundrum

CEP is the process by which simple events—such as a single stock quote—can be correlated with other factors, such as a corresponding option price or other data over a timeframe to create a moving average.

It is these relationships that can enable traders to identify disparities between a stock and its derivatives, fair value or other instruments, and provide opportunities for arbitrage trading. Or more likely, since the volume of data required for analysis is constantly growing and changing at speeds beyond the ability of human traders, the ability to recognize these disparities and to generate trades would be coded into a firm's trading algorithm.

As two of our front page stories this week show, interest in CEP is clearly growing. In fact, the revenues of CEP providers have experienced compound growth of roughly 100 percent over the last four years, to around $49 million in 2006, according to research from Boston-based Aite Group. Furthermore, Aite expects that growth to accelerate during the next four years to potentially reach just under $1 billion by 2010.

Commerzbank has apparently bought into the CEP philosophy in a big way, recognizing that CEP engines are designed specifically to analyze individual events and streams of data and generate an action in response, using basic "if… then…" logic. But instead of being hard-coded into a proprietary application, that logic algorithm can be transposed into the workings of the CEP engine itself—engines such as Aleri's Streaming Platform or Tibco's BusinessEvents software—which has many of the basic structures already in place for creating business-specific trading rules.

For example, Aite Group's research suggests that utilizing CEP engines results in firms using only 20 percent of the resources that they would if they custom coded equivalent logic into in-house applications. And, according to Aite, the more complex the requirements, the greater the return on investment achieved.

Of course, the real returns will come from a system's potential not just to save on development resources, but to make money from trading.

While CEP software's genesis was in the monitoring space for compliance and anti-fraud measures, its ability to extract key data elements from real-time feeds and generate a response, such as a trade order, makes it particularly suited to algorithmic trading—especially because of the volumes and latencies involved. But its other strength may lie in the non-commoditized asset classes where low latency is not the only solution to gaining competitive advantage, and where there are opportunities to create revenue by identifying potential trades based on more complex analysis of data.

While firms scramble to achieve the lowest-possible latency, they are increasingly recognizing that achieving competitive advantage by speed alone gets harder and harder with every microsecond that they shave off the process, and that this position may indeed be untenable in the long term. Therefore, what is likely to drive Aite's predicted CEP revenue explosion will be firms turning to technology that can add greater value through intensive analysis, and which will differentiate firms based on the intelligence of their algorithms and their ability to identify longer-term strategies, rather than the size of their budgets and speed of their ticker plants.

Max Bowie