managing director, Asia-Pacific, SunGard's capital markets business

HOW TO USE COMPLEX EVENT PROCESSING FOR ALGORITHMIC TRADING

Posted by

Complex event processing (CEP) has been labelled as the next revolution in trading technology and is already a prominent fixture in many investment banks, hedge funds, broker/dealers and exchanges. But how justified is its revolutionary status? Is it suitable for everyone? And what exactly are the benefits that CEP provides? These were the questions that were answered in a recent presentation at SunGard’s City Day in Beijing.

CEP is ideally suited to the trading world because it is technology that allows you to deal with information that is in motion rather than at rest. One of the biggest problems facing capital markets firms is the sheer volume of data and the speed at which it is moving. And for firms operating in the high frequency trading space, the speed of execution is a constant imperative.

In such high volume, low-latency environments CEP technology comes into its own. Trading decisions are not binary choices but complex events where a number of concurrent factors have to be acted on at the same time. A CEP engine uses real-time datastreams rather than traditional static databases enabling it to process millions of events per second and for traders to make decisions on all these events in microseconds.

In the Asian capital markets the speed of trading may not yet be up to the level of microseconds but there is every possibility that this will soon be the case. The Singapore exchange is investing in a new advanced trading platform that will make it the quickest in the world and other exchanges in the region are likely to follow suit.

System set-up

A typical CEP system will have inbound and outbound adaptors to get information into the system (such as market data) and out of the system (execution decisions). The CEP engine itself will process millions of events per second and makes decisions based on the logic designed by users in the application. Some systems will also employ visualisation to help with the application development process. For example, CEP vendor Streambase has a graphical development studio to enable users to design their own applications.

The rapid development of applications and for algorithmic traders this is especially important. The shelf life of a high frequency trading algorithm is constantly reducing (some firms will modify their algorithms on a daily basis) so the quicker these algorithms can be developed and brought to market, the greater the competitive advantage for trading firms.

When you change your algos every day this makes risk management much more important on both a pre and post-trade basis. The key to pre-trade risk management in a high frequency trading world is that it has to happen in microseconds and be able to be modified according to changes in regulation and compliance requirements.

Post-trade risk management is also increasingly vital for any firm engaged in algorithmic trading and this is one area where future development of CEP technology is likely to focus. Previous end-of-day P&L reporting, trade monitoring and market surveillance is no longer acceptable in this trading environment, it must be in real-time in order to spot erroneous trades or changes in market behaviour that warrant a change in the algorithm.

The next area of focus for CEP technology is making it applicable for adoption by a wide group of users across the enterprise. For algorithmic trading firms this wide group will include not just prop traders, quantitative analysts, brokers and business analysts but also risk and compliance managers. As CEP becomes more widely deployed across the organisation, firms will also want to consider how cloud computing can be used rather than existing IT infrastructure.

5 Responses to “HOW TO USE COMPLEX EVENT PROCESSING FOR ALGORITHMIC TRADING”

Leave a Reply

  • (will not be published)