The 10 Imperatives of Next-Generation Algorithmic Trading

Dr. John Bates and I recently wrote an article called "The 10 Imperatives of Next-Generation Algorithmic Trading."  The article was well received when it was published last month, and, based on the interest in this blog around "The 10 Myths of EDA and SOA," we thought we would expand upon and blog on algorithmic trading as well.   We're not simply re-posting the article, the internet provides the advantage of hyperlinking some of the ideas and concepts to resources on the web, and, of course, interactivity and linkage to other blogs in the blogosphere.  So let's get started:

Algorithmic trading has been one the most discussed topics within the financial industry over the past 12 months. In today’s hyper-competitive trading world, financial institutions feel the mounting need for technology that aids their unique trading style. The continually shifting landscape means both buy-side and sell-side firms need to adapt to the effects of change. Sell-side institutions are exploring ways to augment the talents of their traders and optimise their client services, while buy-side firms are persistent in their endeavour to control their trading strategies and to hide them from the competition. Algorithmic trading has played a significant part in this.

During 2006, algorithmic trading has entered the mainstream (read the AITE group's report on event processing in capital markets). Algorithmic techniques and the technology that powers them are now highly influential in the way that financial instruments, both in exchange and OTC markets, are traded. Algorithmic trading was initially used in equities; however other asset classes, including futures, options and foreign exchange (FX) have begun to catch up quickly.

Algorithmic Trading Explained

Initially algorithmic trading was defined as any type of computer-assisted trading activity which handles the timing, submission and management of orders. However, recently the term has expanded as a ‘catch-all’, to encompass other terms that describe computer-assisted trading, including ‘program trading’, ‘auto trading’, ‘black box trading’ and ‘high-frequency trading’, across single or multiple pools of liquidity. As the power and flexibility of the new generation of algorithmic systems becomes more widely accepted, these terms are tending to fall under one banner.

There are two elements of an algorithmic trading strategy: the decision of when to trade, or pre-trade analytics, and the decisions of how to trade, or the execution phase of the algorithm.

The decision of when to trade is based on continuously re-calculated analytics and monitored thresholds. This could include, for example, a moving average crossover algorithm that calculates two moving averages, and analyses, in real time, when they cross one another. It then buys or makes the decision to buy or sell, depending on which average is higher.

The decision of how to trade, or the order execution element of the algorithm, can be just as complex as the decision of when to trade. For example, once an opportunity is identified by the pre-trade analytic to buy, for example, 10,000 shares of IBM, the order execution element of an algorithmic trading strategy may slice the order up into smaller parts (blocks of 1,000 shares). In conjunction, it may place the order in multiple liquidity pools to take advantage of the prices and availability of liquidity across a ‘virtual’ exchange with multiple participants (OTC markets).

The Approaches to Algorithmic Trading:  Custom Built, 'Black Box' and 'White Box'

There are three common approaches to implementing algorithmic trading: custom built, ‘black box’ and ‘white box’. Firms building algorithmic strategies on their own, in Java or C++, have the advantage of a completely custom built solution. In doing so, they optimize their control over the behavior of the algorithms and the way they are integrated into their technical infrastructure. Unfortunately, bespoke algorithmic trading systems are very expensive and time-consuming to build. Furthermore, custom-built trading systems require developers to ‘re-invent the wheel’, as they are building from scratch algorithms and infrastructure that have already been built in the past. Although many algorithmic variants are very similar – VWAP algorithms, for example, are available already in hundreds of forms – slight differences can offer competitive advantage.

Another approach to algorithmic trading is to use algorithms provided by a broker or application provider: the black box technique. This is the simplest and fastest way to begin algorithmic trading. However, since firms cannot see how the algorithms work, they cannot add their own ‘secret sauce’, so the trader loses control over their trading strategies in the market. Using a commoditised black box approach to algorithmic trading, it is often difficult to beat the market – if you’re using the same algorithms as your competitor, it is difficult to gain a significant advantage. Also, many black box strategies are ‘fire and forget’, in that a user fills in the parameters of the strategy, initiates the strategy and then simply waits for the results.

The final approach is the ‘white box’ technique – an open, customizable, flexible platform with pre-built trading analytics that come with source code. A white box approach lets the user customize algorithms; combine them at will; and control how the platform is integrated with existing market feeds, OMS and EMS applications. This approach offers the best of both the custom built and black box model – fast time-to-market and flexibility, in one platform. A white box approach is not ‘fire and forget’. It usually provides a flexible and customizable front-end, enabling users to see the progress of a trading strategy on a real-time graphical dashboard and allowing them to intervene to adjust key parameters in-flight. Users can customize the look and feel of this front end to their own trading style. They can also maintain a high level view of all their algorithms as they execute. In this way, a trader becomes the coordinator of a set of strategies, rather than the mechanism to either manually execute the strategies, or just initiate them.

On with the "10 Imperatives of Algorithmic Trading..."

So with a basic introduction to algorithmic trading in place, and the role of white-box and black-box approaches explained, start reading the first imperative of algorithmic trading:  Move First.