• Strategy Automation
• Coding
• API Development

A new approach to FX Alpha strategy development

By Peter Sibirzeff, CEO of Alphacet

Yet despite the influx of talent and volume into the FX market, many of the market's original members have not strayed far from their humble roots. While more established asset classes such as the equities market have gravitated towards new algorithmic trading technologies, many in the FX field have shied away from these innovations, choosing instead to rely on simple analyses often run in Microsoft Excel.

While some of this discrepancy is based on fundamental differences in how FX is traded - over-the-counter (OTC) versus exchanges - there is no arguing that FX trading is behind the times. According to the Aite Group, only 7 percent of FX trading is conducted algorithmically, compared to more than 50 percent of equities trades.

As hedge funds, CTAs and other institutions continue to enter the FX market, it is becoming clear that FX traders must evolve to remain competitive and able to react to rapid market changes affecting FX trading around the globe.

Redefining the FX Trader

The FX market is undergoing a metamorphosis. Hedge funds and CTAs are using their knowledge of quantitative trading from the equities and futures markets to drive FX trades. The current economic recession has led to unprecedented volatility and economic uncertainty. To return profits in today's volatile market, FX traders must be able to adapt fluidly to these frequently changing conditions. Long-term trend following is no longer enough.

Yet, ignoring the problem is not a permanent solution. To fix the situation, some banks are stepping up their response by increasing the use of technology and advanced quantitative methods both on the market making side and the proprietary trading side of the business.

Quantitative Toolsets for FX

With new tick capture and storage solutions that take feeds from multiple data sources, traders can develop higher frequency models that incorporate multiple types of previously untapped data including flow data from other departments in the bank. Systematic trading is becoming more important on the proprietary trading side as banks seek to reduce the overall headcount while scaling the business and having a quantifiable risk-taking approach to help bolster profits.

While quantitative FX traders once relied on large in-house proprietary systems for quantitative trading, budget cuts due to the economic downturn have made building these systems in-house no longer a viable option for virtually all firms. Fortunately, there are solutions that can serve the needs of even the most demanding trading firm, helping to greatly reduce costly IT projects that require expensive personnel as well as the risks of launching new in-house projects.

Many skills once necessary to operate these systems, such as specialized coding and programming skills, are becoming obsolete as new software solutions offer innovative interfaces and toolsets that allow FX traders to build financial models without coding or programming. Today, banks need only a few PhDs to write libraries of code that non-programming traders can share and leverage within a single platform.

And while many solutions originated in the equities market, there are now alpha generation platforms that allow traders to build a single strategy that incorporates all asset classes on multiple frequencies of data. To compete with tech-savvy hedge funds and adapt to changing market conditions, FX traders can turn to these vendors to build models that increase alpha sources and reduce time to market.

Getting technical with Alpha

Whether talking about execution algorithms, automated netting algorithms or pure signal generating algorithms for proprietary trading, the fundamental need is to analyze and manipulate FX time series to find information content that can be systematically exploited by your strategy. The ability to process and reprocess raw and derived data during exploratory research and development is a critical component of this process, as is dealing with data access issues and cleaning.

In the past, people have felt the FX market was too large to be moved by any group of people acting in unison. However, now that such a large percentage of the volume is speculative or not trade-related, this is arguable. One way to look at the market is to think of various populations of traders pursuing similar sources of alpha. With sophisticated tools, it is possible to model and exploit the behavior of these different populations of traders as the impact of their actions has an effect on short term prices.

One way to do this is to use technical indicators as input into more complex strategies to help users find trends within populations of traders, which enables users to find patterns in trading which they can then exploit to increase profitability. For example, if you run a Bollinger band on a Bollinger band signal and you find interesting signal generation possibilities, what does this mean? Really this is similar to running a mean reversion strategy on mean reversion traders. This also means that competitors with more sophisticated modeling tools can likely game your strategies over time.

The key here is the ability to go beyond academic and textbook modeling methods to easily try new ideas. While this would be difficult if it were necessary to code every idea from scratch and think about data access with each backtest, new commercial vendor solutions make it easy to test multiple ideas rapidly.

Leaders of the Pack

Understanding the need to migrate towards new technologies, many buy-side institutions are turning to pre-packaged algorithms from the sell side, which focus mainly on latency and execution. While this could indeed be a way to increase profits, once use of these algorithms reaches a critical mass, the benefits are diminished.

Using more generic toolsets that allow users to build their own strategies and algorithms ensures that returns will not be minimized due to competitors using the same method. And, if there is suspicion that a competitor is using a similar process, these toolsets provide flexibility to change strategies and ideas quickly to move towards more profitable trends in the market.

Open platforms that leverage existing code libraries in multiple languages facilitate the sharing of strategy building blocks across the enterprise, while platforms that come preloaded with extensive libraries of strategy building blocks help traders reduce strategy research and development time.

Additionally, advanced dynamic optimization capabilities through powerful machine learning combined with the ability to easily move from backtest to walk forward (paper trade) on real-time data, help traders transform ideas into profitable strategies. Platforms that enable traders to easily move from real-time walk forwards to production trading, and FIX order generation and routing further decrease time to market.

A New Era of FX trading

While today's market already looks very different from the one many traders entered just a few decades ago, the current economic crisis coupled with an influx of highly sophisticated quantitative toolsets assures us that the current FX market is changing once again.

As financial institutions struggle to discover untapped sources of alpha within an uncertain market, they must rely more heavily on new technologies that will help get banks back on track delivering effective strategies that can generate profits today, tomorrow and well into the future.