The Alpha Engine: Designing an Automated Trading Algorithm
The Alpha Engine is a trading strategy that is the result of nearly three decades of study that began from an effort to enhance economic theory and then apply it to models. Not only is the Alpha Engine a profitable system but it promotes healthy markets as it provides liquidity to financial markets.
Three hallmarks of profitable trading are laid out in the research that will be incorporated into the Alpha Engine system. First, the trading strategy should be parsimonious; it should able to be explained with the fewest possible assumptions. A strategy that is parsimonious has a limited set of variables and as a result is more adaptive to changes in market regimes. The next imperative component of the trading system is self-similarity. The strategy behaves in a similar manner across multiple time frames and thus allows for shorter time frames to be a filter of validity for longer ones. If the trading system is validated on the shorter time frame, it implies that it is also validated on the longer time frame. Furthermore, with self-similar systems, limited data is a not an obstacle to testing as the strategy can be authenticated on a wide range of events across the time frames.
The last hallmark of a profitable trading is modularity. The modelling approach for the system should be modular meaning that it can be built in a bottom up fashion. Smaller blocks can be used to build larger components and therefore create a more complex system.
The instrument universe chosen for backtest is 23 Forex pairs. The Foreign Exchange market is an over-the-counter market that is not constrained by specific exchange based rules which is beneficial to the evaluation of the statistical properties of the system. The symmetry and high liquidity also offer an ideal environment for the development of the strategy.
The paper expounds in detail upon the framework and composition of the Alpha Engine, which is a counter-trending algorithm. There are six main components of the Alpha engine including an internal time scale that dissects the price curve into directional changes and overshoots. A directional change is a change in the price trend and the trend component is the overshoot. The strategy incorporates Scaling –law distributions as well as a probability indicator that determines the trade size placed. The higher the probability of an event, the larger the trade size to be placed and vice versa. This is defined by market activity that deviates from normal behaviour.
The model had an unleveraged return of 21% for eight years for an annual Sharpe ration of 3.06. Furthermore, the max drawdown was around 0.7%. Results of the research show that using leverage of 10:1 yielded approximately 10% per year. When a time series was generated on a random walk, the Alpha Engine yielded profitable results as the model dissected Brownian motion into intrinsic time events.
Access the full research piece by Anton Golub, James B. Glattfelder, and Richard B. Olson to learn more about the framework and background of the Alpha Engine.
Access the source code of the Alpha Engine on Github.
Want to test this Alpha Engine on a demo account?
You can generate a REST API token to apply the Java code to your account:
- Register for a free practice account here.
- Log into the demo account and go to account ID in the upper right corner.
- Click token management and generate your token.
Risk Warning: The FXCM Group does not guarantee accuracy and will not accept liability for any loss or damage which arise directly or indirectly from use of or reliance on information contained within the webinars. The FXCM Group may provide general commentary which is not intended as investment advice and must not be construed as such. FX/CFD trading carries a risk of losses in excess of your deposited funds and may not be suitable for all investors. Please ensure that you fully understand the risks involved.