- The Book Structure
- Strategies Considered in This Book
- Scientific and Empirical Approaches to Developing Automated Trading Strategies
- Rational Approach to Developing Automated Trading Strategies
Rational Approach to Developing Automated Trading Strategies
Most developers of automated trading strategies combine scientific methods and empirical approaches. On the one hand, strategies resulting from such combining are based on strong economic grounds. On the other hand, they benefit from the numerous advantages of optimization and from the impressive progress in computer intelligence. We will call this the rational approach.
Under the rational approach a set of rules, determining the general structure of a trading strategy, is formed at the initial stage of strategy creation. These rules are based on the prior knowledge and assumptions about market behavior. The results of statistical research, either received by the strategy developer or obtained from scientific publications and private sources, can also be used to shape the general framework. Obviously, patterns established during such research introduce certain logic into the strategy under development. At the same time, statistical research may result in the discovery of inexplicable relationships lacking any economic sense behind them. Such relationships should be treated with special care since they may either be random in nature or result from data mining.
The initial stage of strategy creation is based mainly on the elements of scientific approach. At this stage the following must be determined:
- Principles of generating the signals for opening and closing trading positions
- Indicators used to generate open and close signals
- A universe of investment assets that are both available and suitable for trading
- Requirements to the portfolio and restrictions imposed on it
- Capital allocation among different portfolio elements
- Methods and instruments of risk management
At the next stage of developing a trading strategy, the rules laid down on the basis of scientific approach are formalized in the form of computable algorithms. This stage is congested with elements of the empirical approach. These are the essential steps:
- Defining specific parameters. All rules formulated on the basis of the scientific approach should be formalized using a certain number of parameters.
- Specifying the algorithms for parameter calculation. Different algorithms may be invented for calculating the same parameter.
- Establishing the procedures for the selection of parameter values. This requires adopting a certain optimization technique.
Usually the decision on the number of parameters and selection of methods for their optimization does not depend on the economic considerations of the developer, but follows from the specific requirements to the strategy and from its technical constraints. These requirements and constraints are developed with regard to the reliability, stability, and other strategy features, among which the capability to avoid the overfitting is one of the most important properties.
In this book we will follow the principles of rational approach to the creation of trading strategies. The main task of the developer is to combine methods attributed to the scientific and the empirical approaches in a reasonable and balanced manner. In order to accomplish this task successfully, all basic components of a trading strategy should be clearly identified as belonging either to components that are set on the basis of well-founded reasoning or to an alternative category of components that are formed primarily by applying various optimization methods.