Automatic Trading System Design

Introduction

The futures trading became electronic business when widespread use of internet technologies occurred. There exist many electronic markets since 90’s and at present day they are equal in both size (quantity) and importance to the traditional trading. When the electronic trading became possible, the computing power of computers allowed many tasks to be automated. It is possible to automatically process market data and to perform the analysis very quickly, automating part of the decision-making process and leaving only small part of the work to be handled by trader.

From the point of view of electronic trading, it is trading of stocks that are in the centre of interest of scientific community. There are several differences between stock trading and futures trading. Discussion about the differences can be found in e.g. (Williams, 1979, pp. 17-20). But there is much in common for both types of trading and we will not be strictly separating stock trading from commodities (futures) trading. In the following text whenever the notion “trading” is mentioned, it is a futures trading that we have in mind. Many (not all) indicators useful for the futures markets are useful for trading stocks as well and vice versa. The principle is usually similar.

This chapter is focused on lifecycle of design of automatic trading system. Its purpose is to present and summarize rules of design, important notes on the topic and discussion of typical issues. No specific automatic trading system will be presented in this text. However, as we hope, it may bring some interesting insights into this problem area.

The investment decision-making process consists basically of two parts. First part of this process is fundamental analysis, second part is technical analysis. For the purpose of automatic trading it is a technical analysis that is more important for us. But best outcome is archived when both parts takes place in the decision-making process. Because this matter will be discussed in the following text for the moment it could be said that fundamental analysis represents the “what”, while technical analysis is the “when”. The concept of fundamental analysis in commodity trading is covered in (Dunsby, 2008). Dunsby (2008) is focused on selected categories of commodities only – energies, grains and oilseeds, livestock, industrial materials, softs. However, this publication gives reader the right perspective of fundamental analysis importance and application on real-life situations.

The cycle of steps of system design is usually done during the mentioned investment decision-making process (IDMP), allowing some part of this process to be automated. It is not necessary to have automated the whole IDMP. An automatic trading system (ATS) may cover only the technical analysis part and provide the best outcome of the applied indicators to the user in order to allow him to make the final decision. It is also possible to leave whole process to the ATS itself, but the former solution is usually preferred by traders.

In the following text, we would refer to trader as to “him”, instead of “him or her”, as it would be more appropriate. By this, we do not mean any disrespect to female traders. This simplification is used in order to make reading easier.