An alternative optimization method to the exhaustive search for systematic trading strategies

Abstract

A simple way to produce an optimal set of parameters for a trading strategy is through an exhaustive optimization approach which scans the profitability of all possible values that could be used when back-testing a systematic trading strategy. The downside to this approach, as implied by the term 'exhaustive optimization', is the amount of time it takes to calculate all potential parameter combinations. This paper presents an alternative approach to avoid the problem of computational intensity created by exhaustive optimization and to test its potential to provide profitability when compared to the exhaustive approach. The study goes about this by applying both approaches to three simple trade strategies and analysing whether any improvements were made. The results demonstrate that optimizing trade rule parameters allows the trader to apply better adapted trade models to the financial market and that by using the alternative approach an improvement in the time required to perform this optimization is achieved. The concepts discussed in this study could be applied to different kinds of markets, different frequencies of data, and different types of parameters, such as the configuration of a stop loss framework for wealth managers. These trading strategies could then go on to be used as either standalone automated trading systems or as signal generators for the purpose of aiding technical analysis traders.