Performance of generated moving average strategies in natural gas futures prices at different time scales

• Moving average strategies are used on three time scales of natural gas futures.
• Genetic algorithm is used to optimize moving average methods and other parameters.
• The selection abilities of genetic algorithm differ from time scales.
• Strategies which use daily data perform better than weekly and hourly data.

When analyzing investments based on the moving average strategy in natural gas futures markets, the time scale of the data is a notable factor. Most studies on moving average investment strategies have focused on one specific time scale. In this paper, moving average performances based on data at different time scales are compared and analyzed. Weekly, daily and hourly natural gas futures prices from the New York Mercantile Exchange are used as target data. The types of strategies, the lengths of the time periods and the range parameters are coded into a binary string, and genetic algorithm is used to search for suitable lengths, appropriate calculation methods and other parameters. According to the results of three experiments using data at different time scales, the performances differ in type of moving average strategy selection and adoption of range parameter. Generally, experiments based on daily scale data show better performance than weekly and hourly scales in rate of return, return times and return stability. The results from this study could help in choosing data when using moving average strategy in natural gas futures markets.