Efficiency and Stability of Transaction Systems Based on Simple, Exponentially and Linearly Weighted Moving Averages

Abstract

Theoretical background: Most papers are dedicated to the problem of optimizing transaction systems only for a single asset or index. In the literature there is a noticeable lack of comprehensive studies related to the entire group of assets.

Purpose of the article: Optimization of transaction systems based on the intersection of the moving average and the closing price (signal of purchase and sale) for 404 shares listed on the Warsaw Stock Exchange. For each equity, the survey covered 5,000 sessions or less if shares were traded in a shorter time horizon. The moving average types used in the study were: Simple Moving Average (SMA), Linearly Weighted Average (WMA) and Exponentially Weighed Average (EMA). In subsequent parts of the article, a ranking of moving averages was conducted and the stability of transactional systems was assessed.

Research methods: The following methods were used in the study: 1) moving averages optimizing the transaction system – correlation analysis of rates of return and of moving averages lengths, linear regression, 2) ranking of transaction system effectiveness – simple and weighted rates of return rankings, 3) analysis of transaction system stability – correlation of the first and second moving average lengths that bring the two highest rates of return, the determination factor for moving average pairs and rates of return, as well as the WF ratio (average decrease in the effectiveness of the 16 best transaction systems per unit rate of return of the best transaction system).

Main findings: The obtained results clearly indicate that for all types of averages, transaction systems were optimized in the vast majority by short-term averages, which confirms the investors’ tendency to proceed transactions with a speculative rather than investment bias. Conducted ranking of the effectiveness of three types of moving averages (WMA, SMA, EMA) unambiguously indicated that for the most part the highest rates of return were obtained for transaction systems based on WMA, before SMA and EMA. The differences in the effectiveness of trading systems based on WMA and SMA were small, but systems using these two types of moving averages proved to be much more efficient than systems based on EMA.