Identification and Evaluation of Profitable Technical Trading Rules in the Cryptocurrency Market: A Mixed Method Approach

ObjectiveThe purpose of this paper is to identify the most effective technical indicators in the cryptocurrency market, as viewed by market experts, optimize their performance using optimization algorithms, and ultimately compare the performance of the selected trading rules against each other and the buy-and-hold strategy. MethodsIn this research, technical trading rules are extracted and optimized for trading two cryptocurrencies, i.e. Bitcoin and Ethereum, using a mixed qualitative-quantitative method. The efficiency of the extracted rules is evaluated and compared with the buy-and-hold strategy. In the qualitative phase, the most important technical indicators are selected and ranked based on the cryptocurrency experts’ view using the fuzzy TOPSIS technique. In the quantitative phase, the selected trading rules are implemented for a certain period and the parameters of the indicators are optimized using the grid search and particle swarm optimization (PSO) algorithm. Finally, the performance of the trading strategies selected by the experts and optimized using metaheuristic algorithms is evaluated and compared. ResultsIn the qualitative part of this study, 18 technical indicators were selected as candidate indicators in cryptocurrency trading, using the results of related studies. The selected indicators were ranked based on a survey of thirteen cryptocurrency market experts using the snowball sampling method. Finally, four technical indicators were chosen as superior indicators, showing a significant difference compared to the other technical indicators. The selected indicators were the exponential moving average, relative strength index, Ichimoku, and moving average convergence divergence. In the quantitative part of the research, the expert-based trading rules were implemented for the Bitcoin and Ethereum markets from 2018/06/30 to 2020/06/30. According to our numerical results, most of the expert-based trading strategies are more profitable than the buy-and-hold strategy. Then, the parameters of the expert-based trading rules were optimized using two metaheuristic algorithms, namely grid search and particle swarm optimization. The implementation results showed that the trading strategies optimized by these algorithms outperform both the expert-based and the buy-and-hold strategies. ConclusionBased on the experimental results of this research, exponential moving average, relative strength index, Ichimoku, and moving average convergence-divergence trading rules are commonly used and proposed by cryptocurrency market experts. Moreover, the profitability of the well-known technical trading rules could be significantly improved using the grid search and particle swarm optimization algorithms. Also, trading strategies using the combination of several technical indicators always perform better than trading strategies with a single indicator. However, all technical analysis strategies are not necessarily more profitable compared to the buy-and-hold strategy. Using inappropriate technical strategies may not only be better than the buy and hold strategy, but sometimes it causes losses in cryptocurrency trading.