Anzhi Chen¹, Zigan Wang², Mengxin Yang^2,

We present a comprehensive analysis of the profitability of technical trading strategies that were successful within the sample period for the cryptocurrency pairs BTC/USDT and ETH/USDT. The study covers the time period from August 2017 to October 2023 and employs rigorous data snooping tests including reality checks and stepwise tests. This approach ensures that any positive results obtained are not merely coincidental, but instead reflect the intrinsic value of the method. Our results indicate that the previously profitable technical approaches, observed prior to December 2021, generally failed to generate profits during the subsequent out-of-sample period, especially after adjusting for potential data snooping. Based on the results, it is recommended to exercise caution when relying solely on historically profitable trading strategies and advisable for investors and practitioners to validate the performance of such strategies in real-time market conditions before implementing them. The findings of the study highlight the difficulty of identifying profitable technical trading strategies in an out-of-sample context when only data from the in-sample period are available, which lends support to the efficient market hypothesis within the cryptocurrency market.

technical trading, cryptocurrency