Profitability and Data-Snooping Tests of Four Technical Trade Strategies for Cryptocurrency Pair BTC/USDT and ETH/USDT in Cryptocurrency Markets During 2022–2023

The paper finds no profitable technical trading strategies in the examined sample and prediction periods after adjusting for data snooping.

“Our empirical results indicate the lack of profitable technical trading strategies in both the analysis sample and prediction sample periods, even after rigorous adjustments for data snooping.”

For ETH/USDT, the MACD strategy fails to generate profits in both the in-sample and out-of-sample periods.

“This optimal strategy yields an annualized return of -24.076%, a Sharpe ratio of -0.413, and a maximum drawdown of 71.128%.The strategy does not have outstanding performance in the in-sample period as its average annualized return is insignificant and can not reject the null hypothesis.”

For BTC/USDT, the MACD strategy fails to generate profits in both the in-sample and out-of-sample periods.

“The strategy fails to generate profits in both the in-sample and out-of-sample periods, indicating a lack of effectiveness in the approach.”

For ETH/USDT, the RSI strategy proves ineffective in generating profits in both periods.

“Ultimately, the strategy proves ineffective in generating profits within the organization and outside the office, suggesting a lack of effectiveness.”

For BTC/USDT, the Bollinger Bands strategy fails to generate profits in both the in-sample and out-of-sample periods.

“However, this strategy did not demonstrate exceptional performance in both the in-sample and out-of-sample periods, as its average annualized return was insignificant and it failed to reject the null hypothesis.”

The authors conclude that their results support the efficient market hypothesis for cryptocurrency markets.

“Consequently, the primary contribution of this research is its support for the efficient market hypothesis within the cryptocurrency market.”

For BTC/USDT, the RSI strategy fails to generate profits in both the in-sample and out-of-sample periods.

“Therefore, we can conclude that the RSI approach failed to generate profits in both the in-sample and out-of-sample periods, suggesting its lack of effectiveness.”

For ETH/USDT, the Bollinger Bands strategy does not generate substantial profits in either the in-sample or out-of-sample periods.

“These findings indicate a lack of”

The authors say the profitability they observe is largely driven by parameter selection rather than by market inefficiency.

“The study further emphasizes that the observed profitability of technical trading strategies is highly dependent on the selection of specific parameters rather than indicating the existence of market inefficiencies.”

The paper applies White's reality check and a stepwise test to address multiple testing and data snooping.

“We utilize the bootstrap reality check implementation, which was introduced by White (2000), to compute the p-value for the reality check”

• cites White (2000) sourceNot checked — no full text of the source

The authors divide the sample at January 2023 to reduce data-snooping bias.

“To mitigate the data-snooping bias, we employ a cutoff date of January 2023.”

The paper evaluates profitability with annualized return and Sharpe ratio.

“Table 2 presents the test results considering the whole sample period.The left and right panels are based on annualized return and Sharpe ratio as performance criteria, respectively.”

The paper also uses maximum drawdown as a risk measure in its performance tables.

“In the panel titled "Best Strategy (In-Sample)", we enumerate the depiction, annualized profit, Sharpe ratio, maximum drawdown, and the p-values of the nominal, actual inquiry assessment, and gradual inquiry assessment for the most superior approach in-sample.”

The paper states that under weak-form market efficiency, technical analysis is unnecessary because prices already reflect all past transaction records.

“In a weakform efficient market, technical analysis is unnecessary as current security prices fully reflect all past transaction records, including both valuation and volume.”

The paper investigates whether technical trading strategies can be profitable for BTC/USDT and ETH/USDT over January 2022 to June 2023.

“In this paper, we utilize comprehensive data on the cryptocurrency pair BTC/USDT and ETH/USDT, spanning from January 2022 to June 2023, to investigate the profitability of various technical trading strategies.”

The paper describes technical analysis as using historical price and volume information rather than fundamental factors.

“This methodology solely focuses on a thorough examination of historical price patterns and statistical information, disregarding the potential impact of fundamental factors that influence the asset's value.”

The paper uses stationary bootstrap resampling from Politis and Romano in its reality-check procedure.

“We resample G using the stationary bootstrap method of Politis and Romano (1994), with prespecified parameter set X, for B times, and label each resample as G b , b = 1, ⋯ , B.”

• cites Politis and Romano (1994) sourceNot checked — no full text of the source

The study uses Binance as the source of cryptocurrency trading data.

“In this study, we sourced our trading data from the highly reputable cryptocurrency data source, Binance.”

The paper notes that prior studies found evidence that some technical trading rules can be profitable in cryptocurrency markets.

“Detzel et al. (2019) provides evidence that trading strategies based on mean oscillations generate profits for Bitcoin.”

• cites Detzel et al. (2019)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

The study tests technical trading methods with different indicators and time resolutions.

“We employ different indicator selections and time resolutions to develop a range of technical trading methodologies.”

The dataset includes daily and hourly trading data for Bitcoin and Ethereum.

“Our dataset incorporates a wide range of granular information, capturing both daily and hourly trading data.”