Floris Laly

Type of the article: Research Article

Abstract
A high-frequency enough dataset may be able to identify very short-term opportunities that can potentially generate significant profits. The present study aims to test the performance of momentum and contrarian trading strategies in a high-frequency setting. To this end, tick-by-tick data from Refinitiv (TRTH) are employed, and 288 strategies are tested across a sample of 28 highly liquid U.S. blue chip stocks during the period 2020–2021, a timeframe marked by significant market volatility, notably the COVID-19-induced crash of February–March 2020 and the subsequent recovery phase from April 2020 to December 2021. The analysis reveals that, although certain strategies exhibit superior performance relative to the benchmark strategy both prior to and following adjustments for risk, data snooping, and luck, none outperform the benchmark strategy once reasonable transaction costs are accounted for. These findings suggest that sophisticated retail traders may be unable to exploit underreaction or overreaction in stock prices at high frequency. The main contribution of this study lies in the application of stringent robustness checks, including a robustness check based on luck within the context of intraday trading strategies.