Shanaev Savva, Vasenin Mikhail, Stepanov Roman
University of Northumbria at Newcastle, United Kingdom, NE8 1ST, UK.
Heliyon. 2023 Mar 2;9(3):e14236. doi: 10.1016/j.heliyon.2023.e14236. eCollection 2023 Mar.
This study discovers a statistically and economically significant intraday anomaly on Bitcoin markets. Positive returns of 0.58 bps per minute are disproportionately concentrated at the turn of 15-min candles (in minutes 0, 15, 30, and 45 of each trading hour). Average returns in other trading minutes are negative. The effect is consistent across Bitcoin exchanges, in quantile regression models, and TGARCH-M estimations with heavy tails, and persist in out-of-sample tests. A high-frequency strategy that exploits this "turn-of-the-candle" effect can be net-outperforming with initial investment as low as $5,000. The anomaly is detected in the data starting from mid-to-late 2020, is potentially associated with algorithmic trading relying on the arrival of 15-min candle information, and its discovery contributes significantly to the understanding of cryptocurrency adaptive market efficiency.
本研究发现比特币市场存在具有统计意义和经济意义的日内异常现象。每分钟0.58个基点的正回报不成比例地集中在15分钟蜡烛图的转折点(每个交易小时的第0、15、30和45分钟)。其他交易分钟的平均回报为负。这种效应在比特币交易所、分位数回归模型以及具有厚尾的TGARCH-M估计中都是一致的,并且在样本外测试中持续存在。一种利用这种“蜡烛图转折点”效应的高频策略,初始投资低至5000美元时就可能实现净超额表现。该异常现象在2020年年中至年末开始的数据中被检测到,可能与依赖15分钟蜡烛图信息到来的算法交易有关,其发现对理解加密货币适应性市场效率有重大贡献。