Strong clustering and weak barriers at round numbers in Bitcoin markets

초록

This paper examines psychological price points in Bitcoin markets using high-frequency one-minute transaction data from Binance covering more than seven years of continuous trading. Bitcoin prices cluster at round numbers substantially more frequently than expected under random price formation, exhibiting one of the strongest clustering patterns documented across traditional equity, commodity, and foreign exchange markets. Despite this pronounced clustering, round numbers display surprisingly weak barrier properties: prices spend only marginally more time above round numbers than below them, volatility in round zones barely differs from that in non-round zones, and machine learning classifiers achieve predictive accuracy indistinguishable from random guessing. This paradox - the coexistence of strong clustering with weak barriers - suggests that round numbers in Bitcoin markets function primarily as coordination points that attract order flow rather than as meaningful constraints on price movement. The pattern reflects Bitcoin's distinctive market structure, in which retail dominance amplifies behavioral clustering while continuous trading, high algorithmic participation, and the absence of fundamental anchors prevent the reinforcement of price barriers. The findings refine theoretical predictions about when behavioral biases generate exploitable inefficiencies in modern financial markets.

키워드