Intraday mean reversion is most reliably signaled by order imbalance, effective bid-ask spread dynamics, and short-term liquidity shocks. Empirical market-microstructure research links short-horizon price reversal to the mechanical effects of trade flow and liquidity provision rather than to slow information diffusion, so these anomalies tend to show the strongest, most repeatable predictive power for intraday reversal.
Order flow and trade classification
Work by David Easley and Maureen O'Hara at Cornell University demonstrates that persistent order imbalance—a dominance of buyer-initiated or seller-initiated trades—contains information about subsequent returns. When imbalances reflect temporary liquidity demand rather than new fundamental information, prices typically overshoot and then revert as liquidity providers absorb the flow and quotes resume equilibrium. Practitioners who classify trades by aggressor side and measure signed volume often capture this short-horizon predictive signal.
Bid-ask bounce and liquidity shocks
Richard Roll at UCLA Anderson showed that the effective bid-ask spread creates negative autocorrelation at very short intervals: part of observed intraday reversal is a mechanical "bounce" as trades alternate across the spread. Tarun Chordia at Emory University and collaborators link broader liquidity shocks and commonality in liquidity to transient return patterns; when liquidity dries up, prices move more on trade pressure and then partially reverse as liquidity returns. These mechanisms are distinct but can coincide, strengthening the mean-reversion signal.
Market structure and execution environment shape the magnitude and reliability of these anomalies. In highly liquid US large-cap venues, electronic market making and high-frequency liquidity provision can make order-flow and spread effects short-lived but consistently exploitable at sub-hour horizons. In less liquid or emerging markets, illiquidity and fragmentation amplify reversals but also raise transaction costs and execution risk, so statistical predictability may not translate into implementable profits.
The practical consequences include predictable short-term alpha for strategies that account for execution costs, and regulatory implications because aggressive flow can transiently impair price discovery. Traders and risk managers should therefore treat order imbalance, bid-ask spread dynamics, and liquidity shocks as the most dependable intraday mean-reversion indicators, combining them with robust trade classification and venue-aware cost modeling to distinguish true predictive structure from mere microstructural noise.