Using Intraday Volume-Price Correlation in a Stock Ranking Strategy

Summary

This forum post presents a Chinese A-share strategy built from the Pearson correlation between each stock’s one-minute closing prices and trading volumes over a day. It sums the daily correlation factor over a rolling five-day window, then ranks stocks by that aggregate and seeks to hold a small basket of the lowest-ranked names. The example describes daily-frequency trading over a historical period and shows how the factor is passed into a backtest engine.

The post primarily reports a backtest error: the platform cannot encode a pandas timestamp while building a module cache key. It does not provide a fix or any performance results. The correlation construction and portfolio code therefore serve as an implementation example, not evidence of an effective signal. Readers should also note that the example’s five-day aggregation, ranking direction, and portfolio handling require scrutiny, and that intraday correlations can be sensitive to market microstructure and data treatment.

Key ideas

• The proposed factor is the daily Pearson correlation between one-minute closing prices and volumes.
• The strategy aggregates the factor over five days and ranks stocks by the rolling sum.
• The example attempts to trade a small basket using the lowest-ranked names.
• The reported failure is a cache-key serialization error involving a pandas timestamp.
• The post provides no resolution or backtest performance evidence.

Tags

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.