RSRS Timing with Skewed Scores and Price–Volume Filters

Summary

This article extends an RSRS market-timing strategy by transforming its slope-based score and adding market-state filters. It first multiplies the standardized RSRS score by the regression coefficient of determination, reducing the influence of extreme scores from poorly fitting regressions. It then multiplies that adjusted score by the RSRS slope to create a right-skewed score, intended to improve the signal’s usefulness for long entries.

For the CSI 300 over the stated one-year sample, the article reports that a lookback of 18 performed best among tested values from 10 to 30, while a larger standardization window improved returns in its experiments. It attributes differences from the cited research report partly to how early observations are standardized. Two further entry filters are proposed: compare recent 20-day moving averages to judge price trend, or require positive correlation between trading volume and the adjusted score. The article provides no detailed performance figures for these filters, and the reported backtest period is limited. Its results depend on parameter choices and standardization procedure; the suggestions require further testing.

Key ideas

• Multiplying the standardized RSRS score by regression fit quality reduces the weight of extreme scores from weak fits.
• Multiplying the adjusted score by the RSRS slope creates a right-skewed signal intended to help long-only timing.
• The article reports its best tested lookback as 18 and says larger standardization windows improved performance in its sample.
• Recent moving-average comparisons or score–volume correlation can be used as filters for entry signals.
• Backtest outcomes may change with parameter choices and with how the standardization window is initialized.

Tags

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