Dynamic RSI Length Based on Price–Momentum Correlation

Summary

This strategy varies the RSI lookback length according to the rolling correlation between price and price momentum. The length is bounded by user-selected minimum and maximum values: stronger positive correlation moves it toward the minimum, while negative correlation moves it toward the maximum. The resulting dynamic RSI generates a long signal when it crosses above the oversold threshold and a short signal when it crosses below the overbought threshold.

The script specifies example thresholds and lookback bounds, but the author says deep backtesting performance is not strong and suggests that regime filters may be needed. No performance figures or tested filter results are provided. The approach aims to reduce reliance on a fixed RSI period, yet still depends on selected bounds and threshold levels; the document presents it as an initial experiment rather than a validated strategy.

Key ideas

• The RSI lookback changes with the correlation between price and its momentum.
• Positive correlation moves the RSI length toward its minimum, while negative correlation moves it toward its maximum.
• A long signal occurs when the dynamic RSI crosses above the oversold threshold.
• A short signal occurs when the dynamic RSI crosses below the overbought threshold.
• The author reports weak deep-backtest performance and identifies regime filtering as an area for improvement.

Tags

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