Delta-RSI Momentum Signals Using Polynomial Smoothing and Fit Error
Summary
This strategy derives a momentum oscillator from RSI by fitting a local polynomial to the RSI series, smoothing the signal, and measuring its time derivative as Delta-RSI. It compares that derivative with zero or an EMA signal line, and can also trade when the derivative changes direction while above or below zero. An optional normalized root-mean-square fitting-error filter excludes signals when the local fit exceeds a chosen threshold. The method is presented with adjustable RSI length, fitting window and order, and signal-line length.
The document describes the indicator logic and includes BTC-USDT futures backtest settings over a brief period, but supplies no reported performance metrics or comparison against simpler signals. It cautions that smoothing can delay entries, ranging markets can produce false signals, and parameter tuning can overfit. The mathematical procedure may also be computationally demanding, particularly for high-frequency use. Suggested extensions include volatility filters and market-regime or multi-timeframe checks; these are proposals rather than tested findings.
Key ideas
- Delta-RSI is formed by smoothing RSI with a local polynomial fit and calculating its time derivative.
- Signals may use zero crossings, crossings of an EMA signal line, or directional turns around zero.
- A normalized fitting-error threshold can filter trades when enabled.
- The method faces lag, false signals in ranges, parameter sensitivity, computational cost, and overfitting risk.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.