Multi-Factor Regression Bands with RSI Reversal Signals
Summary
This Bitcoin trading system fits a rolling multi-factor linear regression using inputs such as BTC dominance, trading volume, lagged prices, and optional interaction terms. It uses the predicted price and the standard deviation of regression residuals to set dynamic upper and lower bands. A move below the lower band can trigger a long entry when RSI is oversold; a move above the upper band can trigger a short entry when RSI is overbought.
The design includes optional Z-score outlier filtering, percentage-based stop-loss and take-profit levels, ATR trailing stops, and position sizing based on ATR and a preset capital-risk percentage. The published settings describe a one-hour BTC/USDT futures backtest over roughly one month, but no performance results are provided. The document warns that historical regression relationships may fail during sharp market changes, parameter tuning matters, and the approach may fare better in ranging than trending conditions. Its claims of robustness and suitability across markets are not supported with reported evidence.
Key ideas
- A rolling linear regression combines market factors to estimate price.
- Residual volatility sets dynamic bands around the regression prediction.
- Band breaches can signal entries when confirmed by RSI extremes.
- Optional controls include outlier filtering, ATR stops, and risk-based sizing.
- The document provides backtest settings but no performance statistics.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.