Kernel-Smoothed Moving Average Ribbon for Adaptive Trend Following
Summary
This trend-following system builds a ribbon from five kernel-smoothed moving averages, with Beta, Gaussian, and Epanechnikov kernel options. The relative positions of the averages indicate trend direction, while configurable crossovers among neighboring averages contribute to entry signals. RSI filtering, a trend-strength measure based on an average's RSI, and a requirement for conditions to persist across consecutive bars are intended to reduce weak or noisy signals. Long and short trading can be enabled independently.
Exit controls include a hard stop, percentage or ATR-based trailing stops, and an ATR-based profit target, with the document describing an order of priority for these controls. It explains the design and possible adjustments but provides no performance results. The many settings create overfitting and operational complexity risks; crossover whipsaws, delayed reactions to reversals, and premature stops are also acknowledged. The document recommends out-of-sample evaluation and proposes regime filters, volatility adaptation, and higher-timeframe confirmation as possible extensions.
Key ideas
- Five kernel-smoothed averages form a ribbon whose alignment represents trend direction and strength.
- Configurable crossovers, RSI filters, trend-strength checks, and consecutive-bar confirmation govern entries.
- The system supports long and short trades and offers hard, trailing, and ATR-based exit controls.
- The document reports no tested performance, and the numerous parameters raise overfitting concerns.
- Ranging markets, sharp reversals, and volatile price moves can still produce losses or premature exits.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.