Nadaraya–Watson Smoothing for a Weighted Multi-Signal Trading Strategy

Summary

This strategy smooths four market inputs with Nadaraya–Watson kernel regression, then combines them using preset weights: closing price, RSI, ATR relative to price, and price deviation from a moving average. A crossover between the integrated signal and its moving average generates a raw trade signal. A long-term price trend filter and an RSI threshold must also agree before the system enters a position. Percentage-based stop-loss and take-profit levels are specified for both long and short positions.

The document explains the architecture and parameter choices but provides no performance evidence. Its published backtest settings cover ETH/USDT on one-minute data over a two-day period, which is insufficient to establish robustness across conditions. The description cautions that parameter tuning can overfit, multiple filters can suppress valid signals, and kernel calculations may affect real-time performance. It recommends out-of-sample validation and periodic review of weights; the named input dimensions and preset weights are design choices, not evidence that they predict market direction.

Key ideas

• The system smooths four inputs—price, RSI, relative ATR, and price deviation from a moving average—using a kernel regression method.
• It combines the smoothed inputs with preset weights and compares the result with its moving average.
• Long and short entries require both a signal crossover and agreement from trend and RSI filters.
• Percentage stop-loss and take-profit levels are provided for both trade directions.
• The short published backtest window does not establish robustness, and parameter tuning risks overfitting.

Tags

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