Kernel Regression Toolkit for Multi-Speed Trend and Crossover Analysis

Summary

This indicator provides multiple configurable Nadaraya-Watson kernel regression lines for analyzing price at different speeds. Each line can use a rational quadratic, Gaussian, periodic, or locally periodic kernel, with settings for lookback window, weighting, regression level, and crossover lag. Filtering choices include smooth, zero-lag, or no additional filter. The toolkit also supports selecting sources for derived J lines, blending two lines by a chosen weight, and plotting line direction through color changes.

The multiple lines offer a way to compare fast and slower estimates, while crossover lag settings let users tune how early a signal appears. Kernel and smoothing choices affect responsiveness and fit, so they involve a tradeoff between reacting quickly and reducing noise. The supplied material is primarily an indicator implementation and configuration guide; it does not define a complete trading system, provide entry or exit rules, or report backtest evidence. Its usefulness is therefore as an analytical tool whose settings and signals need separate evaluation for the instrument and timeframe in use.

Key ideas

• The toolkit plots several kernel regression estimates that can represent different response speeds.
• Available kernels include rational quadratic, Gaussian, periodic, and locally periodic forms.
• Lookback, weighting, regression level, filter mode, and crossover lag control the estimates' behavior.
• Users can derive lines from selected sources, combine two lines with a weighted blend, and inspect their direction.
• The document describes an indicator rather than a validated trading strategy and provides no performance evidence.

Tags

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