Nadaraya-Watson Envelope

Nadaraya-Watson Envelope

This indicator draws a smooth, adaptive envelope around price using Nadaraya-Watson kernel regression. A Gaussian-weighted average of recent closes forms the center line, and an upper and lower band are placed at a volatility-based distance from it.

How it works:

Recent bars are weighted by a Gaussian curve, so the newest bars count the most. This gives a smooth estimate of fair value.
The band width is the average distance between price and the estimate, multiplied by your chosen factor.
A green arrow (below the bar) appears when price crosses under the lower band, which suggests a possible oversold bounce.
A red arrow (above the bar) appears when price crosses over the upper band, which suggests a possible overbought pullback.

Settings:

Bandwidth: lower values follow price closely, higher values give a smoother line.
Multiplier: controls how wide the bands are.
Kernel Window: how many bars are used in the calculation.
Envelope Width Length: the lookback used to measure band distance.
Colors: choose your own band and signal colors.

How to use it:

It suits ranging or mean-reverting markets best. In strong trends, price can stay outside the bands for a long time.
Confirm signals with trend, support/resistance, or volume.

Note: This version uses the non-repainting method, so plotted values don’t change after the bar closes.