Nadaraya–Watson Kernel Smoothing and Its Noncausal Limits

Summary

The Nadaraya–Watson estimator, attributed here to Nadaraya and Watson in 1964, estimates values with a locally weighted average. A kernel function assigns the weights, producing a smoothed series that can help a reader visually assess a possible trend.

The document emphasizes a central limitation: the indicator is noncausal and recalculates its smoothed values as data changes. It does not extrapolate beyond the observed data. The author recommends discretionary estimation and cautions against using it to generate trading signals. No tests, performance evidence, or parameter details are provided, so the note explains the estimator’s intended interpretive role rather than establishing predictive value.

Key ideas

• The estimator uses a kernel to weight observations in a local average.
• Its smoothed values are noncausal and can recalculate as data changes.
• The described indicator adds no extrapolation beyond observed values.
• The document recommends discretionary use and warns against treating it as a signal.

Tags

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