Multi-DSMA: Adaptive Trend Filtering with Layered Moving Averages

Summary

The Multi Deviation Scaled Moving Average combines eight adaptive moving averages with progressively longer lookbacks. Each layer filters the close-to-close input with a Super Smoother, scales it by a rolling root-mean-square measure, and adjusts its smoothing rate according to the normalized signal magnitude. Averaging the layers blends faster response with slower trend context. A score measures how many of the seven faster layers sit above the slowest layer, and the indicator uses that score for color, opacity, and signal markers.

The guide proposes using the line and score as a trend filter, timing aid, exit cue, or multi-timeframe bias measure; it cautions against treating arrows as standalone entries. It describes score thresholds, adjustable lookback and sensitivity settings, and provides platform-specific code. No empirical performance evidence is presented. The material asserts that calculations use only current and past bars, but does not document out-of-sample tests, trading costs, or robustness across instruments and timeframes.

Key ideas

• Eight deviation-scaled moving average layers with increasing lookbacks are averaged into one trend line.
• RMS normalization makes each layer's adaptive response depend on its filtered signal relative to recent scale.
• A layer-alignment score communicates trend agreement and drives the indicator's visual display and signals.
• The guide treats crossover markers as timing aids and suggests combining them with line direction and price behavior.
• The document explains parameter effects but provides no evidence of strategy performance or robustness.

Tags

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