Multi-Anchored Regression Channels with Flexible Fits and Signal Filters

Summary

This indicator plots linear regression channels and bands using configurable lookbacks and anchor events. Anchors can be based on fixed lengths, price or volume extremes, regression correlation or slope, technical indicator extremes, calendar periods, market sessions, corporate events, or external data. The user can choose among standard, ridge, lasso, elastic net, Huber, least absolute deviation, and Bayesian regression options, as well as several deviation calculations and weighting approaches.

It also supports signal filters based on price position within the channel and regression correlation, with optional chart labels, alerts, tables, and exported signal values. The settings allow a sampling step to reduce loading time, while warning that this can reduce anchor accuracy. The script describes configuration choices and intended interpretations, but the supplied material gives no performance results or validation of trading signals. Results depend on anchor selection, input settings, data source, and chart scale; logarithmic mode is intended to be matched with a logarithmic chart.

Key ideas

• The indicator can anchor regression lookbacks to price events, indicators, time periods, sessions, or external series.
• Several regression estimators and deviation formulas are available to change the fitted trend and channel width.
• Signal criteria can combine price position relative to a channel with correlation thresholds.
• Increasing the anchor search step speeds loading but may reduce accuracy.
• The document describes features and settings but provides no evidence of trading performance.

Tags

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