Composite Standard Error Bands Around a Linear Regression Curve

Summary

The document explains a composite band indicator built around a linear regression curve. Unlike Bollinger Bands, which are commonly centered on a moving average and scaled by price variability, these bands use estimates of regression error. When prices track the fitted line closely, the bands narrow; when residual variation grows, they widen. The indicator plots both narrow and wide upper and lower bands, using a multiplier to set their distance from the regression curve.

It also describes an alpha-beta calculation that estimates the regression slope and intercept, and references an additional standard-error approach intended to provide support and resistance levels. The supplied implementation uses a 21-period window, a first-degree regression, and a multiplier of two. The post explains construction, not trading rules or predictive performance: it gives no entry, exit, or risk-management method and presents no backtest evidence. Band behavior depends on the chosen settings and does not itself demonstrate a profitable signal.

Key ideas

• The bands are centered on a linear regression curve rather than a conventional moving average.
• Their widths reflect estimated regression error and expand as prices deviate more from the fitted path.
• The indicator plots narrow and wide bands above and below the regression line.
• The document supplies construction parameters but no trading rules or performance evidence.

Tags

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