Linear regression channel
A straight best-fit trend line through recent prices, with parallel lines offset above and below it (typically by standard deviation) forming a channel around the trend.
What the numbers mean
The first number is how many bars the regression line is fitted through; the second sets how far the parallel channel lines sit from it, measured in standard deviations of price around that fit. Two deviations bracket roughly 95% of the bars in the window.
- 100 bars — a common window; the whole channel is fitted to exactly these bars
- 2 standard deviations — brackets roughly 95% of the bars in the window
- 1 deviation — a tighter channel that price leaves frequently
- 3 deviations — only genuine outliers reach the edges
How it is used
Used to frame a trend and judge where price sits within it: the centre line is the statistical fair path, the outer lines mark stretched conditions, and a decisive break outside the channel suggests the trend's slope has changed.
- Centre line → the statistical fair path of the trend
- Touches of the outer lines → stretched conditions within the trend
- A decisive break outside the channel → the slope of the trend itself has changed
How it relates to everything else
The regression channel is the trend-aware version of Bollinger Bands: both use standard deviation for width, but Bollinger centres on a flat moving average while this centres on a sloping line of best fit. In a strong trend that difference matters — price walks the upper Bollinger band while sitting comfortably mid-channel here.
- Bollinger = deviation around an average · this = deviation around a trend line
- Same construction as the standard deviation channel; the names are often interchangeable
- Refits every bar, so its historical shape is not what a trader saw at the time
What to watch for
The whole channel re-fits as bars are added, so its shape in the past is not what a trader would have seen at the time. Picking the start bar changes the answer, which makes it easy to fit to a conclusion.