Linear Regression Indicator

The linear regression indicator is the current value of the best-fit straight line through the last N closes. It is the endpoint, not the whole line. A linear regression channel draws the whole fit and bands around it. This indicator collapses the fit to one number per bar, which you can then treat like a moving average that follows the local slope.

If the closes are rising in a straight line, the endpoint sits near the latest close. If price hooks down, the fit tilts and the endpoint drops faster than an SMA, because the whole slope is recomputed. It still cannot see beyond the window. A straight-line assumption on a curve will lag the curve in a smooth way and then jump as the window forgets the base.

Traders use the slope of the regression value as a trend filter, and use crosses of price through it the way they use crosses of a moving average. A flat regression value is a range. A steep one is a trend. Pair it with the channel if you also want to know whether price is stretched away from the fit.

Assuming least squares means the forecast is statistically reliable. The fit is descriptive. The errors are not a clean normal sample, and the next bar is not guaranteed to respect the line.

One outlier close inside the window tilts the entire fit. A bad print or a one-off gap moves the indicator more than a median-based smoother would. Know what is inside the window.