Linear regression R-squared is the charting application of R-squared, the statistic that measures how much of the variation in a set of data is explained by a fitted line. As a technical indicator it is calculated on a rolling window: a straight line is fitted to the closing prices of the last bars, and the indicator plots the share of price variation that the line accounts for, from 0 to 1 or 0 to 100 percent.
A reading near 1 means prices have moved along an almost straight path, which is taken as evidence of a strong, orderly trend. A reading near 0 means prices have been scattered with no clear direction. The indicator says nothing about whether the trend is up or down, so it is paired with the linear regression slope, which gives the direction. A common approach treats a slope signal as valid only when R-squared exceeds a level judged significant for the window length.
Traders also watch for R-squared falling from high levels as a sign that a trend is losing coherence. Its role resembles that of the Average Directional Index, which also measures trend strength without direction. The indicator's value at each bar is the end result of the same fit used by the linear regression indicator.