Linear Regression

Ordinary Least Squares Trend Line

IndicatorTrendOverlay

Overview

Linear Regression fits an ordinary least-squares (OLS) line through the entire input series and plots the predicted value at every bar. Unlike a moving average, the line is not lagged — it is redrawn over the full visible window each time new data arrives.

In addition to the predicted values, the node outputs the slope (rate of change per bar), the intercept, and the R² coefficient of determination, which measures how well the straight line explains the price movement (0 = no fit, 1 = perfect fit).

Formula

slope = (n·ΣxY − ΣxΣY) / (n·Σx² − (Σx)²)

intercept = (ΣY − slope·Σx) / n

predicted[i] = slope · i + intercept

R² = 1 − SS_res / SS_tot

Where x = 0, 1, 2, …, n−1 (bar index within the window) and Y is the input price series. SS_res is the residual sum of squares and SS_tot is the total sum of squares about the mean.

Parameters

Linear Regression has no configurable parameters. The regression is computed over the entire visible window set by the chart's window range control.

Inputs & Outputs

Port	Direction	Type	Description
input	Input	(number \| null)[]	Any numeric series — typically close prices.
values	Output	(number \| null)[]	Predicted (fitted) values along the regression line.
timestamps	Output	number[]	Bar timestamps aligned with values.
slope	Output	number	OLS slope — price change per bar. Positive = uptrend.
intercept	Output	number	OLS intercept (predicted value at bar index 0).
r2	Output	number	R² coefficient: 0 = no linear fit, 1 = perfect linear fit.