Regression Trend

The regression trend channel applies ordinary least squares (OLS) regression to price data over a given lookback period. The center line is the linear regression line — the best-fit straight line through the data that minimizes the sum of squared residuals (the vertical distances from each price bar to the line). Upper and lower bands are plotted at fixed standard-deviation distances (default ±1σ) from the center line, creating a statistically anchored volatility envelope.

This tool was popularized by John Bollinger in his volatility framework (though Bollinger Bands center on a moving average, not a regression line) and further refined for trend analysis by Timothy Hayes. The key advantage over subjective trend lines isobjectivity: every trader using the same lookback period and data sees the identical regression line and bands. There is no room for drawing bias or wishful thinking.

The regression channel also provides the R-squared statistic, which quantifies how well the linear trend fits the data on a scale from 0 (no linear relationship) to 1 (perfect linear relationship). An R-squared above 0.8 indicates a strong, reliable linear trend. Below 0.3 suggests the trend is weak, non-linear, or the market is ranging. This statistical context is one of the most valuable features of the tool.

The Regression Trend tool requires 2 anchor points (minPoints: 2). Click at the lookback start point, then drag to the current bar or the end of the trend being analyzed. The regression window includes all price data from the start point to the endpoint. The regression line and standard deviation bands are calculated and rendered instantly using OLS.

After placement, drag either endpoint to adjust the lookback period. The regression recalculates automatically for every pixel of movement, with the bands adjusting in real time. Drag the center regression line body to translate the entire study (preserving lookback length and slope). In multi-select mode, multiple regression channels move together.

On hover, the regression line, bands, and endpoints all highlight. The R-squared value can be displayed as a floating label on the chart. A “follow price” mode automatically extends the end point as new bars form, making the regression channel a live, updating tool that evolves with the market.

The regression trend channel offers flexible styling for the center line, bands, and fill, plus statistical display options.

Mean Reversion at the Bands: When price touches the upper band, it is statistically overextended from the trend — a pullback toward the regression line is probabilistically favored. Similarly for the lower band. Enter counter-trend at the bands with a target at the center line and a stop beyond the band (band + 0.5σ). This works best when R-squared is moderate (0.3–0.7) and the market is cycling within the channel.

Trend Continuation at Center Line: In a strong trend (steep slope, R-squared above 0.8), use the regression line as a dynamic entry on pullbacks. Enter when price touches the regression line and resumes in the trend direction. The slope confirms the trend direction; the bands serve as trailing profit targets. This is the most reliable strategy in trending markets.

Statistical Extreme Signals: A close outside the ±2σ band is a rare event (approximately 5% probability under normal distribution assumptions) and often precedes a sharp reversal or acceleration. Look for exhaustion candles (doji, long wicks, bearish/bullish engulfing) at ±2σ levels. A ±2σ touch combined with a momentum divergence (RSI, MACD) is a powerful counter-trend setup.

1. 1
   Select the lookback start

   Click at or before the start of the trend you want to analyze. The regression window begins here.
2. 2
   Set the end point

   Drag to the current bar or the end of the trend you are analyzing. The regression line and bands update instantly.
3. 3
   Adjust standard deviations

   Change the σ multiplier in the settings. Default ±1σ contains ~68% of data. ±2σ contains ~95%.
4. 4
   Check R-squared

   Enable R-squared display. Above 0.8 confirms a strong linear trend. Below 0.3 suggests weak trend or range.
5. 5
   Evaluate the slope

   The regression line slope shows the average price change per bar. Compare to the visual trend steepness.

The regression channel gives you three independent pieces of information: (1) the trend direction and strength from the regression line slope and R-squared, (2) the current overextension or underextension from the band touches, and (3) the statistical probability of price remaining within the channel.

• Price at the upper band suggests statistical overextension — a pullback toward the center line is favored.
• Price at the lower band similarly signals a potential bounce back toward the regression line.
• Price walking the band (hugging the upper band for multiple bars) = strong trend with accelerating momentum.
• A flattening regression line indicates trend deceleration — the linear trend is losing conviction.
• R-squared above 0.80 = strong, reliable linear trend. Below 0.30 = weak or non-existent trend.
• A widening band over time indicates increasing volatility. A narrowing band indicates decreasing volatility.

Multi-Lookback Stacking: Overlay three regression channels with different lookbacks — for example, 20 bars (short-term), 50 bars (medium-term), and 200 bars (long-term). When all three regression lines slope in the same direction, the trend is confirmed across multiple horizons. When they diverge, a trend change may be imminent.

Regression Slope Divergence Analysis: Compare the regression slope angle with the angle of the most recent price swing. If the recent swing is steeper than the regression line, momentum is accelerating beyond the mean trend — expect reversion toward the regression line. If the recent swing is shallower, the trend is decaying — expect the regression line to flatten further.

σ Band as Volatility Stop: In a trend-following strategy, use the opposite σ band as a trailing stop. For a long position in an uptrend, trail the stop at the lower (−1σ) band. If price closes below −1σ, the trend has deviated significantly and the position should be closed for a statistical violation.

• 1.Linear Regression & Standard Deviation Channel — StockCharts ChartSchool
• 2.Technical Analysis of Financial Markets — John J. Murphy
• 3.Bollinger on Bollinger Bands — John Bollinger
• 4.Linear Regression Forecasting — TradingView Help
• 5.Statistics for Traders — Timothy Hayes
• 6.Understanding R-Squared in Trading — Investopedia