Linear Regression

The Linear Regression indicator fits a least-squares straight line through the last N closing prices and returns the endpoint value of that line. It projects where the trend "should" be based on the statistical best fit, effectively acting as a forward-looking moving average.

LinReg(N) = a + b × N
where b = (N × Sum(i×Close_i) - Sum(i) × Sum(Close_i)) / (N × Sum(i^2) - Sum(i)^2)
      a = (Sum(Close_i) - b × Sum(i)) / N

The indicator outputs the regression value at the current bar (the right edge of the window).

• Predictive element — unlike MA which averages past data, LinReg extends the best-fit line to the current point.
• Less lag — tends to lead price during steady trends because the slope projects forward.
• Slope = trend strength — steep positive slope means strong uptrend, steep negative means strong downtrend.

Trend direction

• LinReg rising — uptrend.
• LinReg falling — downtrend.

Price vs. LinReg

• Price above LinReg — price is outperforming the statistical trend (bullish momentum).
• Price below LinReg — price is underperforming (bearish momentum or reversion expected).

| Parameter | Default | Description | |-----------|---------|-------------| | Period | 14 | Number of bars for the regression calculation |

| Condition | Meaning | |-----------|---------| | close > LINEARREG(14) | Price above regression line — bullish | | close cross_over LINEARREG(20) | Price reclaims regression — potential reversal | | LINEARREG(14) > LINEARREG(50) | Short-term trend steeper than long-term — momentum building |

The slope of the regression line (coefficient b) tells you the trend speed:

| Slope | Interpretation | |-------|---------------| | Steep positive | Strong uptrend — price gaining fast | | Gentle positive | Moderate uptrend — steady climb | | Near zero | No trend — consolidation or range | | Gentle negative | Moderate downtrend — gradual decline | | Steep negative | Strong downtrend — price falling fast |

A change in slope direction often precedes a change in price direction, making LinReg a leading signal.

Combine LinReg with standard deviation bands to create a regression channel:

• Upper band = LinReg + 2 × StdDev(N) — overbought zone
• Lower band = LinReg - 2 × StdDev(N) — oversold zone
• Enter long when price touches lower band and slope is positive
• Enter short when price touches upper band and slope is negative

This creates a powerful mean-reversion setup that adapts to the current trend direction.

• LinReg works well for mean-reversion strategies — if price deviates far from LinReg, expect a pullback.
• Combine with standard deviation bands around LinReg for a regression channel.
• Short periods (5-14) are popular for intraday; longer (20-50) for swing.
• LinReg is especially useful on assets with steady trends (indices, blue-chip stocks).