Direct Answer
A regression residual is the difference between an observed value and the value fitted by a regression line. It isolates the part of a series that the explanatory variable does not account for. A positive residual means the actual value is above the fitted line and a negative one below.
Regression Residual: Formula, Meaning, Signals, Examples and How to Use It
How Is Regression Residual Calculated?
A regression residual is the difference between an observed value and the value fitted by a regression line. It isolates the part of a series that the explanatory variable does not account for.
Formula: residual = y - (a + b * x), where a is the intercept and b the slope of the fitted line
Inputs:
- Dependent series y
- Explanatory series x
- Fitted intercept and slope
Worked example. With a fitted line of 10 plus 0.8 times x and observations y of 52 and x of 50, the residual is 2.00. The inputs are illustrative numbers, not market data.
Reading it. A positive residual means the actual value is above the fitted line and a negative one below. Residuals scattered around zero with no pattern suggest the line captures the relationship, and persistent runs on one side suggest it does not.
Limits. Its scale depends on the units of y, so it needs standardizing to compare across series. It also depends on the fit window, and the latest point has influence on the line it is measured against.
What is Regression Residual?
A regression residual is the difference between an observed value and the value fitted by a regression line. It isolates the part of a series that the explanatory variable does not account for.
How to use Regression Residual
A positive residual means the actual value is above the fitted line and a negative one below. Residuals scattered around zero with no pattern suggest the line captures the relationship, and persistent runs on one side suggest it does not. A reading alone is never a buy or sell signal; it needs context from price structure, volume, volatility and the broader market regime.
Strengths
- It isolates the unexplained part of a series in the series' own units, so a residual of 2.5 means 2.5 units from the fitted line.
- A residual carries a sign, so above-line and below-line deviations are separated and can be tracked as one series.
- Reproducible when the formula and data convention are fixed.
Weaknesses and limitations
- A single linear regressor may leave out other drivers, so a large residual can reflect a missing variable rather than a true departure.
- Where the relationship is curved, residuals show a systematic pattern, so the line's misfit is mixed in with genuine deviations.
- Correlated indicators can create false confidence by repeating the same underlying information.
Editorial and risk note
This page is educational content, not individualized investment advice. It should not imply guaranteed prediction, accuracy or outperformance.
Frequently Asked Questions
Is Regression Residual a buy or sell signal?
No. A positive residual means the actual value is above the fitted line and a negative one below. Its scale depends on the units of y, so it needs standardizing to compare across series. A trade built on it still needs a trigger, an invalidation rule, a position size and realistic execution assumptions.
What is the best setting for Regression Residual?
Regression Residual is computed from dependent series y; explanatory series x; fitted intercept and slope, so the choices that matter are the data convention and the window it is read over. Test those across instruments, regimes and out-of-sample periods instead of fitting one historical winner.
Can Regression Residual be used by itself?
Its scale depends on the units of y, so it needs standardizing to compare across series. It also depends on the fit window, and the latest point has influence on the line it is measured against. For that reason it is read next to measures of direction, participation, valuation, liquidity and risk, not alone.
Does Regression Residual work on every timeframe?
The calculation (residual = y - (a + b * x), where a is the intercept and b the slope of the fitted line) can be run on any bar size, but behavior is not portable: a reading on a five-minute chart describes a very different market window from the same reading on a daily chart. Validate the exact timeframe before relying on it.
Why does Regression Residual give false signals?
A single linear regressor may leave out other drivers, so a large residual can reflect a missing variable rather than a true departure. Where the relationship is curved, residuals show a systematic pattern, so the line's misfit is mixed in with genuine deviations. Lag, noise and regime changes add to this, because market participants react to new information after the reading is calculated.
How should Regression Residual be backtested?
Rebuild Regression Residual from dependent series y using point-in-time data, write the rules before testing, include delisted securities where relevant, model realistic fills and costs, reserve a validation sample and compare with a simpler baseline.