Direct Answer
A residual z-score standardizes the error left over after a regression fit. It shows how many standard deviations the latest observation sits from the value the fitted relationship predicts. A large positive value means the series is above what the relationship implies, and a large negative one means below.
Residual Z-Score: Formula, Meaning, Signals, Examples and How to Use It
How Is Residual Z-Score Calculated?
A residual z-score standardizes the error left over after a regression fit. It shows how many standard deviations the latest observation sits from the value the fitted relationship predicts.
Formula: z = (residual - mean of residuals) / standard deviation of residuals, where residual = y - fitted value
Inputs:
- Dependent and explanatory series
- Regression window
- Residual mean and standard deviation
Worked example. A residual of 1.8 with a residual standard deviation of 0.9 gives a score of 2.00. The inputs are illustrative numbers, not market data.
Reading it. A large positive value means the series is above what the relationship implies, and a large negative one means below. It is widely used to judge how stretched a spread or a relative-value pair is.
Limits. It assumes the fitted relationship is stable, and a structural break turns a signal into a model failure. Residuals from an in-sample fit understate real out-of-sample error.
What is Residual Z-Score?
A residual z-score standardizes the error left over after a regression fit. It shows how many standard deviations the latest observation sits from the value the fitted relationship predicts.
How to use Residual Z-Score
A large positive value means the series is above what the relationship implies, and a large negative one means below. It is widely used to judge how stretched a spread or a relative-value pair is. A reading alone is never a buy or sell signal; it needs context from price structure, volume, volatility and the broader market regime.
Strengths
- Standardizing residuals puts the miss of a fitted relationship in standard deviations, so mispricings across pairs share a unit.
- Subtracting the fitted value first removes the part explained by the regressor, leaving only the unexplained component to be scored.
- Reproducible when the formula and data convention are fixed.
Weaknesses and limitations
- Residual scatter can change with the level of the regressor, so a single standard deviation understates the miss in turbulent spells.
- Each score depends on regression parameters that change as the window rolls, so a past residual is not fixed in later recalculations.
- 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 Residual Z-Score a buy or sell signal?
No. A large positive value means the series is above what the relationship implies, and a large negative one means below. It assumes the fitted relationship is stable, and a structural break turns a signal into a model failure. 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 Residual Z-Score?
Residual Z-Score takes these inputs: dependent and explanatory series; regression window; residual mean and standard deviation. There is no universal best setting for the tunable ones, so test them across instruments, regimes and out-of-sample periods and favor stable parameter regions over one historical winner.
Can Residual Z-Score be used by itself?
It assumes the fitted relationship is stable, and a structural break turns a signal into a model failure. Residuals from an in-sample fit understate real out-of-sample error. For that reason it is read next to measures of direction, participation, valuation, liquidity and risk, not alone.
Does Residual Z-Score work on every timeframe?
The calculation (z = (residual - mean of residuals) / standard deviation of residuals, where residual = y - fitted value) 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 Residual Z-Score give false signals?
Residual scatter can change with the level of the regressor, so a single standard deviation understates the miss in turbulent spells. Each score depends on regression parameters that change as the window rolls, so a past residual is not fixed in later recalculations. Lag, noise and regime changes add to this, because market participants react to new information after the reading is calculated.
How should Residual Z-Score be backtested?
Rebuild Residual Z-Score from dependent and explanatory series 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.