On this page
Start from the original paired data
| Period | Benchmark (%) | Asset (%) |
|---|---|---|
| 1 | -2 | -3 |
| 2 | -1 | -0.5 |
| 3 | 0 | 1 |
| 4 | 1 | 2.5 |
| 5 | 2 | 6 |
Read different measures
| Measure | Value | Meaning |
|---|---|---|
| Slope | 2.1 | Fitted asset response to one benchmark-return unit |
| Intercept | 1.2% | Fitted asset return at benchmark zero |
| Residual standard error | 0.632456 percentage points | Response-error scale calculation with n−2 divisor |
| R² | 97.350993% | In-sample variation ratio |
| Correlation | 0.986666 approximately | Signed standardized association |
Use the points and period table
A high fit score leaves room for deviations and model issues. Compare each observed asset return with the fitted line, then inspect signed residuals and the relevant period identity. The line is not extended as a future forecast.
Preserve the evidence for interpretation
Record the actual benchmark, source window and return basis. A change in date pairing or unit convention can alter a slope without changing the separate input means. Inference requires a specified model and uncertainty method beyond these descriptive outputs.