In the Goodness of Fit table, each parameter is rated on how close it comes to the ideal value for each Goodness of Fit metric. The rating is indicated by the number of gold stars or red crosses next to the parameter, from three red crosses being the worst to three gold stars being the best.

Note: Root Mean Square Error has no rating because it is not a bounded characteristic.
Goodness of Fit metrics, when calculated for different types of response surface points, measure different aspects of your response surface. You can interpret your Goodness of Fit table as follows:
| Type of Point: | GOF Measures: |
|---|---|
| Learning Points | Quality of the interpolation |
| Verification Points | Quality of the prediction |
| Cross-Validation Learning Points | Stability or reliability of the response surface |
For the Genetic Aggregation response surface, goodness of fit metrics calculated for learning points via the cross-validation technique provide an especially effective way to assess the stability of your response surface without needing to add new verification points. If the GOF metrics for cross-validation learning points are good, you can be confident in the quality of your model. If these GOF metrics are not, you know that you need to enrich your model by adding new refinement points.