Native regressions of timing-aware stopping
Completed 4/4 original-start fits; 4 pass numerical and physical checks. These generated-data regressions add no aircraft observation or statistical confidence claim.
| Case | Earlier fit passed | New fit passed | Evaluations | Later horizontal RMS | Later height RMS |
|---|---|---|---|---|---|
| 1 | True | True | 26 | 1.082 m | 1.997 m |
| 2 | None | True | 26 | 1.082 m | 1.997 m |
| 3 | None | True | 10 | 0.942 m | 1.939 m |
| 4 | None | True | 10 | 0.942 m | 1.939 m |
Earlier results retain their original labels. Every input, weight and charged operation is preserved. Broader optimizer qualification remains a separate gate; calibration does not restart automatically.
All fits · Failed checks · Start comparisons · Frozen plan · Next analyses.