Working abstract
GV measures whether a system's ability to recover after disturbance is degrading persistently over time. Synthetic falsification tests show that GV stays quiet on recoverable noise, misses abrupt collapse without precursor degradation, and warns when recovery force weakens before visible failure.
Current operational form
Gv ≈ f(dR/dt, recovery_time, persistent_degradation_probability)
Boundary results
- Stable α region: 0.75–0.85
- False-positive boundary: ≈ 0.90
- Synthetic precursor lower bound: ≈ 0.001
- Abrupt collapse without precursor: missed
Reality check
- This is not proof of a universal law.
- Synthetic systems are still controlled.
- Real-world validation is early.
- GV must survive external datasets.