Physical interpretation of MGIC_RV model-comparison results

Determine why the BIS activity diagnostic is more effective than the other tested diagnostics for modelling the stellar activity in V1298 Tau radial velocities and identify what produces spurious 3σ planetary-mass detections in the Gaussian-process modelling framework.

Background

The paper applies the MGIC_RV criterion to compare one-dimensional and multidimensional Gaussian-process models of the highly active young star V1298 Tau. The criterion strongly favours models using the bisector inverse slope (BIS) activity diagnostic, but those models do not produce any 3σ planetary detections, whereas models using other diagnostics yield apparently significant masses that may be spurious.

Because V1298 Tau is among the first complex systems to which this model-comparison criterion has been applied, the authors explicitly leave unresolved the physical reason for the differing diagnostic performance and the origin of the apparently spurious detections. Resolving these issues is important for assessing whether the inferred planetary masses are genuine or artifacts of activity modelling.

References

Since V1298 Tau is one of the very first complex cases to which the MGIC$_{\rm RV}$-based criterion is applied, we consider it appropriate to leave the detailed exploration of the physical meaning of the model comparison results for a future work, specifically to understand why the BIS proves to be the most effective stellar activity diagnostic compared to the others we tested, and what would produce spurious $3\sigma$ detections in our modelling framework.

— The GAPS programme at TNG LXXIX. New mass constraints for the infant planets in the V1298 Tau system through a long-term radial velocity monitoring with HARPS-N  (2609.30038 - Damasso et al., 24 Sep 2026) in Section 4.1, subsection “Model comparison analysis using MGIC_RV”