Bayesian analysis of multiplanet systems with multiple datasets

Application of Bayesian model inadequacy criterion for multiple data sets to radial velocity models of exoplanet systems

Some interesting conclusions: firstly with regard to Gliese 581 they end up with a 5-planet model being most probable, but there is not yet enough evidence to reject the four planet model.

For Upsilon Andromedae there seem to be inconsistencies between various published datasets. Their final set of parameters gives a lower period and mass for the fourth planet, but I have to wonder whether this can be stable given the high masses of the second and third planets determined by astrometry...

Even accounting the fourth planet is only RV-detected? Maybe astrometry could furtherly reveal a higher mass for the outermost planet, as well as likely different inclination.

Eccentricity of gliese 581 d in my opinion seems to be very unlikely over 0.3 with no large disrupting gas giant at all in that system.There is even more likely existence planet f than g