is there any accountability regarding the accuracy of these models .. Does anyone do post hoc analyses on their predictive validity ? Someone needs to investigate.. The models should be getting better with time … not worse !