Yes, a little Occam's Razor would go a long way here. Just far too many arbitrary assumptions, inferences, extrapolations, etc, required for their model. None of which have stood up to robust scrutiny, let alone the model as a whole.
The best fit model is still the simplest and most obvious...