Can we properly model the neutron monitor count rate?