I'm having a hard time understanding how the PVVT and PVAVS systems would
end up with dramatic differences in heating energy consumption and wanted
to know if anyone had ever compared the two before?

I'm calibrating a model for an existing building. The installed system is
one of the common Trane SWUD units with single-stage electric heat and
three-stage cooling. If I model the system as a PVVT system to capture the
staging, the energy use in June/July/August is within 5% of actual
consumption but is understated by ~25% in the shoulder months and almost
50% in Dec/Jan. If I model the system as a PVAVS system, heating months
are within 5% but cooling months are overstated by ~15%.

The only keywords I adjust between the two are the ones that eQuest
automatically adjusts when the system type changes. One possible cause of
the problem from what I can see is that the PVVT system requires a control
zone and that control zone defines temperature for all of the other zones
(resulting in a lot of unmet load hours - up to 88% depending on the
control zone selected). Any thoughts?
Jeremy R. Poling, PE, LEED AP+BDC