Jim they are both right, but for the sake of argument the in-phase is right, and it is right for a long time, but when the switch closes there is a good chance that the combination of left over flux, and switch closing angle will try to drive the flux past the normal peak and when things go to one extreme the core saturates extremely. In which case the inrush is so big that it swamps the load current, but likely when it is real high the transformer is no longer a “transformer”.

I’m not sure it I the same image, but on one you could see the small load current partially ?going to the right, but at the left you see the load current rise then the ?inrush swamps the waveform., so it was not the worst case. If the transformer was really left saturated, the current would shoot up with no delay with the voltage. I recorded a 30X inrush on a Honeywell visicorder. That core was being driven very hard. Too many volts per turn. That was a early class H potted transformer, maybe 25 or 50 kVA. Fusing as a real problem for it.

?