It is a common misconception despite being perpetuated and mis-quoted frequently.
Its not that voltage differences do or do not exist during starting, but the fact that they are small in comparison to the internal differences in battery impedance. You are going to all that trouble to save 1/10 volt during starting an
engine when in fact the inherent battery parameters are an order of magnitude larger. If they had not been in parallel and you tested each battery for voltage supplied during starting loads, their voltage would differ much more than the minor drops due to wiring layout.
Within a second after starting is completed, the batteries will all be at the same voltage due to battery to battery current so any imbalance only lasted a few seconds. Each battery contributed energy for the starting function but once completed they all are AT THE SAME VOLTAGE so each had contributed its share of the energy that it could release between the initial voltage prior to starting and the rest voltage after starting.
Eventually you took xyz amp-hours out of the battery bank to start the
engine. The net effect of wiring differences boil down to
heating loss which again is insignificant. For maximum efficiency the wiring layout should minimize wiring resistance.