I'd never thought about it much before, but yeah, this kinda has to be the case. During the power stroke, the rod is not parallel to the cylinder bore, it's angled, and therefore is transferring a significant side load that must be the primary torque acting on the block. You could mitigate this...