I highly suspect that interplanetary flight would be more accurate in real time, since it allows shorter time steps in the simulation and IIRC Orbiter uses a lower order propagator at high time acceleration.
For instance, I suspect that some to most inaccuracy in the IMFD n-body predicting Map is due to numerical error in the simulation itself, and not the fault of the map.
It would be interesting to see how much delta-V savings there would be in MCCs by doing a mission in real time versus under high acceleration.