I always build planes with no right thrust in the actual airframe, (firewall offset) and depend on the trimming process to show me what's required. Flying is the only real proof. Also, since different engine sizes and prop combinations might affect the amount of offset thrust required, a straight airframe offers more choices.

I don't think comparing these needs to full-size aircraft is necessarily valid. After all, (especially with aerobatic types) we fly at radically different Reynolds numbers, with aircraft that often are turning props at 5 to 10 times the speed of the full-size aircraft, and moving the aircraft at scale speeds often equal to several hundred miles per hour.

I've found that for true "neutral" response, as in a Pattern plane, right thrust is nearly always required, when the airframe is straight, symmetrical, and balanced. Trying to compensate for this with control trims would not only lead to other problems with aerobatic precision, but it would increase drag as well.

One example:
The Raytheon King Air, which I've documented for a scale model I'm building ( http://www.nextcraft.com/b200_construction02.html ) has a one degree twist built into the vertical stabilizer, producing a "right rudder" effect.