I agree completely with the above statements. I have heard and read many comments/articles about how to trim a pattern/IMAC airplane so it will not yaw/roll under all conditions and it cannot be done with a single-engine propeller driven airplane. The very large tail moments of pattern planes are attempts to mask the yawing/rolling effects generated by a spinning propeller, but they can never be completely eliminated under all flight conditions, not even with careful program mixing. If you set the plane up for zero yaw in hands-off level upright flight, the airplane will yaw left in a non-vertical positive climb unless the pilot adds right rudder, and will yaw right slightly in a non-vertical, positive power off downline, unless the pilot applies left rudder. The amount of rudder to apply depends on the power being produced, the airspeed and the AOA (AOA used loosely here).

The following discussion assumes the airplane is trimmed to fly wings level and zero yaw during aerobatic "cruise". At the top of a loop, the airspeed is less than at cruise, unless there is so much excess power available that the pilot "cruises" at less than maximum airspeed and adds power as the airplane climbs, in order to maintain a constant speed. Since without instrumentation and real-time telemetry this is probably impossible, I think we can assume the airspeed at the top of a loop is at least slightly less than at the bottom, if the pilot uses full power for the entry and at least the first half of the loop. Anyway, since the airspeed is less at the top, yet the power setting is normally the same, the aileron trim opposing torque is less, and the airplane will attempt to roll left near the top of a loop. The spiral slipstream is also slightly more compact, so there might be a bit of left yaw as well. Gyroscopic Precession might be more pronounced for the same reason (though the angular velocity is less since the airspeed is less and the radius of loop is constant), and opposes the left yaw force due to slipstream. P Factor is minimal since AOA is near zero at the top of a loop unless it is very tight and high positive G is maintained. There is no way to mix this correction in without messing up most other flight conditions, unless we have attitude telemetry between the plane and transmitter, or some form of autopilot. Every loop I have performed in full-scale propeller-driven airplanes (including full-scale aerobatic competition) has required right aileron near the top to keep the plane from rolling. It is interesting that the left-rolling torque effect is greater than the right-rolling stator effect on the wings/tail at the top of a loop, at least for the Decathlons, Citabria, and C-150 Aerobat I have flown, as well as every model airplane I can think of, including mid-wing airplanes. The F-16 I flew (once, for 20 minutes during a military incentive ride) did not, since there is no appreciable torque, P-effect, or spiral slipstream produced by the turbine engine.

As I alluded to earlier, it is possible to "cruise" at a low power setting and increase power so that the speed at the top of a loop is the same as or greater than at the bottom, which will cause different corrections to be required. However, even if the airspeed is exactly the same at the top as at the bottom of a loop due to power variation, there will still be a slight yaw correction required (which you could see if you were sitting in the airplane), because P-Factor will be less, due to the lower G loading and therefore lower AOA at the top. The higher power setting will also affect the slipstream and I don't know for sure what effect that will have.

These effects are not huge, so the casual sport flyer might not bother to correct them, or even understand or notice them, which is fine. Full-scale IAC aerobatic judges are generally well-trained, and often are full-scale competitors themselves, and they know what to look for, so if the pilot doesn't make these corrections, it will be reflected in her score.

P.S. A good computer model includes wind and gust effects, as well as the four primary yawing/rolling effects generated by propellers: Torque, P-Factor, Spiral Slipstream, and Gyroscopic Precession.