Even on a positive 45d up-line, the wing is positively loaded, though not as heavily as in level flight, and adverse yaw works in the conventional way, i.e. too much down aileron causes yaw in that direction. But here's a twist: In a 90d up-line, the nose will yaw slightly toward the aileron with the greatest deflection, or presented area. So if both ailerons move up more than down, in this case the nose will yaw right when applying right aileron, and left when applying left aileron, ceteris paribus. The above method uses a 45d line as a compromise between level flight and a 90d upline. I think a 60d upline might also be a good compromise, since it will result in even less differential, which suits me. But if you have an airplane that is symmetrical with neutral incidences, and symmetrical surface hinging, just set the aileron throws as equal as possible, then fly the thing. These effects cannot be eliminated for all airplanes and flight conditions; they can only be made small enough that they are not noticeable to the judges under most conditions.

Incidentally, I don't use the above method because of the relatively strong P-Factor effect that exists in a 45d upline. You end up ignoring P-Factor, while deciding how much differential to program. What about the 45d inverted uplines? Now you've fouled things up, though maybe not to a large degree. Full-scale acro pilots (I'm one of them) don't spend time trying to optimize aileron differential and rudder mixing. They spend time practicing using the rudder to correct for these effects. That way, with practice, they can use the right amount of rudder for any flight condition, instead of trying to find an automatic compromise that only works perfectly in one flight condition. I know this flies in the face of the conventional wisdom that you have to set up and trim your airplane the way the big dogs do to hope to compete in pattern/IMAC. To each his or her own!