Bill,
In case Al doesn't pick up on this one:
When he spread the word about UP-line forward, he was building large, very scale looking stunters - and using pretty large props.
The principle is that propellors act as flywheels to a significant extent. Flywheels are like gyroscopes. Gyroscopes demonstrate something called 'precession.' Set a gyroscope spinning and try to tilt its axis, axle, spindle (whatever you call it.) It moves as if you had pushed it 90° later in its direction of rotation.
At the RPMs we use, and the weights of our props, this can be a significant force. In 'ordinary' flight, with engine turning counter clockwise as we see it from the front, and with upright flight also counter clockwise, starting to rotate the model nose-up in pitch pushes the nose out, away from the center of the circle.
Think of pitching up as pushing back on the top of the prop disk. Seen from the handle, the upper half of a prop's rotation is away from the center. Precession makes this "push at the top" into a push at the outer edge of the disk, i.e., nose-out.
Same-same in reverse on pitching the model down - precession tends to yaw the nose in.
Up and down are not related to where the ground is: UP is where a pilot's head would be; and DOWN is toward his feet. Wherever the model is pointed...
We pass control forces to the model through the lines. Except for a few experimental over and under leadout setups, there's always some fore and aft distance between the up line leadout guide and that for the down line. The pull we feel acts on the model by aiming through the mid-point between leadout guides directly at the CG but this only applies when the pull on both lines is equal as in straight flight level or up a long straight maneuver track like wingovers or the diagonals in the hourglass.
When we put in control forces, the flaps and elevators meet airloads, and that shifts the mid-point of the pull force a bit toward the loaded line to overcome them. So, with CCW engine and CCW upright level flight, the nose wants to yaw out on up control. If the UP line is forward in that case, it shifts the middle of the total pull force a bit forward and tries to pull the nose in.
Same-same for DOWN - the loaded line should shift pull aft a bit to try to yaw the nose out while precession tries to push it in .
For the magnificent models Al Rabe has made for over 30 years, even that wasn't enough to overcome the "compromises" away from pure stunt layout (His scale model fuselages look HUGE compared to typical stunters...), so he also devised the Rabe rudder, which steers in the same manner as the counter-precession action from shifting pull.
And, if UP line is forward, we flip the bellcrank over (pushrod nearer the leadout guides), because that way the leadout cables can't possibly saw themselves apart.
Of course, if the flap horn were under the wing, and the elevator horn above the stabilizer, we could leave the bellcrank the way it was first used, 65 years ago by Jim Walker in a Fireball. He didn't have the stunting performance of our stunters, so the slight effect for a small, overpowered, essentially right-side-up-ONLY model wasn't important.