A few considerations for a movable weight device:

1. Moving a weight affects the inertia of the airplane as well as CG. Since you said that you wanted to be able to move the CG aft for "wild aerobatics", you would probably want to have the least airplane inertia in the aerobatic mode. That means the movable mass would need to be closest to the airplane CG during aerobatics. Thus the weight would be up in the nose for stable flight and back near the CG for aerobatic flight. If you move weight back into the aft fuselage for aerobatic flight, you will be increasing your inertia. That slows down any rotary responses in pitch or yaw. Not what you want.

2. A variable CG system will change both pitch and yaw stability, and have no effect on roll. Is that what you want ? Your neutral point in yaw probably isn't at the same CG as the NP for pitch. But the CG moves the same distance when the weight moves. So having moved the CG say, one inch, you might be at the neutral point in pitch but still quite stable in yaw. Something to think about.

3. A lighter and perhaps less complex alternative might be to add movable canard surfaces to the forward fuselage. Deploying horizontal canards up front would destabilize the model in pitch. And of course deploying a vertical up front would destabilize in yaw. By tailoring the size of the canards you could have just the desired stability reduction in pitch and yaw, no more and no less.

Dick