I have to dissagree with Alasdair. I spent many years in the flight testing of real aircraft and missiles and can guarantee that everything will flutter given the input that will excite that structurer's natural frequency. The only way to eliminate flutter is to make sure that the natural frequency of any of the components lie outside the range of stimulus. You do that by structurally designing the component parts so that their natural frequency lie outside that range. Mass balance has nothing to do with that unless, in the process of adding it, you have changed the stiffness of the structure. The reason for having the inputs at an odd division of the span is to make sure that the control inputs have not inadvertently created a stimulus at a harmonic that is resonant with both halves (if the control horn were centered), just using the laws of probability to make the possibility of induced flutter a little less. Where mass balance helps, especially on ailerons and elevators, is to take the shock loads off the servos during high G impacts (read that hard landings) allowing the hinges to take the full force of the stresses and not stressing the push rods or linkage.