Obviously back to the basic concepts Ian, you first need to determine just what you have available as reserve ballast. Obviously the boat needs to be stable and ballasted without load so I would start off with her in this condition at the correct water line. Then pump out all the ballast you can into a measured container and see just what weight you can replace with the load.
That would indicate the weight you can replace however you must also take into consideration the stability. Although you are simply replacing the same weight with the load the ballast is low down in the hull but the load is quite high up so you will have significantly reduced the righting moment (GM). This will have the effect of reducing the weight that you can add as load to maintain the suitable stability.
I think, bearing in mind the scale proportions, that a quite light load will be the answer to start with to maintain the required stability. Tests in the bath will be cruicial but don't simply observe the level at which the model floats, you must also induce instability by pushing the top of the mast over and observing the speed at which it returns to upright. If it is slow to recover then it will not be suitable to sail. Unfortunately this is one area that scale works against our models and we need our models to return to upright dramatically quicker than you would expect in a real vessel. This is dictated by the fact that waves you will encounter on the pond with be striking the model at a frequency of perhaps 2-3 per second whereas in the real world the ship will be experiencing wave frequencies of somewhere in the region of 10 seconds. Consequently you need your model to return to upright substantially quicker than a real ship as your righting moment has to be proportionally greater.