I thought about the shortcoming of my First RockSim model and decided to approach it differently. May or may not be better, but I did it anyway!
Instead of a BT-60 rocket with a 3" ringtail, this one is a 3" rocket with a central BT60. Since you can't attach a nose cone to an inside tube, the exposed BT-60 is separate from the BT-60 section that is inside the 3" body. The two funky parts of this one is that only the exposed part of the strakes are modeled and the front BT-60 abuts to the 3" tube without a transition section. It's not clear how RockSim handles this.
This version is much more stable than the first. With a C11-3 the margin is ~2.5 with the CP on the rear nozzle section. The expected altitude is 218' and the optimal delay is almost 3. With a D12-3, the altitude is 447' and the optimal delay is essentially the same. This still doesn't compute to me. Maybe it is because it is moving faster and RockSim thinks the drag is higher?
Since this version behaves better, my confidence is a bit higher. I may go back to the C11 just to keep the altitude down on my semi-local field.