Last night, I was making a RockSim model of another strange design (more later) and ran into the often asked question: RockSim or Barrowman? As an aside, RockSim can either compute the CP using the venerable Barrowman equations or using a proprietary method that accounts for some additional factors omitted from those equations. Anyway, RS Version 9 defaults to Barrowman, even when you add things that can't be simulated using that method. In this case, I remembered to select the RockSim method. However, it made me think that I didn't do so in the design of my Toypedo. Although there are no components that shouldn't work using the Barrowman setting, it appears the program was in fact barfing on it. That explains why I was getting CPs ahead of the model with the 'stock' fins. I just changed the setting and now the CP with the stock fins is about in the middle of the rocket...a result which seems reasonable in my meager experience.
Update: What timing! The differences between the Barrowman and RockSim methods are the subject of the latest Apogee Newsletter (#238).