To avoid this, I’d change all of the normal distributions that you may end up dividing by to log-normals.
Okay, I’ve now done this.
Okay, I’ve now done this.