On Fri, Jun 26, 2015 at 11:15:33AM +0200, Knapp wrote:
> I would think predicted position would be a cone for a ship with a
> fast vector and a sphere for one sitting, "still".

No -- as it happens, they're the same shape.  The relative motions
available to a moving body in vacuum are identical to those of a
stationary one.

> The cone might bend or even split (more like a hole in the cone), if
> it is interacting with a gravity well.

Gravity wells have predictable effect, so over reasonably short
timescales have negligible effect on uncertainty in position.

- Tim