If you did it the way FC is suggesting, it should show the movement - ball low when you're low, and vice versa. The question is, will it show enough movement to be accurate? I'm wondering how the movement is affected by the ball being closer or father behind the datums and also how your distance to the datums (from your aircraft) affects movement. We should arrange it so that the closer you get, the more sensitive it becomes, but I'll need to take a few minutes and use some (gasp) brain cells to wrap my head around this one.
The tricky part would be determining the appropriate distance for the ball to be behind the datums, and of course, figuring out the correct glidepath for the centered ball condition.
SKippyBing's suggestion about the angled boxes also seems practical - at least if we could get them into the ship models. If we can, it should work out nicely.