ChazInMT- the way i understand it, the top view that you're depicting is different than looking a the car from the side. unlike from the side, the top view is not in a ground plane and it therefore not a half body but a full body of rotation. When the car pushes the air molecules around the car, they go up, left, and right. Aside from slight underbody flow, these are the three planes for the air molecule to move in(excluding swirly and moving forward and backward). Therefore, there must also be 3 seperate planes to allow the air to snap back into place behind the vehicle.
To determine where to place the template, you align the bottom of the template with your stagnation point(in this case, the middle of the car splits flow to both side) and then the you align the top with the point of max side camber. When looking down at the car, you want it to look like a tear drop.
On the side view, the reason the bottom of the template is not placed directly on the stagnation point is because the underbody flow is not considered. Therefore, the ground plane becomes the stagnation point.
The method you've been using does not hurt, but you might not be reaching you full streamlining potential. it's always safer to go longer than it is to go shorter.
Hope this all makes sense. It's the way I've interpreted the streamlining information and I've yet to earn my high school diploma, so I could be wrong.
