Pat,
I am not saying that what you advise is wrong regarding the intersections of the overhead eights. What you suggest may be a learning tool for a pilot to improve on his eights. But the pilot should be aware that putting "D" intersections in any of the round eights is an error, and depending on the degree of error, the maneuver score should so reflect. The problem is that so many flyers murder this part of the maneuver, that most judges sort of expect that that is the way the eights are to be flown. (Sort of like watching everybody do 120-degree wide horizontal eights - everybody does it and when somebody flies the 90-degree horizontals, it looks "TOO TIGHT".) Anyway back to the intersections of the overhead eights
First, intersections of all of the round eights are just a point of tangency. In my clinics, one of the things I empahsize over and over is that this point of tangency is just that, it is a point. The airplane is turning one way and then it should instantly be turning the other way. It is physically impossible for the pilot to change the handle position fast enough to do this properly. At least, the change in direction should be as quick as the pilot can flick his wrist from an inside turn to the outside turn, and vice versa. (That flick of the wrist, though it can be a very short part of a second, it is still not instantaneous.) A good experienced judge should look for that instant (or at least as close as possible) change in turn. When a good judge sees that straight line as you recommend, there should be a major downgrade in the maneuver score for each intersection where that straight path occurs, whether on the horizontal round or vertical or overhead eights. Also, an experienced judge can easily tell if the orientation/direction of the intersection is the same each time the model flies through those intersections of any of the round eights. The judge does not need to be pefectly aligned with the direction of those tangent points in those vertical eights. The judge has to watch and pay attention to what aspect he sees of the model at those tangent points. The judge will see one of three aspects of the model at that point. The judge might see only the tail view of the model (when the judge is perfectly in line with that tangency direction, or the judge might see some aspect of the top of the model or some aspect of the bottom of the model. Whatever is viewed, that aspect should be the same each time the model is flown through that intersection. If that same aspect if not seen, then it is a measure of how poorly the model has "X'd" the intersections or over-rotated while flying through the intersections.
Judges are not in a good location to see the "roundness" of the loops in the vertical eights. But good judges can see if there are incorrect straight flight paths in those intersections and how the length/duration if that straight path is held in the intersections and the maneuver scores will so reflect "D-shaped" loops in any of the round eights should not get good scores from a judge, and judges can see these, even in the overheads.
Even though roundness cannot be accurately seen on the overheads, the judge can watch for turn rates through the loops even when the pilot is constantly adjusting those turn rates as the heading changes relative to possible windy conditions. Again, an experienced judge will be able to transpose in his mind what those loops really look like on the surface of the hemisphere. In addition, and specifically on the overhead eights, the judge can tell how close the edges of each of the loops are to the specified 45-degree elevation. When you see those bottom extremities of each loop in the eights (which are the lowest part of the overhead eights), the judge can quickly visualize how tight another loop would have to be in order to fit beneath that bottom extremity (specified to be 45-degrees) and the the 5 foot "normal flight altitude". Here again, with that tool in the judge's mind, it can easily be seen if the loops of the vertical eights are too big (which they normally are, even with experienced flyers), or if they come close to the specified 45-degree diameters. The pilot can also use this to determine if the loops of his overhead eights are too big. Just visualize at that lower extremity how tight a loop would have to be performa a loop from that lower extremity and the ground.
All of this sounds like a lot of things to think about while flying or judging. I believe that experienced judges naturally see these things without having to mentally go through a check list of things to look for during each maneuver. Experience will flag whatever errors might appear. It is just that experience is needed together with a comprehensive knowledge of what errors can be flown and sometimes are intentionally flown - like "D" intersections in ANY of the round eights.