The door has fielded panels, i.e. a much thicker stronger raised field in the middle with a much thinner weaker bas relief surround. Bullet D passes through the bas relief within millimetres of the raised field, i.e. the the panel's weakest point. D is the only bullet hole to coincide with a crack which "damaged" the hole where they meet demonstrating that bullet D was unequivocally made before the bat strike that cracked and broke the panel in that area. This was confirmed by Vermuellen in cross and would be confirmed by anyone who has a fair knowledge of even basic DIY woodwork.
The thinner surround gives a much shallower bullet hole (around half the depth of the three others passing through the raised field) which as bullet hole D was also damaged by the bat and/or the crack it will have significantly more slack for the rods to wobble around in making difficult to ascertain a precise angle at which the bullet passed through the door (the deeper the depth of the hole the more "wall" to guide the rods through at the correct angle). Trajectory D therefore is IMO the least reliable because the slack and damage afford the rod more movement and a change of even just a few degrees can significantly change the trajectory as in the two photos you liked in which bullet hole D shows two completely different trajectories.
