Kert,
I understand conceptually what you're getting at, but there are a couple of things to consider....
1. The springs in your drawing would be a lot thicker and more substantial than what you've shown.....ie. stiffer. And the forces aren't pushing the spring UP...they are moving laterally, trying to compress the distance between the spring eye and the springpad.
2. CamoK5's wrap bar, even with a broken lower link is not a good design for validating a single-link antiwrap bar. Look again at how close the upper link pivot is to the top of the pumpkin housing...it only looks like it's maybe 2" higher than the pumpkin itself....and too close to the axis of rotation to be effective. You were right in speculating that it does not have enough "link separation" from the axle leafspring mounts to create good triangulation.
I think the real answer is that in real life, with more link separation and normal leafsprings....there will be substantially more stiffness than you are illustrating. Ntsqd will have to provide the "heavy math" that proves WHY if he's so inclined.