I had the same experience, Moss. At full lift, and maximum dump, having dumped, it would not curl the bucket back up (retract the cylinder) until you started lowering the lift arms. Then, it would curl up as normal.
I don't have much bucket experience with this old 1430, mostly mowing, so I'm not sure how it moves through that arc. I do know that the loader arms are significantly different on the new 1430s - the angle of the bends are sharper, and the QA plate/pivot points seem to be a bit closer to the front wheels. Per Terry, swapping to new arms was not an option to get a QA setup, so I didn't compare much farther.
The whole geometry of these PT loader arms is pretty complex, and it would take software modeling (or a smart engineer) to fully illustrate it. For example, the height/distance of the arms on that rollover bar is greater than the distance between the pivot pins on the QA and the point where that top link attaches to the QA plate. So, the loader arms (and their pivot points) do not form a parallelogram with the top link (except at a certain point in curling up and down). Plus, those arms on the rollover bar are moving in an arc all the while. This is nothing like the normal bucket dump/curl on a regular tractor, where the only change is the attachment pivoting on the pivot pins. Now combine that with the arms underneath the rollover bar where the curl/tilt cylinder connects are shorter (at least on the 425) than the arms on top where that top link attaches. One inch of movement from the cylinder does not equate to one inch travel of the toplink - the toplink travel is amplified by the ratio of those different lengths.