The spot will go through the same convolutions if you move it around the screen with the shift controls. This rules out any cross coupling in the amplifiers, hum related things or dry joints. X and Y waveforms at the deflector plates look good on another scope. Anything electrostatic somehow stored in the tube would not have survived the wash of the beam for this long, even tubes meant for storage would not hold anything for more than a few minutes with the read beam on. I suppose it could be something stored magnetically on a ferrous mesh, but I've got some fair magnets about the place and I can't shift it much, and then only temporarily. My bet is a mesh that has become detached, but then again, laying into the side of the tube with a rubber mallet would be expected to shift it some, but no. I shall eventually open the tube and find out for sure, once any hope of getting a replacement tube from Mitsubishi for its novelty value has gone.