From the 2d viewpoint, during contact each subject is watching a small orange sphere within the silhouette of a larger black sphere. The subject is, presumably, employing a control strategy that is based on (1) visual location of that orange dot within the black and (2) force feedback. My guess is that the control strategy is one that tries to maintain an equilibrium, with the orange dot at the center of the black. So you’d expect to see the subject exert a force that is always aimed roughly towards the center of that region, even as the orange and black spheres are subject to perturbation (random muscle movement, latency effects, etc).
If we can plot force-exerted vs. displacement-from-center for each trial, it would be interesting to compare this footprint accross conditions.