Ball bearings...that would be just fine under normal circumstances...except that it would not be practical, since the speaker would be easily moved by accident[a feather-light touch, for example].
I doubt the accuracy of your report, it does not add up. Nor could I come close to replicating it, with extreme conditions[that were used in order to give you the benefit of the doubt]. Did the speaker have a 500 gram cone, or some other unusual circumstance? Did you purposely measure at a resonant mode related to the coupling[which is another matter]? What was the coupling you used? Such a value of movement as you report is not supported by any calculations of which I'm aware, under any normal application.
Since you pose the ball-bearing example, I did just that. I took an old 10" car subwoofer, threw it in a spare, half-completed subwoofer cabinet, which weighed in at 38.6 lbs on a digital scale. I placed a sheet of acrylic down onto a piece of smooth wood on a level surface. I then placed 6 ball bearings on the acrylic. I then placed another layer of acrylic on top of the ball bearing. I placed the cabinet on top of this second sheet.
http://www.linaeum.com/images/sm1.jpg
At this point, the cabinet could be moved with a feather-like touch[no suprise].
http://www.linaeum.com/images/sm2.jpg
I then used a 20Hz sine wave, and adjusted the amplitude until the speaker cone moved +/- 6mm (12mm total travel). I placed a tape measure, with tape, onto the bottom piece of plexi, so that it aligned the measurement marks under the cabinet.
http://www.linaeum.com/images/sm3.jpg
The mm units will make it easy to see a 0.015"(0.381mm) movement by eye. I also video taped the action using a macro lens, so that I could review it on a monitor/zoomed. No visible movement with bare eye viewing. NOt on the video, either. 0.015" of movement did not occur. Not even close. And this, at high amplitude, with a relative light-weight enclosure, set on ball-bearings for minimum friction in the axis of speaker/air movement.