My cheap theory is that the drain theory is not a theory, but marketing assumed to be most palatable for common-sense physicists.
From Stereophile's Bad Vibes!, Shannon Dickson, November, 1995:
"Unfortunately, once we've built or purchased our dream platform, we then have to connect it to a stand or floor and place a component on top. This is the kicker: When you couple the most ideal practical platform to the floor with cones, spikes, or any other rigid footing, even at the ideal locations with respect to each, the best vibration performance you can achieve is nearly 100% transmission of floor-borne vibrations through the platform, without amplifying them or generating any new resonances in floor or platform! The same applies to component-generated vibration. At the very best, the combined structures will roughly approximate the "ideal rigid body" we mentioned earlier, moving through space in synchrony relative to each other so that the motion of the floor is matched by the motion of the shelf, with nothing added.
Any technique that does not provide isolation of external vibrations will only vary the amount of resonant stimulation added to the components concerned. It cannot reduce at all the level of baseline vibrations in the floor or those coupled from the air!
This principle is illustrated by both the "ideal rigid body" line in the compliance curves shown in sidebar 1, and the horizontal unity-gain line (labeled "1.0" in the various transmissibility graphs of sidebar 2). A perfectly rigid structure would not diverge from this unity-gain baseline in either direction, indicating nearly complete transmission of all vibrations between both the floor and the coupled elements.
At first glance, transmitting nearly all of the floor vibrations to a component might seem to be of no benefit at all. On the contrary, this would be a significant accomplishment compared to most real-world coupling schemes, due to an appreciable reduction in random levels of resonance affecting key components, as described above.
Indeed, it is the degree of deviation from this ideal that defines the wide variety of subjective sonic changes experienced by audiophiles using various non-ideal rigid coupling devices, stands, shelves, and components in actual audio systems. Also, when you consider all the ramifications of this scenario, it appropriately undermines the claim by certain purveyors of cones and spikes that these devices have a directional "diode-like effect," forcing discrete vibrations to flow like water from a dam: out of a component, through a coupled shelf, and then into the floor, where they are finally dissipated..."
Read the
rest; its good.