Taking into account all rules of science and observation made by the great minds of the past. One could safely conclude that a marble falling through a liquid would have the same impact upon reaching the bottom regardless of it's size
Then you surely won't mind sitting on the bottom of a swimming pool with me (holding our breath or using scuba gear), and playing "catch the stone." My assistant will drop a pebble in the water to me, and he will gently unload a boulder from a forklift to you. Are you in?
I believe that richmke's analysis is pretty good. However, I don't think (but am not honestly sure) that we are in the regime where the friction does not differ in an important way between what we are calling "large" and "small" marbles. Due to the viscosity effect that Bkisel cites, a marble falling through a liquid
will eventually reach a steady-state speed (the terminal velocity). I doubt that a 23L carboy is tall enough that this will happen for marbles in wine, but don't really know. Interestingly, though, the terminal velocity scales as the square of the radius of the marble (R^2). This is because the force due to gravity scales as R^3 but the frictional force (i.e., the
Stokes drag) scales only as R. (Tenbears, this already takes into account Archimedes's principle.) So the terminal velocity of marbles in a viscous medium is much larger for large marbles.
It gets worse! If we consider the energy of the marble, as rchmke did, we find an astonishing difference. As I said above, the terminal velocity V scales as R^2. But the kinetic energy (0.5mV^2) scales as V^2. And, as rchmke pointed out, the mass scales as R^3. So, put this all together, and the energy scales as R^3 (for mass) times (R^2)^2, or R^7. This is a big difference for big vs. small marbles. Comparing 5/8" marbles to 1" marbles, the ratio of the energies is (0.625)^7=0.037. So, if terminal velocity is achieved in both cases, a 5/8" marble carries only 4% of the kinetic energy of a 1" marble.