you're correct. In stats/thermo class there's a commonly taught "what's the probability of all the molecules of air in a room spontaneously moving to <extremely small location>". The problem is that shoving all those molecules into a tiny location woudl increase the pressure tremendously, sending all the particles in directions that would eventually return to a uniform distribution.

IIUC my professor right the probability is non zero but is practically impossible for a large number of incompressible spheres.