So, yeah, an energetic water molecule ought to be able to make a visible difference in the motion of a colloid particle.

Then you add the success of the statistical math Einstein presented for the RMS displacement as a function of time and the theory looks really solid.