One of the first papers to try to explain the Brazil nut effect hypothesized that the tendency of smaller grains to fill in voids underneath the larger grains during the “free fall” portion of the shake (when the container is accelerating downwards at a rate greater than g) is the dominant physical mechanism behind the effect [1]. A Monte Carlo simulation was performed to test this hypothesis, with each “shake” being simulated as a random (but local) rearrangement of grains. Because voids left under large grains could only be filled by small grains, and because the only way a large grain could move back down would be if many smaller grains moved out, the larger grains moved inexorably upward. In this way, the simulated system found a local potential energy minimum, but not the global minimum attained by having the larger grains on the bottom.