I still disagree: the problem is not the number of dimensions, the problem is that the physics is wrong.

If one flips the rubbersheet, the geodesics are the same, but clearly now the model doesn't match our intuition. Indeed the objects are attracted by the mass because of the Earth's gravitational potential, not because of the rubbersheet curvature. The balls are following the variation in height of the sheet (in a sense a first order variation) not the curvature of it (the second order variation). The model is bending the wrong component (space, instead of time), and with the wrong sign (the geodesics are repelled by the mass).

As pointed out by Ron Maimon in this answer: http://physics.stackexchange.com/questions/1019/common-false...

a much better explanation can be offered introducing proper time and the effect of gravity on it.