If the sphere has uniform density, then the amount of mass in each segment of the sphere is ~ density x pi x D**2 x A x T and ~density x pi x d**2 x A x T, where A is angle subtended by the chunk of the sphere at r, D and d are the distances from r to the sphere, and T is the thickness of the shell.
Since the force due to gravity is -GMm/distance**2, the two forces are ~ G x density x pi x A x m and ~ G x density x pi x A x m -- the distances cancel out and the two forces are equal in magnitude but opposite in direction -- they cancel each other (as advertised)!
F = ma ===>m dv/dt = - GMm/r**2 ===>m dv/dt = - 4 pi G density m r / 3
===> dv/dt = - 4 pi G density r / 3
===> v = r(0) dR(t)/dt and density = [some constant density = rho(0)] / R(t)**3
Note that there is always a force unless the density is 0 ===> we cannot have a static universe which contains matter. The universe must be doing something.
Hmmm, this is quite suggestive. The constant k in the above theory is simply a measure of the total energy (kinetic plus potential) of the mass m. A GR calculation leads to the same relationship given above, however, the constant -k/2 defines the curvature of the universe (and one can add a term characterized by the cosmological constant). This suggests that the curvature of the Universe is similar to the energy of our Newtonian universe.
Combining the above relations leads to
rho(now) < H(now)**2 / [8 pi G / 3]
then the Universe has exceeded its escape speed and will expand forever. This is a nice test because all we need to do is measure the current density and expansion rate of the Universe. Much of our current observational efforts are directed toward this seemingly simple task. I will much more to say about this exercise later.