Some controversial results that arise in the problem of the scattering of gravitational objects are analyzed in this paper. First, the approach leading to possible use of the approximation of pairwise (binary) collisions in the theory of scattering in plasmas and in gravitational systems is examined. It is stressed that this possibility is not a consequence of neglecting long-range (distant) collisions. Next, results obtained for the number of scattering events of a test particle in an infinite gravitational medium with a fixed velocity change are discussed. The conclusion that the function describing the number of particle collisions is divergent is shown to be incorrect. This is because the finite time duration of collisions with large impact parameters is neglected. In concluding, we show that the previously proposed variant of a modified Holtsmark distribution is erroneous.