The superconductivity in alkali intercalated C-60 crystals is analysed on the basis of conventional B.C.S. theory. It is shown that, in the limit of negligible electron phonon coupling between C-60 molecules, the coupling parameter lambda of B.C.S. theory can be rigorously factorized as lambda = NV where N is the density of states at the Fermi level and V is a pure intraball contribution. We describe several calculations of V (Tight binding, LDA, MNDO) which all lead to a value in the range 50-60 meV. This, together with reasonable values for N (similar to 15(eV)(-1)/C-60) and for the Coulomb pseudopotential mu (similar to 0.2) leads to T-c = 20 K as observed for K3C60. We show that the dependence of T-c upon lattice parameter strictly reflects that of the density of states and confirms the factorization of lambda. We also compare our predictions to available experimental data (phonon self energies, isotope effect...). We discuss the origin of the difference in T-c between the A(3)C(60) and intercalated graphite. Finally we comment on the validity of the B.C.S. approach showing that different corrections due to the narrow bandwidth tend to compensate each other.