Yes, d8 transition metal ion always form square plane with exception of Ni2+ which with strong field ligands forms square plane and weak field ligands tetrahedral. The reason why all other d8 ions form tetrahedral lies in the fact that as they are for the second and third row of transition metals (think of Pd, Pt, Au, Rh, Ir) form complexes with bigger Δ than the first row transition metals (like Ni) because of having larger orbitals (3d<4d<5d), causing more repulsion, increasing Δ, just like strong field ligands do.
d4 I think have more tendency to form tetrahedral complexes because in order to form square plane they first need to form an octahedral arrangement and distort in order to stabilize electrons formerly occupying degenerated eg orbitals. When one makes the diagram it's easy to see that d4 high spin octahedral complexes can distort, according to Jahn-Teller's theorem, in order to stabilize the z axe, increasing the length of the metal-ligand bond in that axe, and hence forming a tetragonal structure. But it can hardly derive on square plane complexes because it only has one electron (an odd number of electrons) in eg orbitals, needing at least 2.