[Radu]

 We have the electronic configurations: Ni:[Ar]3d⁸4s²;
                                                         Pd:[Kr]4d¹⁰
                                                         Pt:[Xe]4f¹⁴5d⁹6s¹

     Can anybody explain me why Pt chooses to be paramgnetic? I know that the energy differences between different shields are lowering with the increase of n, so that's why 4d gets fully occupied at Pd. But why Pt:[Xe]4f¹⁴5d⁹6s¹ and not 5d¹⁰?

[Enthalpy]

Magnetism is a molecular property. Atomic orbitals won't tell you much about the metal.

Examples: austenitic stainless steel, with Fe Cr Ni, isn't ferromagnetic, unless cold-worked.
But Mn Zn "ferrites" are used everyday as magnetic cores.

[Radu]

 I don't think I understand. Are you saying that the magnetic measurements used to determine the electronic structures have been conducted on sheets of metal on atoms rather than on atoms and that the results were showing the behaviour of the "molecular" Pt instead of an Pt atom? Or what?
  I am looking more for a quantum mechanics reason.

[Corribus]

Maybe you'll find this link helpful?

http://www.chemguide.co.uk/atoms/properties/3d4sproblem.html

[Enthalpy]

Oops, atoms.

We had a discussion on chemicalforums about the scandium to zinc group, and it was already a mess. By including the repulsion of electrons within a single orbital, plus some interaction among two shells, it was possible to reproduce more or less the filling sequence. But I doubt any prediction can be made for heavier atoms.

"QM reason"... The filling sequence resulting from repulsions, hence from geometric details, it won't be a simple consideration on quantum numbers.