The problem of microscopic nuclear structure theory in large single particle basis systems is reviewed. Several approaches are discussed, which attempt to approximate the large model spaces numerically inaccessible in complete shell model expansions of the nuclear wavefunctions. All of them use symmetry projected Hartree-Fock-Bogoliubov quasiparticle configurations as basic building blocks of the theory. They differ, however, in the degree of sophistication of the variational procedures which are used to determine the corresponding mean fields as well as the configuration mixing, u p to a level, on which the construction of the configuration space itself is entirely left to the dynamics of the considered system. The mathematical formalism underlying these models is briefly summarised and the steps towards a numerical realisation are discussed. In several examples the possibilities and the power of the models are demonstrated and their limitations are shown. The models may provide a powerful tool for the analysis of experimental data as well as for predictions in still unexplored regions. On the other hand they may lead to a much better theoretical understanding of effective nuclear interactions as well as the underlying fundamental forces.