In #160 a principled way to deal with multi-particle statistics was introduced. Moreover, it introduced an alternative backend implementation for fermion statistics (FermionBitstring). It should be documented, benchmarked, and potentially made into the default for fermions.