Dynamical Dark Energy (DE) is a viable alternative to the cosmological constant. Yet, constructing tests to discriminate between Lambda and dynamical DE models is difficult because the differences are not large. In this paper we explore tests based on the galaxy mass function, the void probability function (VPF), and the number of galaxy clusters. At high z the number density of clusters shows large differences between DE models, but geometrical factors reduce the differences substantially. We find that detecting a model dependence in the cluster redshift distribution is a hard challenge. We show that the galaxy redshift distribution is potentially a more sensitive characteristics. We do so by populating dark matter halos in Nbody simulations with galaxies using well-tested Halo Occupation Distribution (HOD). We also estimate the Void Probability Function and find that, in samples with the same angular surface density of galaxies in different models, the VPF is almost model independent and cannot be used as a test for DE. Once again, geometry and cosmic evolution compensate each other. By comparing VPF's for samples with fixed galaxy mass limits, we find measurable differences.

We show that using the redshift dependence of the deep cluster abundance to detect the nature of Dark Energy is a serious challenge. We compare the expected differences between flat LCDM models, with different Omega_mo, with the difference between LCDM and dynamical DE models. In the former case, cluster abundances in comoving volume and geometrical factors act in the same direction, yielding a significant difference between the expected angular densities. On the contrary, when we keep a constant Omega_mo and change the DE nature,abundances in comoving volume and geometrical factors act in the opposite direction, so that the expected differences in angular densities reduce to small factors.