When a given high-spin complex is jahn-tellerJahn–Teller distorted, this does not change the relative position of the total energy of the complex (assuming d10$\mathrm{d^{10}}$ configuration).
Indeed, the distortion only causes the degenerate triplet to split into a doublet and a singlet, and the degenerate doublet splits into two singlets.
In the case of a d4$\mathrm{d^4}$ high-spin complex, however, the total energy of the system is lowered, since the one ex-e_g$\mathrm{e_g}$ electron is now in a singlet, which has a lower energy relative to the doublet.
What stops this process, which is lowering the total energy more and more? Why isn't every possible complex for a jahn-tellerJahn–Teller distortion automatically distorted in such a way, that we end up with a quadratic planar boundary case, as it can be observed e.g. in [AuCl_4]^-$\ce{[AuCl4]-}$?