Ah, the dreaded MTF charts!
Maybe now would be a good time to refresh the links to a couple of explanatory (PDF) papers on the Carl Zeiss website:
As for Sigma, on the page you link to they state that 'the advantage of using “Geometric MTF” data is that it is easy to measure and calculate since it does not consider the diffraction quality of light'. While I'd agree that it is easy, or at least easier
, to calculate Geometric MTF I'd love to know how Sigma manage to measure such an MTF chart! Do they have a lens testing laboratory in some alternate universe where the laws of optics, and in particular the effects of diffraction, no longer apply.