Let me try and get my head around this from a different angle and tell me if this is accurate;

The 1D X in the largest format makes a pic 5184 x 3456 while the 5D III makes one 5760 x 3840. To get the crop he takes a sample out of the picture that is the identical size from each. Because the 1D X is starting with a smaller pic the crop will appear smaller.

The MP count comes in to play in that it changes the size of the picture.

Is this correct?