Well, first, to be clear, I never said bond lengths were the only factor - only that in this case, they were the reason iodine had so much less steric interaction, not that iodine is one atom while methyl is 4 atoms, as you said.
The I atomic radius is huge. Two eclipsed I atoms will have a huge steric interference because they have the same bond length. The steric interference is so small in the case in question because the I, though huge, is positioned more or less by itself above everything.
Same reason the 1,3-diaxial values for I are so small (when compared to H,H).