While Frogman's statement is certainly true, the converse is also true. Any information that is lost or distortion that is added by a downstream component or cable cannot be corrected by anything further upstream!
IMO the lack of consensus in this thread confirms what Philjolet said early on: "There are too many variables in personal stereos to have a fair consensus for comparison."
Many technical factors could be cited in support of that position, including the dependency of many cable effects on the input and output impedances of the components that are being connected, balanced vs. unbalanced interconnections, ground loop susceptibility of the components, the rfi/emi environment, use of feedback in the components, power supply design, AC power characteristics, the lengths of the cables, etc., etc.
Regards,
-- Al