Notwithstanding the hilarious frequency response claim, I’m pretty sure the internal wiring to and from the internal portions of the RCA input and output jacks of the attenuator is likely going to be more bandwidth limiting than any external RCA interconnect(s).
While I’m not familiar with the internal construction of George’s Lightspeed attenuator, I would expect that what limits high frequency bandwidth is the low pass filter formed by the interaction of the impedance "looking back" from the junction between the series and shunt LDRs and the capacitance "looking forward" from that point.
Given the far greater length of an interconnect that would be usually be connected to the output of the Lightspeed, compared to the length of the wiring between the LDRs and the output of the device, that capacitance will almost invariably be dominated by the capacitance of the interconnect. Hence the statement I quoted above appears to me to be incorrect, and George’s statement that "interconnects are the determining factor here" is likely to be correct.
Perhaps, though, the statement I quoted above was intended to refer to the bandwidth of the interconnect in itself. If so, under any reasonable circumstances involving a home audio system I would expect that number to be so high as to be irrelevant to the bandwidth limitation imposed by the interaction of the capacitance of the interconnect and the Lightspeed’s impedance that I referred to.
Regards,
-- Al