The counterargument is to look at current patterns of candidate visits. They already concentrate a huge amount of time, effort and money on a non-representative

The counterargument is to look at current patterns of candidate visits. They already concentrate a huge amount of time, effort and money on a non-representative sample of the US, because the electoral college system prioritizes that non-representative sample. Replacing one arbitrary non-representative sample with an arbitrary more-representative sample is not necessarily bad.