The impact of gerrymandering on Congressional districts makes the use of the NE/ME method too risky - it could lead to States which are currently very sensible about their redistricting, being targeted by national parties for gerrymandering.
Gerrymandering for presidential elector purposes could be harmful for congressional elections. You want a number of narrow majorities for your party to be effective. But this could put your representatives in jeopardy in close elections, or make it harder to knock off an incumbent of the other party.
Even if you wanted proportional apportionment of electors, you wouldn't want to use the calculation that was used in Colorado.
If a state had 6 representatives and 8 electors, then under a rational system (St.Lague) a candidate would get 4 electors with between 43.75% and 56.25% of the vote. In a reasonably moderated state, you would usually have a 4:4 split, unless there was a national landslide, in which case it wouldn't matter how the state split.
And if a state had 7 representative and 9 electors then you can present a nice example showing how if a state had a 55:45 popular vote split, then it would have a 5:4 elector split, which is also 55:44. But any victory from 50+:50- to 61:39 is going to produce the same 5:4 split.
For a different example, Nebraska would be ignored because it would be locked in 3:2, with almost no chance of 4:1 or 2:3 split, while Idaho might get attention because there would be a chance for a 3:1 or 2:2 split, if the Democrats could keep within 25%. But when Idaho gains a 3rd representative, and Nebraska lose its 3rd, the situation will be reversed.
But the proposition in Colorado didn't even use a rational scheme. It used an ad hoc scheme that would round each candidate's number of electors independently (if the fractional apportionment was greater than 0.5 or not). If this resulted in too few electors, then any extras would be given to the leading candidate. If rounding resulted in too many electors, then any extras would be taken from the trailing candidates.
For example: 44:43:6:4:3% popular vote; 3.52, 3.44; 0.48; 0.32; 0.24 raw electors; 4:3:0:0:0 after rounding. The leading candidate is given the 8th elector, disregarding that they had already been rounded making the final split 5:3.
Or: 44.2:44.1:11.7% popular vote; 3.54, 3.53; 0.94 raw electors; 4:4:1 after rounding, and the extra elector is taken from the last place candidate, even though they he had almost 1/8 of the popular vote.
This is not the method proposed in Colorado. But even under St. Lague, any winning percentage between 50% and 70% produces the same 3:2 split.
Oddly enough
Colorado would have split 5:3.
More interesting is whether more voters would have voted for Nader if they thought that they could gain an electoral vote, or not cost Gore the entire state delegation (see for example Washington, Oregon, Colorado, Califormnia, Minnesota, Wisconsin, New Mexico, Illinois, and Ohio). Could Georgia and North Carolina have kept Nader off the ballot, if the number of petition signatures was anywhere comparable to the number of votes needed to get an elector?
If 1% of those who voted for Gore, switched to Nader, Bush has a popular vote victory. Would this have been reflected in the distribution of the EV?
Or notice that the 25th EV in Florida (and 270th overall) would have hinged on the recount, and how many other recounts would have been triggered? Picking up 5 electors in NM was inconsequential in 2000, but picking up one might have made the difference under a proportional system. And Bush was 1519 votes short of an 11th elector in Ohio.