Just a reminder: I don't support the use of a threshold to get at-large EV and would include all the CDs a ticket has won when it comes to allocating the at-large EV.
Yes, and a quarter (or just over a quarter) of the voters in one congressional district (or just over half half the voters of a candidate who would get just over half the vote in that district if everyone voted honestly) is just over 1/(4
n) of the statewide vote in a state with
n congressional districts if the total vote in in that district was 1/
n of the total statewide vote, which it usually will be approximately. In a two way race (not likely to happen in a presidential election in any state with a quasi-proportional method in place but anyway), one candidate would get an electoral vote assuming no overhang under the D'Hondt/Jefferson method for every 1/([e] + 1) or 1/((
n + 2) + 1) of the statewide vote he received in a state with
n seats in the U.S. House and thus
e electors. An overhang elector (even if that elector would really vote for one of the two leading candidates) would reduce the denominator for the remaining electoral votes by 1.
x/((
n + 2) + 1) - 1/(4
n) is greater than (
x - 1)/(
n + 2) if... well I have to get ready for bed but I'm sure there is a wide range of
x and
n such that a candidate's supporters could theoretically at least get an allied elector elected from a decoy list if they took full advantage of the instant runoff voting method used to determine the district winner while still having enough support left over to get an
xth elector (counting the puppet elector from the decoy list) that they wouldn't have gotten otherwise. A candidate could technically get a second elector in a state with four electoral votes this way if they had just over 50% support in one district and between 25 and 30 percent support in the other (again assuming the same number of votes are cast in each district), with the remaining vote going to one other non-decoy candidate. If it can theoretically work with only two non-decoy candidates than it could with more candidates, although a larger third-party vote might limit the ability for a candidate strong enough to carry a district under IRV to net an additional elector in this way. It would also be difficult to limit the decoy voting to one or a small number of congressional districts within the state (too much of this could cause the candidate to lose one of the statewide seats). But the fact that it is theoretically possible makes the method seem slightly flawed in my opinion. No voting method is perfect, however, and I appreciate defe07 for proposing this method which I have had fun pointing out this arguable defect of.