By how much could a candidate lose the popular vote and still win the Presidency? Let’s say that every state the winner won he only won by one vote. Isn’t it possible then that the loser could win a landslide in the popular vote yet lose the election? For instance, a candidate could recieve every single popular vote in the west (except California) and lose California by one vote and end up behind. What’s the most votes someone could (theoretically) get yet still lose?
Couldn’t this actually end up in the multi-millions?
If you want to go all theoretical, say only one person votes for candidate “A” in California, New York, Texas, Florida, Pennsylvania, Illinois, Ohio, Michigan, New Jersey, North Carolina and Virginia and you’d get the requisite numbers in the Electoral College with only 11 votes while candidate “B” could get a 100% turnouts and 100% vote in the remainder. So multi-millions, is theoretically possible.
As to the probability of this occurring, I’ll leave that to someone else. 
A presidential candidate can win the electoral vote by winning as few as 11 states. The 11 most populous states have 271 electoral votes (California, 55 electoral votes; Texas, 34; New York, 31; Florida, 27; Illinois, 21; Pennsylvania, 21; Ohio, 20; Michigan, 17; New Jersey, 15; Georgia, 15; North Carolina, 15), with 270 votes necessary for an electoral-vote majority.
Theoretically, if a candidate won those states by the narrowest possible margin with extremely low turnout, and lost the other 39 states and the District of Columbia by an overwhelming margin with extremely high turnout, then he or she could win a clear electoral-vote majority while winning less than 1 percent of the popular vote.
A little more realistically, let’s use some actual numbers from the last presidential election (but still using the latest available electoral-vote allocations, from the 2000 Census, which were not yet available during the 2000 election). Let’s say that one candidate wins the 11 most populous states by the same margin by which Governor Bush won Florida, the state with the narrowest margin between the two major-party candidates (0.009%). Let’s say that the other candidate wins the other 39 states and the District of Columbia by the same margin by which Vice President Gore won the District of Columbia, the widest margin between the two major-party candidates (76.2%). Finally, let’s assume that each third-party candidate gets as many popular votes as in the 2000 election. The result: the candidate winning the 11 most populous states receives 27,767,084 popular votes in those states, and 4,368,393 popular votes elsewhere, for a total of 32,135,476 popular votes out of 105,363,298 votes cast. That candidate wins the election with 30.5 percent of the popular vote.
My numbers for electoral votes are based on the allocation of electoral votes based on the 2000 Census as posted on the Federal Register’s website. The numbers for popular votes are based on the 2000 Presidential Election: Popular Vote Totals as posted on the Federal Register’s website.
Multi-millions indeed.
I think you (not you personally) could lose by 39,321,000 votes out of 205,815,000, but that’s assuming that of voting age votes, and that all the felons are kids (like having your candy stolen by a baby).
Well, of course, the highest number you could lose by is 116,898,961 out of a total of 116,899,000, or in other words you won with 0,00003% of the votes cast; if you carried the same states (39 of them), but only one person turned up to vote in each of those states (meaning that most of the electors didn’t bother to vote themselves) and in the others all the voters turned up and voted against you.
Arrived at the answer by dividing the number of potential voters for each state [small](source, see hicack below)[/small] by the number of electoral votes, to find the states with the lowest number of voters per elector (or whatever they’re called).
Carrying the states of …well, the list of who you don’t have to take is shorter: Ohio, Virginia, Illinois, Michigan, North Carolina, New Jersey, New York, Arizona, Georgia, California, Texas, North Carolina. Taking all the states except those will get you to 270, exactly one vote ahead.
You don’t need all votes in the states you carry, though, you can afford to lose one vote less than half of the voters needed for one elector. In the states you don’t win, you’ve lost all the votes, since we’re trying to find out how many you can possibly lose by.
[hijack]How can Montana and Alaska have over 100% of the potential voters registered to vote?[/hijack]
Correct me if I’m wrong, but since small-state votes “count” more per person, then if you wanted the largest difference possible, you’d give the small states to the winner (of the electoral vote), and the more populous states to the loser. No?
Achernar, I’m one step, or half a second, ahead of you. =)
Note that my equation didn’t have any third party candidates, either - or at least they didn’t vote for themselves.
Not necessarily. If you take woolly’s approach, then you want the lowest possible number of states–and thus only the most populous states–in the electoral-vote winner’s column.
There are some states that require electors to vote for the candidate they represent, but some of these may not void their vote but only penalize the elector. And there are some states that do not reguire the electors to vote in any fashion.
So it may be possible to win without any popular votes at all!
Another possibility is a very close electoral election, and a third party candidate’s name appears on only one state like New Hampshire, wins the state, throwing the election into the Congress, winning there.
Hell, that happens just about every election in Chicago!