Another is the nature of determining a winner. One success of a system like the College is that a plurality vote winner will likely be magnified to a majority winner. That has the positive effect of producing a winner in single round of voting, yet blocking really weak plurality winners. It's often been noted that Clinton never broke 50% of the vote, and in 1996 the House would have sided with Dole. A runoff would have been a large and unnecessary expense since Clinton had a substantial lead over Dole. The Electoral College proved superior to a straight popular vote majority system in that election.
Muon is arguing for the electoral college using the opposite criteria as Ensworth. Ensworth was arguing that the electoral college is good because it is more likely to be close, and make your vote count, Muon is arguing that the electoral college is good because it is less likely to have a close election. Can't have it both ways, guys.
Anyways, the electoral college gave the win to 3 popular vote losers. I'm not counting 1824, which was just weird.
I would count both 1824 and 1876 as exceptions due to othe circumstances. That leaves 1888 and 2000 as examples where the Electoral College failed to identify the popular vote winner. In both of those elections the top two popular vote-getters were within 1%. In the other close popular election (1884) the Electoral College selected the popular vote winner.
I'm not sure I see Emsworth and I arguing forom exactly opposite sides, but I will contend that the mechanism of the Electoral College should mathematically increase the percent margin. This could be expected from the statistics of any random partition of the orginal sample. Once the popular vote margin exceeds a small amount it becomes increasingly unlikely that the partitioned votes (the College of state votes) won't magnify the difference.
In 9 elections since 1860 the popular vote winner had less than 50%, and more than a 1% margin over the second-place candidate. In 8 the Electoral College gave a comfortable majority to the winner. In 1876, only fraud prevented that election from also comfortable electing the winner. The College was successful in avoiding a runoff election for those 9 cases.
One change to the College I would favor is to bind the electors to their candidate. It seems that the voting freedom of the electors has lost most of its utility since all states have gone to direct election of the electors.