I keep looking at the election numbers, trying to make sense of it all, and the more I look and count and add, the more a couple of things become clear:
1. The absentee vote wasn't just about Ed Lee. Clearly, the Lee forces got their troops out and did an absentee drive, but the total absentee votes for mayor (62,446) were about the same as the total votes for district attorney (63,354) and most of the propositions.So the people who voted early voted the entrie ballot.
2. The election-day votes were so dramatically different from the absentees that several factors had to be at work. One of them was the phenomenal campaign for John Avalos, which moblized thousands of people and demonstrated how much of a force progressives can be. Keep in mind -- Avalos, who had no independent expenditure groups and less money than many of the other candidates -- actually came in first on election day. His team worked hard and smart and pulled off a near miracle.
3. The drop-off in support for Lee between the absentees and election day suggests that his popularity was, indeed, declining fast in the past few weeks. The voter fraud scandals  had something to do with it, but so did the attacks on Lee by the Herrera and Yee campaigns and by IE groups supporting those two candidates. If Lee hadn't been so far out in front a month ago, he might not have won. As it is, if he holds on, it won't be with the kind of mandate he would like to claim.
When the Department of Elections runs the first pass at ranked-choice voting, we'll get a better idea of how much Lee's support has fallen; RCV won't be such a big deal with the absentees since Lee got so many of those first-place votes. The election-day votes will be more telling; when Adachi, Yee and Chiu are eliminated, where do those seconds go? How many will go to Lee -- and how many will go anywhere but?