I'd like to refer us back to this time in '08, when Mark Begich trailed Ted Stevens by just over 2pt (48.22%-46.20%, a margin of 3,986 votes)
http://www.dailykos.com/...
Begich ended up winning 47.8%-46.5%, with a margin of 3,953 votes, gaining 7,939 votes of margin after Election Night.
Where would that get him against Sullivan? 210 votes short. This is still definitely worth watching... also, AK is showing 5,000 FEWER early votes, despite all of the new early voting locations in rural Alaska, so I'm betting those votes look good for Begich, especially since we're showing basically no early vote in those areas at the moment.
https://drive.google.com/...
And for a good apples-to-apples comparison, Stevens led in-person election-day voting by about 6,500 votes; Sullivan leads (with 100% reporting) by just under 8,200. So if Begich underperforms by the same raw amount (and higher %) with absentee/early once the dust settles, he would still squeak by with a win.
Let's buckle in for the next few weeks.
5:57 PM PT: UPDATE: Found a bit of bad news for Begich in looking at '10 for comparison... % of ballots cast absentee fell from 34% of in-person election-day votes to 22% in '10... that would translate to about 33k absentee outstanding (plus questionable/remaining earlys) instead of 57k (if the proportion matched '08). Any ideas why absentees less popular in midterm year? Anyhow, at this point, Begich had better hope that there's a lot out in Inuit areas AND that they're breaking more blue than they did in '08-- I know some have speculated that the rural Alaskan vote is pro-incumbent in general because of federal largesse, but this makes the hill seem much steeper to climb.