Okay, the part of Kentucky on Central time is now closed. Votes are being counted.
Mongiardo and Conway are currently trading leads, with just 5 percent reporting. Note -- Conway is expected to do better in urban areas, and he's clearly mopping up in Jefferson County (Louisville). But still early, and still lots of votes left to be counted.
Update: With 13% in, Conway has a 47.2-40.0 lead, or slightly less than 5,000 votes. A third of Fayette is in (Lexington), which has given Conway 1,300 of his vote margin right now. But the big chunk of that lead comes from Jefferson, which is 30% in and has given Conway an 8,000-vote margin.
There's still lots of rural Kentucky left, so whatever margins Conway can rack up in Jefferson and Fayette will be critical in holding off Mongiardo's expected strength in the rural parts of the state.
(Note, this will be reversed in Pennsylvania, where Sestak will do well in rural PA, while Specter will do better in Philly and Pitt.)
Update II: As noted on the comments, Democratic turnout is MUCH higher than GOP turnout. With 17.5 percent of precincts reporting, Dems have had about 125,000 voters. GOP? Only about 82,000. And the GOP had a teabagger-fueled primary that was supposed to generate bigger turnout.
Also, Conway now leads 50.6 to 37.0, and 16,000 votes. BUT... Jefferson County is two-thirds in. This is where Conway will run up his margin and hope that it's enough to offset the vote from rural areas where Mongiardo runs stronger. So don't get too excited for Conway yet.
Update III: Rand Paul has a 21-point lead in the GOP primary. Safe to say, he's going to be the GOP nominee.
Update IV: Nate Silver:
Based on a crude extrapolation of the vote in from Fayette County (Lexington), Jefferson County (Louisville), and "rest of state" so far, I show final results of Mongiardo 214,701 (43.7%), Conway 214,075 (43.6%). It might be very close!
I bet Nate would say that this extrapolation is within the margin of error!
Update V: In Jefferson County, Conway has a 17,000-vote lead with 2/3rds in. In the entire state, Conway has a 16,000-vote margin, but with only 29 percent in. In other words, Jefferson is currently Conway's entire margin, and he's running out of ballots there far quicker than the state at large.
Update VI: Nate updates his crude calculations based on new data:
Conway's numbers are holding up better in other rural areas than they were in the Eastern portion of the state. The crude projection method I describe a moment ago now shows him winning 44.8%-42.7%.
Still within the margin of error!
Update VII: Guys, you do realize that if Conway pulls this off, we're going to have to go balls out to help him defeat Rand Paul, right? I know people are burned out and dissatisfied with our current Senate, but picking up a seat in KENTUCKY (of all places) in this political environment, against a nut like Rand, is too good to pass up and sit on the sidelines.