Pennsylvania: Specter, although this one should be close in the end. Hoeffel has not done the things necessary to make an impression on the electorate. The distracting presidential election has not been to his benefit, either. In an off-year this could have been a toss-up race.
North Carolina: Bowles is a known quantity from his senate run two years ago, and is currently even with, or (and more likely) trailing his opponent Richard Burr. This is a BIG disappointment. Bowles led throughout the long march and only fell behind in the waning days. Unless his campaign is aggressive in the final week, and there is some lift from a better-than-expected Kerry-Edwards performance in the state, it's a loss.
South Carolina: Another disappointment. Not AS big a disappointment though. This is, after all, South Carolina. But after delivering a series of gaffes that would have had him pulled from the race in any state above the Potomac River or bordering the Pacific Ocean, and badly damaged in many others, DeMint, the Republican, appears to be pulling away from charmless rival Inez Tenenbaum.
Georgia: I actually think this will be closer-than-expected. Black voter turnout should be quite high. Still a loss, of course.
Florida: Castor will win this one running away. She's currently a few points ahead and, get this, her weak spot is south(!)Florida, where Kerry is polling exceptionally strong. He should improve her performance drastically there. Curiously, Castor is performing quite well in the conservative-leaning Tampa area, and in conservative stronghold north Florida. This adds up to a Bill Nelson-esque victory.
Kentucky: Though the DSCC trots out polls showing Bunning dead-even against rival Mongiardo, I firmly believe this is a 53-47 race that Bunning will win. The responsibility for this loss, I believe, belongs collectively to Dan Mongiardo and to the Democratic Party in Kentucky. Mongiardo for running an undistinguished campaign that has left him with poor name ID 7 days out from balloting, and the KY Democratic Party for not finding and recruiting better candidates to take on a clearly vulnerable senator.
LA: Vitter could very well tie himself tightly to Bush and pass, hand-in-hand, the the 50 +1 mark needed to make himself a senator, sans the runoff. If he wants to be a senator, he better, because if this race surprises me and goes to a runoff, he'll lose. That's the best way for Republicans to lose elections in LA. Maybe if Breaux got out there and convinced his conservative followers to cast their ballots for John or Kennedy, he might force a runoff.
Wisconsin: This MIGHT have been close if Feingold had little money, as he did in 1998. But he has a lot of money. And this one's appropriately over.
Colorado: This should be tight, but Salazar might very well run ahead of ALL the federal statewide contenders.
Illinois: Obama. Democrats might almost be as excited as Kenyans.
California: Did not live to the hype.
Washington: See California.
South Dakota. Daschle. He's battle tested. But, then, so is Thune. The difference: Dashcle has won his battles, Thune lost his. And Thune has only faced one true battle in his entire career. The presidential race makes this one that much more tougher for the majority leader, but I'd be very surprised not to have him back next year.
Oklahoma: This one will be unnecessarily close, solely because of the presidential race. Bush will win here in a blowout beyond all blowouts, and could end up dragging Coburn across this finish with him. Carson is the smart money, though. His campaign has adeptly avoided the Kerry label, and he has run well in the rural tier of Oklahoma, where (to my astonishment) Democrats pull many, if not most, of their votes. Cities and especially suburbs in OK are ultra-conservative.
Alaska: Knowles has consistently led. I love consistency. Consistency is the surest sign of victory. Knowles.