This years' races are particularly tough to call, and I'm really only 100% comfortable with two of these five calls, and only one of the margins.  That said:

NJ: Christie 46 Corzine 44 Daggett 8.  The polls pretty much consistently showed a last-minute movement toward Christie.  Whether this is due to some kind of Halloween/World Series effect remains to be seen.  Moreover, Dems probably have a superior GOTV effort here.  I think this is about the most likely outcome, but I don't think Christie's likely to go much higher in his win, and there's a great chance that he goes lower.

ME rejects the legislature's approval of gay marriage, 51-49.  Normally I'd say a ballot initiative that isn't consistently above 50% on the "yes" option is going to lose.  But I think these gay marriage initiatives are different, and the opponents are unusually primed to turn out and vote.

VA: McDonnell 56 Deeds 44.  We'll see what happens in Northern Virginia.  Don't think McDonnell will go much lower than 55 percent, and could go as high as 60.

CA-10:  Garamendi 56, Hamer 43.  With one poll out there, its hard to make any predictions.  I'm actually particularly interested in this race, for reasons I'll go into later.

NY-23:  Hoffman 48, Owens 46, Scozzafava 6.  Polls are all over the place here, but I think Hoffman pulls it out in the end, but by a closer margin than many are predicting.  Incidentally, this is exactly how conservative candidates go to Congress in New York -- a liberal Republican and a Democrat split the left-leaning vote.  Buckley and D'Amato (who was also on the Republican line) went this way, and I believe (and I'm sure to be innundated with e-mails telling me I'm wrong) that this is how William Carney was first elected in NY-1 in 1978.

--------------------------------------------
Follow the RCP Blog on Twitter.
Become a fan of RCP on Facebook.
--------------------------------------------