ubikcan predictions

OK, here are the ubikcan predictions for tomorrow's midterm elections. I don't have any special data or insights so of course this is just my opinion based on all the polling, the pundits, and especially the blogs at the local and national levels.

Overall
House to go Democratic, gain of 25 seats (15 needed)

Senate to stay Republican; Democrats gain +4 (6 needed)
PA, RI, OH, and 1 of MO, MT, VA (wimp!). OK, if pushed I predict VA.
No GOP pickups.

Specific races
GA-Gov: Purdue (R) (duh)
GA-04: Hank Johnson (D)
GA-05: John Lewis (D)
GA-08: Jim Marshall (D)
GA-12: John Barrow (D)

CT-Sen: Joe Lieberman (I)

FL-Sen: Bill Nelson (D) (duh, sorry Katherine)

MD-Sen: Ben Cardin (D)

NJ-Sen: Bob Menendez (D)

OH-Sen: Sherrod Brown (D)*
OH-15: Mary Jo Kilroy (D)
OH-18: Zack Space (D)

PA-Sen: Bob Casey (D)*

RI-Sen: Sheldon Whitehouse (D)*

TN-Sen: Bob Corker (R)

VA-Sen: Jim Webb (D)*

WA-Sen: Maria Cantwell (D)

*Pickup

Interestingly, polls can be quite wrong in that there may be 2-3 point difference between the poll and the actual result due to GOTV. Here's an insightful post by Kos for example. He's talking about the recent poll numbers in Virginia:

In 2005, the Democratic ground game swamped the GOP's -- even with Rove's full attention and stewardship. I have no reason to believe that this year should be any different.

Let's go back to the open 2005 governor's race in Virginia:

Rasmussen 11/4/05. Likely voters. MoE 4.5% (10/27 results)

Kaine (D) 49 (46)
Kilgore (R) 46 (44)

Mason-Dixon 11/2-3/05. Likely voters. MoE 4% (10/18-20/05 results)

Kaine (D) 45 (42)
Kilgore (R) 44 (44)

Surprisingly comparable numbers, both heading into the election with the narrowest of 1-point leads. The final results? And remember, the GOP went to the hilt with their 72-hour program.

Kaine (D) 51.72
Kilgore (R) 45.99

So final polls showed 45-44 but actual numbers were 51.72-45.99, and Kos attributes this to GOTV. So races, such as VA which are close, can be determined by the GOTV.