The polls are all over the place in races for the Senate – which means they are useless. I have tried to explain this before, but if a segment of people make a major shift the polls do not necessarily pick it up. The fact that polls are all over the map, swinging ten plus points in only weeks, just tells this engineer the models are bad and something is happening. Pollsters rely on historical trends to extrapolate the small samples into statewide respresentations. Any small errors in their models can result in being off widely. And then there is the constant issue of reaching the voters in this age of blackberries, cellphones and voice-over-IP. I am just not buying the pollster are gaining accuracy as the technology base shifts. Overall the trend is clear, the Reps are pulling even as the races tighten.

Virginia: Anyone thinking VA is still in contention is not thinking straight. Polls always under represent the Cons by 3-5% in this state – always. I would give Allen a good chance down three. He is up five. This is a likely Rep hold (and off my watch list as is Arizona – a one time ‘at risk’ state).

Maryland: If there is any place to not trust the polls, this is the state. The black community in Maryland doesn’t hate Michael Steele. They just don’t. And they have been pining for a major black leader in the State and in the Congress for years. They, like us in Virginia, are seasoned political watchers. They will say anything to a poster to retain some connection with their Democrat roots. They will pull the handle for Steele. Cardin needs to be up more than 5% in polls for me to believe he has a chance. His base is just not there. Rep pick up.

New Jersey: Again, I think the polls are measuring some return to the Democrat Menendez, but New Jersey is fed up with their pathetic leaders and Keane will probably squeek this one out. Menendez is to far down for an incumbant and a Democrat in a dark blue state. Rep pick up.

Ohio: Dewine has brought this race back to even, surprisingly. Again, unless the Democrat is ahead by more than 5% points I think the Rep will win (given their superior GOTV abiltilies). Rep hold.

Tennessee: Same thing. Ford is a strong candidate, but Tennessee is a fairly red state and he needs at least a 5% lead in the polls to have a chance. Rep Hold for now.

Missouri: Same thing. Talent is bouncing between a tie and a 3% lead. If the polls stay in this range then Talent will pull it. In 2002 Talent went into the election with a 4% lead and sqeaked out a win. I would bet if he goes in with a 3-4% lead again he will win again. Leans Rep.

Pennsylvania: This is the best shot at a Dem pick up outside Rhode Island. I keep thinking this is a lost cause but insiders don’t seem all that panicked. I still have to put this in the leans Dem category.

Rhide Island: Tough call. I would say this one is a toss up. But let’s give it to the Dems since I think if it had to tilt, it would tilt left.

Montanna: Another one which should be going to the Reps, and may. Conrad Burns may not be the favorite son, but the state knows that a Democrat Congress is much, much worse. As lng as Tester is above 5% (which he is) then this is a Dem pickup.

If my math is right I see two Rep pick ups and 3 Dem picks for a one seat loss in the Senate. My earlier predictions in the summer was 1-2 Rep pick up, and that could still happen if we see Washington State get tight. I am watching PA as the bellweather. If Santoruim wins it will be on a wave of rejecting Democrats and the Reps could pick up 1-2 seats.