we will struggle greatly to hold the Senate.  This cycle bears some resemblance to 2006 in the sense that there a large number of Democratic incumbents below 50, but with the lead.  In 2006 virtually everyone of the Republican incumbents went on to lose who were below 50. 

If that holds for '10, we will lose the following:

Arkansas

Colorado

Nevada

Wisconsin

California

It is tempting to note that with respect to Wisconsin and California that the seats are being fought on favorable ground to the Democrats.  This didn't save Brown, and my guess is the race is going to get worse for Democrats if the economy continues to struggle.  Another incumbent, Murray in Washington, is flirting with 50%, and I would rate as a tossup right now given the strange SUSA polling.

In addition to these 5, there are a number of open seats.  The Democratic seats in play are Indiana (almost a certain loss), Illinois (a true tossup) , Delaware and Pennsylvania.  Delaware and PA are moving in opposite directions, but my guess is both seats are losses. 

The only Democratic pickup that looks like a reasonable shot is Kentucky, and Paul just might lose the race.  The others are trending badly (Ohio, Missouri, NH), though we may have a shot in NC. 

And then there is Florida - a wild race if ever there was one.  Assuming Meek wins the nomination, the real question is whether Crist can hold onto enough GOP support to overcome the almost certain loss of African American votes.  I think predicting Florida at this point is next to impossible - but three way races are inherintly unstable.

BTW - my house model predicts the Democrats will lose 63 seats in the House.  The model is based on the models used to predict swings in British elections, and doesn't account for things like underfunded challengers.  In this sense it is almost high given the predicted 5 point advantage in the generic ballot.  But then there is no reason to think this advantage won't grow.

The model is here (I use the Cooke PVI if the incumbent isn't running)

https://spreadsheets.google.com/ccc?key=0Aq5zx_FMTYpxdGxhUWQ1WW5mbUl6cWdQVElfcjVjdkE&hl=en#gid=0

This election is most similar to 1992.  In 1992 the Recession technically ended well before November, and yet Bush's approval rating continued to fall as unemployment remained high.  The lesson to learn is that GDP is irrelevent in politics, and if people view unemployment as likely to say high for a significant period, that view will steadily erode Democratic support.  Reagn's approval rating reached 36% in the depth of the '82 Recession, and I would be surprised not to see Obama's fall in that range.

Lastly, I will say that there is huge volatility in these races. 

I have about politics is this: people are far more likely to underestimate rather than overestimate the volatility in a given cycle.  The model I developed, crude in some ways, reminds us of the fact that the generic ballot is showing an average of a <b>16 point swing from 2008</b>, and that if it were to hold the size of the lanslide will exceed 1994.

...a strong rec for excellent writing.

for alomost a year--still sticking with 50-60 seat pickup with the GOP, though I'm now leaning to the higher end of that range.

I think one has to back to the 1920s to find the Dems under 200 seats in the House; I'll have to check that.

As for the Senate, I no have a 50-50 Senate, down from 52-53 Dems.  Crist won't win as Meek draws too may Dems away from him; in fact I think Crist will finish third, despite what Q says.

Senate control comes down to:  WA, CA, CT, NV, CO, IL and WI.  I think GOP will not win all of them, but could win four or five.  Yes I've included CT, and NC would be on this list if the DSCC would put some effort into that race.  They must not like Elaine Marshall.

The only Democrat who could have won Indiana was Evan Bayh.  Democrats would have been better off just nominating Tamara D'Appitoo(sp) and have Ellsworth stay in his House seat.  Coats may be flawed, but he is in a state that hates Obama and wants a Republican Senate.  Nothing will stand in the way of that.