In a way, it's good that some of the polls came out in the last day or so, such as NY-20 and CO-5, came out. Because I think people were getting carried away by the suspect polls coming out of the Constituent Dynamics polling organization. Instead, I'd like to focus on the races Democrats have a realistic chance of winning from the Republicans.
I think there are 31 Tier 1 races. These are races because of the nature of the district, past performance or lack of an incumbent that the Democrats have the best chance of winning. These are also districts where the Democrat has been shown to be tied or in the lead in more than one poll with a track record.
CT 2
CT 4
CT 5
NY 24
NY 26
PA 6
PA 7
PA 8
PA 10
VA 2
KY 3
KY 4
NC 11
FL 13
FL 16
FL 22
TX 22
OH 1
OH 15
OH 18
IN 2
IN 8
IN 9
IL 6
IA 1
MN 6
WI 8
NM 1
AZ 8
CO 7
WA 8
These are the races that we, as well as the DCCC, should be focused on in terms of money and attention. Assuming the Democrats do not lose any seats, winning half of these races would give us control of the House.
There are 7 races that I would classify as Tier 2. All these races either have a shaky Republican incumbent, an unusually strong challenger or a district profile conducive to a Democrat. In addition, they are races where more than one reputable poll has show the Democrat to be within the margin of error.
NY 19
NY 20
NY 29
NC 8
OH 2
CO 4
AZ 5
The remaining races people tend to discuss are Tier 3 longshots, in my view. These tend to be races where the Republican is coming off a devisive primary, such as ID-1, CO-5 and NV-2. Or where the incumbent Republican is in ethical hotwater but the district is strongly Republican, such as CA-4 and CA-11. Or districts which would favor a generic Democrat but the Republican is a popular long-time incumbent, such as IA-2, NV-3, PA-4, NY-25 and VA-10. This is not to say that the Democrat can't win some or all of these races. It's just not likely.