Using Charlie Cook’s historical & current ratings to predict next month’s midterm……

I’m not one who’s provided a race-by-race breakdown to predict the House elections.  But I finally decided to come up with my own rudimentary model, with Charlie Cook’s ratings as a guide.

My model relies on Cook’s late September ratings in 2008 and 2006, both wave elections, and compares them to the most recent Cook ratings now.  Relying on late September ensures an apples-to-apples comparison.

The hardest races to forecast are the disfavored party’s tossups and “lean” races in an anti-majority party wave.  Those are the races that decide whether the House flips.

What I found is interesting, and discouraging for us.

Cook’s 2006 ratings in late September had 18 GOP-held tossups and 16 GOP-held “lean R” seats.  Of those, 10 from each category flipped.  Also flipping were 6 of 19 “likely R” GOP-held seats, as well as 2 GOP-held open seats Cook already had flipping in late September.  And 2 “safe” seats from Cook’s late September ratings flipped, those being Boyda over Ryun in Kansas and Altmire over Hart in PA-04.  NO Dem-held seats flipped, and indeed in late September Cook had all Dem-held seats as lean, likely, or safe, with NO tossups.

Cook’s 2008 ratings in late September had 19 GOP-held tossups and 14 GOP-held “lean R” seats.  Of those, 13 tossups and 6 leans flipped.  None of 20 “likely R” seats flipped this time, nor did any safe seats.  Meanwhile, Dems had more vulnerable seats this time in Cook’s late September ratings, and 2 of 10 Dem-held tossups flipped as did the lean D seat of Tim Mahoney due to his late-breaking sex scandal, and in a runoff the safe D seat of Bill Jefferson due to his being a crook.  Also flipping but excluded from my consideration was Don Cazayoux’s seat, which I exclude because he won it as a Dem pick-up in a special election earlier in the year before losing it in November, and that makes it awkward to include in any count discussing 2008 gains or losses.  I note, too, that

Here’s the interesting thing per Cook’s late September ratings:  the total number of seats the Rs lost from Cook’s tossup and lean R columns almost perfectly matched the number of R-held seats in Cook’s tossup column.  In 2006, Cook listed 18 GOP-held tossups, and the GOP lost 20 seats total from the tossup and lean R columns.  In 2008, Cook listed 19 GOP-held tossups, and the GOP lost exactly the same number total from the tossup and lean R columns.

The difference in 2008 was that no likely R seats flipped, compared to 6 in 2006.  The reason for this is obvious:  the likely R seats in 2006 were much lower-hanging fruit than the 2008 likely R seats, since the remaining Republcan-held seats were much more conservative and safer after the Dems already had made big gains one cycle earlier.

Applying the same princples to 2008, Cook in late September had 43 Dem-held seats as tossups, and 31 as lean D.  If the election follows the same pattern as the previous 2 waves, we should lose 43 seats total from those 2 categories.  We also should lose all the Dem-held seats that Cook counts as lean R or likely R, and that’s 10 more.  That’s a gross gain of 53 for the bad guys.  But there are 4 GOP-held seats we should pick up by everyone’s predictions, seats that Cook lists as lean D or tossups, and that knocks down the net GOP gain to 49.

That would give the Republicans a 228-207 majority.  And sadly it’s a very reasonable prediction that lines up perfectly with ALL the published predictions out there.

Here’s where I think we either can have some confidence or where we’re deluding ourselves, election day determining which it is:  when I look at Cook’s “lean D” seats, it’s just really hard to see hardly any of them flipping.  My “feeling” is that we hold almost all of them.  And even on our tossups, I did a quick count and found 25 I’ll say are gone.  That adds up to total losses of 15 fewer than my rudimentary model would predict, and of course it means we hold the House with 222 seats for the good guys.

The optimistic seat-by-seat breakdown is essentially what conspiracy and StephenCLE and others are engaging in with their own breakdowns posted in occasional diaries here.

And I can see exactly how they get there.

But sadly history shows that a lot more tossups and leans flip in a wave, and that’s where we might find out we’re deluding ourselves.

I just hope our candidates and party committees continue hammering the opposing candidates and getting voters to reject enough of them to keep us at 218 on election night.  But I just don’t feel good about it.