# The 90-50 Formula For House Swing Projection

Like most of you here, I’ve kept a watch on Neil Stevens’ Swing-O-Meter and his various poll-based projections. (Always a worthwhile read, Neil.) And like many of you, I also have enjoyed watching more and more names mount and shift right on Cook’s Political Report. Being a numbers guy myself, I thought it would be fun to create my own gauge for what kind of swing the Republicans might end up with this time around, so I have come up with the 90-50 formula.

We start with the current version of Cook’s House Projections. As has been noted before, Cook is notoriously hesitant to shift any race with an incumbent beyond “Toss-Up”, which means that when he lists a race as such, there is much more than a 50% chance that the other side is going to win the race. And if you believe what you read from many other sources, many of the “Lean D” races are quite competitive, with some polls indicating a Republican lead. As for the Republican side, Charles Djou’s seat is at worst a toss-up for us (recent polls show him leading), and there is no seat that is currently listed as “Lean Republican” that looks like an upset. So we’ll leave those out.

So the formula works out like this: 90% of “Democratic Toss-Ups” will shift, 50% of “Lean Democratic” will shift. Add to that any seat that Cook has already placed on the other side, and you have your number.

So let’s see what we end up with:

Democrat Toss-Ups: There are currently 46 of them. 46 * 90% = 41.4, so we’ll round down and say 41 of these seats shift.

Leans Democrat: 28 Democrat seats fall into this category. 28 * 50% = 14 seats.

R seats leaning D: There are 3 of these (IL-10, DE-AL, and LA-2).

D seats leaning or likely R: There are 17 in the “Lean R” column, and 6 more in the “Likely R” column, for a total of 23.

So to sum it all up:

Toss-ups: 41
Leans: + 14
R to D: – 3
D to R: + 23
——————-
Total: 75 seats

That seems like a good number to me. Feel free to play along at home when Cook’s next report comes out.

• distantvoter

If you feel less optimistic, you could say 75%-25%. You would STILL get a gain of 56 seats, which exceeds 1994.

We just don’t know what the turnout will be, and whether conservative Ds and independents will get scared by Democratic tactics and lies. It’s possible 75 is too low. Cook has been moving lots of races in our direction with every update, so if the percentages hold, it is likely to be better than his current ratings would indicate.

• IJB

About the best thing you can say for the Dems in regards to early voting is that Indies aren’t turning out very much. But I’m not sure that helps them all that much, as I don’t think Indies are big ‘early voting’ types anyway.

But Republican turnout is up across the board in early voting, even coming close matching or exceeding Dem turnout in some surprising states (e.g. NV, NC, ME(!), FL(!!)).

Then you start hearing anecdotal stories about how “Dem early voting in Chicago is way down which is worrying IL Dems” or how GOP early voting is way up in PA. Etc.

So far, I don’t see anything that’s helping to mitigate this ‘wave’ for the Dems. If anything, it’s the opposite.

• distantvoter

RCP has a link to a story that early voting is A) down IN Chicago and B) up OUTSIDE of Chicago (in the suburbs).

That is the scariest of the tossups to me. In Illinois, I think you probably have to win by about 3% statewide because of how they’ll cheat in Chicago. So that kind of turnout can put Kirk over the top. Not that he’s my favorite candidate, but he’s far better than the alternative.

And if we get that kind of turnout boost, who knows what happens in the House?

Does anyone really know how many Democrats are vulnerable?

• http://www.hakubi.us/ Neil Stevens

I saw this diary title and thought to come in to pump the Swingometer in the comments. Too late.

• froster

polling showing them down. If we could only get some more polls…

• distantvoter

It’s just a little hard to be patient, isn’t it?

• IJB

If you poke around Cook’s site, you find this page with an interesting table – an academic looked at the success of Cook’s ratings over the years (for his c. Sept. 1 ratings of a given House election year), and put them in a handy table.

So I’m going to apply the numbers in that table from 1994 (and I don’t think I can fit that table into this post, or I would have reproduced the 1994 table-row here…) to Cook’s Sept. 2, 2010 House ratings, and see what I get.

Note: This is about the most *conservative* projection you can get from Cook’s own numbers!

For the purposes of this analysis, let me lump all the GOP-held or -leaning seats together (remembering that there were, in fact, 6 held-Dem seats in the “Lean/Likely GOP” categories back on Sept. 2), and just apply a “flat” 95% “win for the GOP” percentage to them (which is conservative – 98% might have been a more likely choice…):

If I do that, I calculate the D’s will win just *one* of 21-GOP leaning seats on Sept. 2. Let me be even more generous, and give the Dem’s *2* of these (say, LA-02 and IL-10). That leads to a net +4 R from these categories.

Now, according to the table cited above, in 1994:

The GOP won over 80% of the “Dem Toss-Up” seats in 1994. On Sept. 2, 2010, there were 42 “Dem Toss-Up” seats – 80% of 42 is about 33-34, so lets call this +33 R.

The GOP won about 30% of the “Dem Lean” seats in 1994. On Sept. 2, 2010, there were 29 D-held (and 1 R-held) “Dem Lean” seats – 30% of 29 is about 9 (counting -1 for the R-held flip elsewhere), so that nets +9 R.

The GOP won about 15% of the “Dem Likely” seats in 1994. On Sept. 2, 2010, there were 25 D-held “Dem Likely” seats – 15% of 25 is about 4, so that’s +4 R.

And, finally, the GOP even won about 1% of the “Dem Solid” seats in 1994. On Sept. 2, 2010, by my count, there were 153 “Dem Solid” seats according to Cook – 1% of 153 is between 1-2, so lets call this +1 R.

So, adding all of this up:

R to D: – 3
D to R: +4
Toss-ups (D to R): + 33
Leans (D to R): + 9
Likely (D to R): + 4
Solid (D to R): +1