?

Log in

No account? Create an account

Rhetoretician -- Fiction etc.

Final Pre-Election Map and Predictions

Final Pre-Election Map and Predictions

Previous Entry Share Next Entry
Embarrased Chimp

Here's the last electoral map before the election, plus my "states to watch" list and my own (ulp!) predictions for the outcome.

 

I could actually do better by waiting for the next set of polls to be released later tonight and early tomorrow, but enough is enough.

 

 

 

Here's the map as it exists tonight, based on polls completed as recently as yesterday:

 

 

 

You can probably get a bigger view by clicking it.

There have been changes in five states, all visibly in McCain's direction: Georgia, Pennsylvania, North Dakota, North Carolina and Indiana. (Actually there have been movements in 12 states toward Obama, and in 21 states toward McCain, but only these five are enough to cause a color change.)

 

 

Things to Look For

 

Thirty-seven states (including DC) are so strongly for either Obama or McCain (10% margin or more) that they're essentially a sure thing. For Obama, this adds up to 238 electoral votes; for McCain, it adds up to 132.

 

The remaining 168 electoral votes, the votes that will decide the election, are contained in the following fourteen states:

 

State

EV

%O

%M

Current Lead

Recent change

New Mexico

5

53

44

Obama by 9%

Flat

Colorado

9

52

44

Obama by 8%

Obama up 2%

Pennsylvania

21

52

45

Obama by 7%

McCain up 4%

Nevada

5

50

44

Obama by 6%

Obama up 1%

Virginia

13

51

46

Obama by 5%

McCain up 1%

Ohio

20

49

47

Obama by 2%

McCain up 2%

Florida

27

49

48

Obama by 1%

McCain up 3%

Missouri

11

48

48

Tie

Flat

North Carolina

15

48

49

McCain by 1%

McCain up 2%

North Dakota

3

46

47

McCain by 1%

McCain up 1%

Arizona

10

46

48

McCain by 2%

Obama up 2%

Indiana

11

47

49

McCain by 2%

McCain up 2%

Montana

3

46

49

McCain by 3%

Obama up 1%

Georgia

15

47

52

McCain by 5%

McCain up 2%

 

 

I've run every combination of these 14 states through a spreadsheet (there are 16,384 of them, thank you), and made the following observations:

 

1.     Since Obama needs only 32 additional electoral votes to win the election, he begins with a considerable advantage. There are only 285 combinations (1.7%) which allow for a McCain win.

 

2.     While we're on the subject, there are actually 33 combinations (0.2%) in which there is an electoral tie, which means that the election would be decided by the House of Representatives (almost certainly an eventual, if painful, win for Obama).

 

3.     If Obama wins any six of those fourteen states, he wins the election. Period.

 

4.     There are only 11 scenarios in which McCain can win the election without winning Pennsylvania, only 13 scenarios in which he can win it without Ohio, and only three (3) in which he can win it without Florida. If Obama wins any one of these three, the election is probably over.

 

5.     There are no individual "must win" states for Obama.

 

So, Tuesday night, watch those fourteen states, and count how many Obama and McCain each wins. Look very carefully at PA, OH and FL.

 

(Nate Silver of fivethirtyeight.com says that Virginia is crucial too, but it doesn't look quite so crucial to me.)
 

 

My Predictions:

 

Based on the current lead and the trends in each state (see above), I predict that:

 

  1. Of these fourteen states, Obama will win New Mexico, Colorado, Pennsylvania, Nevada, Virginia and Ohio, which collectively have 73 electoral votes, for a total of 311. Of these six states, the one I’m least sure of is Ohio.

 

  1. McCain will win the other eight states, Florida, Missouri, North Carolina, North Dakota, Arizona, Indiana, Montana and Georgia, which collectively have 95 electoral votes, for a total of 227. Of these eight states, the one I'm least sure of is Missouri.

 

This prediction gives Obama fewer electoral votes than Nate Silver's model, which currently projects 346 EV for Obama. But Silver's own simulations show that 311 is the number that comes up most often.

 
Here's what an electoral map reflecting those predictions would look like:


 

Having said all this, please remember that we're dealing with imperfect data. You could write a book about the limitations on opinion polls to predict the outcome of an election, and that’s probably especially true this year, for various reasons.

  • Wow...I thought for sure McCain would lose Florida. The heart of the US is still red. Reminds me of the old (?) Chinese communist anthem.."the East is Red"

    You know, I could see a senerio in which Obama wins the popular vote, but McCain wins the electorial and the election. Sound familar?
    • I could be wrong about Florida; Obama's still in the lead there. But the movement towards McCain has been so pronounced that I think he has a good shot at it. Also, I'm tending to give McCain the benefit of the doubt in states where the vote is a tie or infinitessimally close (e.g., Missouri).

      That scenario you suggest is possible, but unlikely.
  • Thanks. This is very interesting and somewhat reassuring. My general feeling is that if Obama wins VA and PA it is pretty much over for McCain. Once I see those come in I may be able to relax a bit.
    • VA and PA would be the whole election; together they're 34 EVs, which puts Obama over the line. But McCain has been putting everything he's got in PA, and made impressive gains in the last two weeks; he could still win it.

      • You are, of course, right. I am just hoping. :sigh:

        It would be nice to be able to just start celebrating early!
  • All of it makes my stomach hurt. But Starbucks is giving out free coffee and babeland is giving out free mini-vibrators. I'm going for both and not turn on the TV till 11:00 pm!
    • You know, Shelly, you have just managed to take my mind off the election for the first time in a week. My goodness...
  • The presidential race seems safe enough now, but that's not a reason to be complacent and not vote. And I'm still losing sleep over CA Prop.8...
  • And on an entirely different note... my hubby has just told me that there's some stat that supposedly goes back eons, and if the Washington Redskins win the weekend of the election, the party who has the presidency keeps it. If they lose, the other party wins. Guess which team is losing tonight? ;-) Let's see if it holds true! And if it does, heck, maybe the parties oughta just let the football teams decide it and save the money and aggravation of campaigning ;-)
    • Didn't work in 2004, though.

      There are lots of predictor-myths like that one, some of them based on reality. Taller candidate wins, candidate with longer name wins, incumbent party wins if GNP has grown by more than x% during previous four years, Missouri predicts the result, Missouri and NC together predict result, one particular precinct predicts the result...

      Did you know that the most popular given name for presidents is James?
      • I think I'd heard that before (about James), but can't remember for certain.

        And yeah, there are plenty of "myths" of all sorts about elections. It's always interesting how the people who "report" on things like that so often leave out the times when it's incorrect, too! I'm interested to see how much you got right this time, though... I'm betting you're right on a good portion of your predictions!
  • Happy election! And may it be a happy result for the rest of the world, too...
  • I'm not sure I have quite as much insight into the election as you do, but I've taken an amateur crack at it and posted my own map based on the interactive one at the New York Times. We come out quite similar in our predictions, though I swear I based mine on my own personal hunches! Have fun this eve.
Powered by LiveJournal.com