Final Pre-Election Map and Predictions
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:
- 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.
- 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.