Okay, we have a lengthy discussion in the comments below about polls and the state of the race (Is McCain surging? Is Obama holding steady? etc.) I thought I'd start a new post devoted to the horse race in the final two weeks. If people like, I can make this a daily feature.
First of all, here's a map that I consider to be the baseline for both candidates. Maybe you have quibbles about this. Maybe you think, for instance, that Indiana should be a toss-up or that Missouri should be for McCain, but I think these are reasonable. (The map, BTW, is taken from the LA Times electoral vote map)
I think this is a reasonable baseline. Obama wins all the Kerry states plus Iowa and New Mexico, where he has essentially never trailed (at least not since March). So the election comes down to seven states -- and Obama only needs to win one of them. (Nevada would result in a tie, but a tie goes to Obama for reasons we have previously discussed). So all that matters are the results in those seven states. As of today, according to the various trusted websites, here are the numbers:
| State | RCP | Pollster | 538 | 538 Win % |
| NV | +2.3 | +3.3 | +2.6 | 74% |
| CO | +5.4 | +6.7 | +5.6 | 89% |
| MO | +2.7 | +1.3 | +1.4 | 65% |
| OH | +2.8 | +1.7 | +2.2 | 72% |
| VA | +8.0 | +8.9 | +5.9 | 91% |
| NC | +1.5 | +3.8 | +1.1 | 62% |
| FL | +2.0 | +2.8 | +2.6 | 75% |
The "538" column refers to the fivethirtyeight.com projection, and the "538 %" column refers to 538's estimated percentage that Obama wins that state.
In any case, you can see what I'm getting at here. If you believe my assumptions, then McCain needs to win all of these seven states; Obama just needs to win one. And as of now, by any decent metric, Obama is leading in all seven of those states, in some cases by considerable margins.
Should I do this again tomorrow? Is this a helpful thing to look at?
10 comments:
If teh 538 probabilities are correct - and I have no clue how they got them - then McCains chance to win the election is a measley 0.000024.
My only quibble is with Colorado. I think that's now in Obama's camp by at least 5 points. This means that McCain can't win even if he gets all the tossups, he must also get one of the states away from Obama.
That is probably why he decided to go after Pennsylvania. If you have to pick something off to win, it, might as well go for something big.
Monkeyman --
That's only true if you assume the states are independent. But in reality, they're not. If McCain somehow wins Colorado (for instance), he will probably win the rest.
LTG -- I don't think that report is true. McCain has no shot at Pennsylvania. He's better off going for Colorado (which is what his campaign says he's doing).
Just a quick quibble with your map: what about North Dakota and West Virgina? Some polls in ND have Obama up, as with WV.
But I like your logic and your reasoning. Knock on wood it all works out.
Yes, the electoral college map looks great for Obama.
The concerns I expressed in the comments thread you talk about were about the national poll numbers which often lead the state polls. As it happens those national numbers are returning to a pro-Obama state so the changes in McCain's direction that were starting to emerge last week have stopped.
If McCain is pullling out of Colorado, it would be tantamount to a concession. If that's where their thinking is, I'd expect to see Palin to start publicly distancing herself from McCain and spending a lot of her time working on her own image with an Alaska audience in mind.
I'm doing my bit to help turn Ohio blue in this race. I spent last Sunday canvassing door to door in Portage County, Ohio. This area is one of the top 2 or 3 areas for "undecideds" in the entire state (according to the Obama campaign). They told us that every 14 contacts made would mean 1 vote for Obama. Given that I managed to hit 40 that's almost 3 more votes. I'm going back out this coming weekend and I've done some phone banking as well so... hopefully we can overcome the mouth-breathing, knuckle dragging racist element in the area.
Quick note: Although your map is taken from the LA Times electoral vote map, what you have pictured here is not their current scenario. The LA Times poll analysis shows five of your tossup states (NV, CO, MO, VA, and FL) as leaning to Obama. LA Times lists just four tossup states: OH and NC (your other two states) plus IN and WV. I think you just meant to credit LA Times with the graphics, but it sounded like you were citing them to support your baseline.
As to your question, Bell Curve... While I really appreciate the effort you took to compile the data from RCP etc., to be honest, I check all these sites regularly anyhow--and I think all of us on this blog check them regularly also. So I guess I am not sure it is worth your while to go through the effort to summarize the statistics for us in tabular form. Much more valuable, I think, is the sort of meta-analysis you provided.
For example, while I would probably shift IN from red to tossup, I fully agree with your blue Obama baseline--and I appreciate the analysis that Obama basically needs to win just one of the toss-up states. That shows how ugly the map is for McCain. (He's talking now about making a play for PA, but I don't buy it.)
monkeyman: I think you might be misreading the 538 site. They currently give McCain a 6.6% chance of winning (it's the second pie chart down in the left-hand column.)
As to how they get their probabilities... For each state they essentially conduct a sophisticated regression analysis across all polls to estimate two key statistics: a mean and a standard deviation. With those two statistics, it is straightforward to calculate the probability that a given candidate will receive 50%+ of the vote. One of the things I adore about fivethirtyeight.com is that they show all the data out in the open. And they have a lengthy explanation of their methods at the end of their faq.
The RCP polling average is returning to where it was last week. We're seeing a certain amount of stability coming into this race.
As for what Palin will do, don't expect her to have the political smarts to do anything other than what she is told right now. She's so far out of her league it's not funny. Seriously, it's not funny.
DrS, thanks for the explanation. Also, you and Bell Curve are right of course. Guess I got a bit overenthusiastic when calculating the probability, but clearly the independence assumption doesn't hold. Nice catch!
Post a Comment