Poll Vault

[Undesired Walrus]

I suppose all Obama has to do is retain all Kerry States and win Virginia and Colorado, being able to afford losing Ohio and Florida. That comes in at a slim electoral college victory, but a win nonetheless. Now that Virginia and Colorado appear to be safe, all he needs to work for is a landslide.

 

[Pookster]

Latest on Indiana ...

Indiana (11)
October 21 - PPP (D): Obama by 2 (48%-46%)
October 9 - Rasmussen: McCain by 7 (50%-43%)
October 7 - CNN/Time: McCain by 5 (51%-46%)
October 7 - Research 2000: Tied (46%-46%)
October 1 - SurveyUSA: McCain by 3 (48%-45%)


Another national tracking poll out this morning ...

October 21 - Hotline/FD Tracking: Obama by 6 (47%-41%) -- Obama lead +1 from Oct 20

 

[Malerin]

Early turnout is heavily favoring Dems in Florida and Nevada:

"FLORIDA: Long lines greeted early voters in the Sunshine state. "Registration numbers released Sunday show a 600,000-voter edge for Democrats over Republicans in Florida: 4.7 million versus 4.1 million, with 2.1 million people identifying with neither party. The McCain campaign acknowledged it expected more Democrats than Republicans to vote early, but says GOP voters have requested 295,000 absentee ballots statewide compared with 199,000 Democrats."

"NEVADA: Jon Ralston on the early voting in the state: “Third day exceeds second. Turnout will hit double-digits Tuesday. Democrats still have huge lead -- 40,625 -17,509. That’s 59 percent to 25 percent. (Monday it was 13,099-6,419.) In CD3, it’s 2-to-1 Democratic voter lead. Senate 5 (Heck) just under 2-to-1 Democrats; Senate 6 (Beers) about 2-to-1, too. The Democrats essentially are doubling the GOP turnout -- or more -- every day. Can they possibly keep this up?"

http://firstread.msnbc.msn.com/

 

[BenBurch]

I suppose all Obama has to do is retain all Kerry States and win Virginia and Colorado, being able to afford losing Ohio and Florida. That comes in at a slim electoral college victory, but a win nonetheless. Now that Virginia and Colorado appear to be safe, all he needs to work for is a landslide.

"Winning Ugly" is still winning, and all that really matters.

McCain will have to fight for every state. Nothing is conceded. This is a "50 state strategy" and is what is bankrupting McCain and the GOP at a time when they are at a huge disadvantage in funds.

 

[MattusMaximus]

I WILL let you know. As little as I like repeatedly defending McCain in this thread, I can't let this flagrant abuse of statistics go unaddressed.

The calculations you're using assume that the events are INDEPENDENT, which they certainly are not. If John McCain wins Virginia, he's almost certainly won North Carolina as well, and with the situation as it is now, probably Ohio and Florida as well.

I imagine his real odds would be more like a substantial fraction of his weakest state that he needs to win. Which is what the markets appear to predict.

Damn, I knew I screwed something up. That's what I get for drinking and posting late at night with too little sleep

ETA: Thanks for the catch!

 

[MattusMaximus]

Then when McCain's poll lead vanished, the prediction markets followed right behind it. The numbers are utterly irrelevant in terms of prediction who's going to win the race. It's just a fun thing to follow, like fantasy football.

Actually, prediction markets are better predictors than many polls. Here's a good article from Scientific American which explains why this is the case...

Super Tuesday: Markets Predict Outcome Better Than Polls

Internet-based financial markets appear to forecast elections better than polls do. They also probe how well the next George Clooney drama will do at the box office and how bad the next flu season will be.

*In 1988 the University of Iowa launched an experiment to test whether a market using securities for presidential candidates could predict the outcome of the election.

*In presidential elections from 1988 to 2004, the Iowa Electronic Markets have predicted final results better than the polls three times out of four.

*Despite the track record of the Iowa market, a fund*amental understanding of how prediction markets work remains elusive, and economists are still trying to develop a body of theory to provide definitive answers. ...

 

[Undesired Walrus]

Do you have a graph that compares previous prediction markets to previous poll figures? I'd like to believe their accuracy, but with the McCain bump in September, it seemed to show itself as simply comforming to the poll figures. It showed no maverick spirit, as it was.

 

[Pookster]

Another North Carolina poll ...

North Carolina (15)
October 21 - Civitas/TelOpinion (R): Obama by 3 (48%-45%)
October 20 - FOX News/Rasmussen: Obama by 3 (51%-48%)
October 20 - PPP (D): Obama by 7 (51%-44%)
October 18 - Research 2000: Obama by 2 (46%-44%)
October 15 - InAdv/PollPosition: Obama by 2 (48%-46%)
October 14 - PPP (D): Obama by 3 (46%-43%)
October 13 - FOX News/Rasmussen: Tied (48%-48%)
October 11 - WSOC-TV: McCain by 2 (48%-46%)
October 9 - Civitas/TelOpinion (R): Obama by 5 (48%-43%)
October 9 - Rasmussen: Obama by 1 (49%-48%)
October 7 - SurveyUSA: McCain by 3 (49%-46%)
October 7 - CNN/Time: Tied (49%-49%)
October 6 - PPP (D): Obama by 6 (50%-44%)

 

[Pookster]

Actually, prediction markets are better predictors than many polls. Here's a good article from Scientific American which explains why this is the case...

Super Tuesday: Markets Predict Outcome Better Than Polls

Polls aren't predictors. They're a snapshot of opinion at a point in time. Prediction markets utilize polls much like those in the stock market use the financial reports of businesses.

 

[Pookster]

Gallup's latest national tracking polls ...

October 21 - Gallup Tracking (Likely Voters Trad.): Obama by 7 (51%-44%) -- Obama lead +2 from Oct 20
October 21 - Gallup Tracking (Likely Voters Exp.): Obama by 10 (52%-42%) -- Obama lead +1 from Oct 20

http://www.gallup.com/poll/111298/Gallup-Daily-Obama-Holds-Lead-Various-Scenarios.aspx

Among all registered voters, there has been fairly little variation in recent days, with Obama receiving between 50% and 52% of the vote over the last five reports and McCain in a range between 41% and 43%. In the current three-day rolling average of registered voters Obama remains ahead by 52% to 41%, exactly the same as Monday's update.

Gallup's modeling of likely voters indicates the race is slightly tighter if we assume that voter turnout patterns will be similar to those seen in most presidential elections from 1952 through 2004. Using this "traditional" definition of likely voters, which takes into account respondents' history of voting, as well as their current interest in the campaign and self-reported likelihood to vote, Obama leads McCain by seven points, 51% to 44%. This is tied for Obama's largest lead among this group since Gallup began reporting likely voters in Gallup Poll Daily tracking.

 

[aerosolben]

That's true, but there are plausible scenarios where Mccain could win Virginia, due to heavy Republican turnout, but still lose in N. Carolina, due to heavy black turnout. These state elections may not be independent events like coin tosses, but they're not joined at the hip either.

That's why said almost certainly.

Damn, I knew I screwed something up. That's what I get for drinking and posting late at night with too little sleep

ETA: Thanks for the catch!

My pleasure.

 

[Pookster]

Another poll for Florida ...

Florida (27)
October 21 - PPP (D): Obama by 1 (48%-47%)
October 20 - FOX News/Rasmussen: McCain by 1 (49%-48%)
October 17 - SurveyUSA: McCain by 2 (49%-47%)
October 17 - Research 2000: Obama by 4 (49%-45%)
October 15 - Datamar: Obama by 5 (47%-42%)
October 15 - CNN/Time: Obama by 5 (51%-46%)
October 15 - InAdv/PollPosition: Obama by 4 (48%-44%)
October 13 - FOX News/Rasmussen: Obama by 5 (51%-46%)
October 10 - Research 2000: Obama by 5 (49%-44%)
October 10 - Strategic Vision (R): Obama by 8 (52%-44%)
October 9 - Rasmussen: Obama by 3 (50%-47%)
October 7 - Mason-Dixon: Obama by 2 (48%-46%)
October 6 - FOX News/Rasmussen: Obama by 7 (52%-45%)

 

[Malerin]

Pew has Obama up by 14. RCP average at 6.9. Weird thing is, there's now a spread of +1 to +14.

 

[MattusMaximus]

The latest from the prediction markets...

IEM-WTA
Obama 87.2
McCain 13.6

IEM-Vote Share
Obama 55.5
McCain 46.0

NewsFutures
Obama 80
McCain 20

Intrade
Obama 84.2
McCain 15.6

Yahoo Political Dashboard (YPD) Electoral College
Obama 364
McCain 174

YPD Battleground Markets:

Ohio - Obama 62.7 McCain 39.9
Virginia - Obama 79.0 McCain 22.0
North Carolina - Obama 58.5 McCain 43.9
Florida - Obama 65.0 McCain 37.6
Missouri - Obama 60.0 McCain 40.2
Colorado - Obama 82.5 McCain 15.1
Nevada - Obama 73.5 McCain 26.5

 

[MattusMaximus]

So much for the PUMAs

Well, so much for the hordes of PUMAs out there who were threatening Democratic party unity. It seems the Dems are more united now than the GOP - oops

Sorry, Jim, the PUMAs are Dead

Over at the Campaign Spot, Jim Geraghty wonders how the new ABC/Washington Post poll could have the temerity to suggest that Barack Obama is winning more support among Democrats than John McCain is among Republicans:

"The ABC/WashPost poll in late September found 86 percent of Republicans for McCain, 88 percent of Democrats for Obama. PUMAs don't exist anymore? Colin Powell and Christopher Buckley are leading the exodus of Republicans for Obama? I suppose it's possible, but I have my doubts."

I suppose I have my doubts too, except that this finding is now reflected in any number of polls. In fact, six of the eight trackers that published today included the support that each candidate is winning within his respective party. ...

... Among Democrats, Barack Obama is now winning 88 percent support, comparable to John Kerry in 2004 or Al Gore in 2000. And there are a couple of points' worth of undecideds left in there, so it's possible that Obama could scrape up against the 90 percent number on election day.

By contrast, John McCain is winning the support of just 85.3 percent of Republicans, well down from Bush's 93 percent in 2004 and 91 percent in 2000. There are some undecideds in there as well, so his numbers should improve some, but McCain is likely to underperform Bush by several points. ...

... If Obama wins by the same margin among Democrats that McCain does among Republicans (and to reiterate, right now Obama is doing a bit better within his own party), and Rasmussen's most recent party ID breakdown is correct (D 39.7, R 33.0, I 27.3), then McCain will need to win independents by about 20 points to earn a draw in the popular vote.

Ouch - major ouch. There is no way McCain will take Indies by +20 points.

Incidentally, there are comments beneath the article. Below is one of the most funny that I read

There never really was such a thing as PUMAs, save for a couple of people who would get on the teevee and claim to be a PUMA.

If the McCain campaign's picking of Palin really was designed to garner the PUMA vote, then that decision goes down in history as one of the most damaging instances of believing one's own propaganda.

 

[Malerin]

Obama +10 in NBC/WSJ poll. Palin is a huge albatross now.

 

[BenBurch]

Obama +10 in NBC/WSJ poll. Palin is a huge albatross now.

I PRAY they nominate her in four years. Or I would if I believed in that sort of thing...

 

[Puppycow]

They're down to hoping to win Pennsylvania. You read it correctly. Pennsylvania. I hope everyone has they're forks ready.

Doesn't make sense to me. Didn't Bush win without Pennsylvania?
Which Bush state do they think they have a worse chance in than Pennsylvania?
Virginia?

 

[MattusMaximus]

Obama +10 in NBC/WSJ poll. Palin is a huge albatross now.

I'd say that Palin is more like a massive lead weight hanging around McCain's neck right about now. At least an albatross can flap its wings for a little lift...

 

[Dr Adequate]

Palin is a huge albatross now.

It is an ancient Republican,
And he appealeth to four in ten.