The Official 2008 Electoral College Prediction Thread

[Zkribbler]

Originally posted by Zkribbler
Obama 303 McCain 235.

I'm moving Florida, Indiana, New Hampsire and West Virginia into Obama's camp and thus changing my prediction to: Obama 376 & McCain 158

The only thing that can now save McCain's candidacy is a major terrorist attack on the U.S., say if al Qaeda nukes New York or something.

[Ben Kenobi]

God your an idiot.

Ok.

I'll take bets from all the polytubbies now, that Obama will win by 13 points come election day?

[EPW]

In Pennsylvania? I'll take that bet. If I win you have to become Asher's personal sex toy.

[Zkribbler]

God your an idiot.

Techically, that should be: "...you're an idiot."

[Ben Kenobi]

Fine by me. If I win, you have to wear an avatar of my choice.

[EPW]

Originally posted by Zkribbler


Techically, that should be: "...you're an idiot."

[EPW]

Originally posted by Ben Kenobi
Fine by me. If I win, you have to wear an avatar of my choice.

What, you don't like President Taft riding a Water Buffalo?

[siron]

That's to answer Kuci's question. As of 2007, the least efficient vote is in Texas, and the most efficient is in Wyoming.

This "efficiency" has nothing to do with the probability to change the outcome of the election. It is a well known result in social science that California is way overrepresented. But it is related to science, so repugs won't care.

[Ben Kenobi]

This "efficiency" has nothing to do with the probability to change the outcome of the election. It is a well known result in social science that California is way overrepresented. But it is related to science, so repugs won't care.

Then you shouldn't have any problems using empirical methods to demonstrate your conclusions.

[siron]

It is a decades old result to use the penrose method to achieve nearly same probabilities.

If this is not done you end up with Californian voters 3 times as powerful as Montaniacs:

http://www.cs.unc.edu/~livingst/Banzhaf/

[Boris Godunov]

Originally posted by Ben Kenobi
Ok.

http://www.electoral-vote.com/evp2008/Pres/Graphs/pennsylvania.html

PA back in September,

This doesn't back up your claim, Ben. It shows that for a very brief time in September, Obama and McCain were tied--not that McCain was ever ahead. This is attributable to McCain's convention bounce, it was never a longterm trend.

Furthermore, compare the composite of this year to 2004. Notice how Bush and Kerry pretty much were neck-and-neck the last two months, while the same trend is NOT in effect now? Obama has a comfortable lead, much more so than Kerry ever enjoyed. So there is again no logic to saying Kerry could win it with less-favorable polling numbers than while Obama can't.

Beyond that, RCP is indeed a Republican-leaning site that cherrypicks the polls it uses based on bias. Pollster.com is a much more reliable site.

They are now predicting North Dakota will go Obama.

I'll willingly take bets with any and all Polytubbies that ND will go McCain.

No, they are not. ONE poll came out showing Obama with a lead. That is not trumping all the polling history showing McCain with a lead. You clearly don't have a clue as to how polling aggregates work. None of the composite polling sites has moved ND into Obama's column, so I have no idea who "they" are, unless it's those voices in your head.

[quote]

http://www.electoral-vote.com/evp2008/Pres/Graphs/minnesota.html

Minnesota has been given to McCain in October.

No, the graph clearly shows the best he managed was a tie in one SUSA poll.

Wisconsin in September

No, the graph clearly shows that McCain never had a lead in September. You can't read a graph, can you?

Michigan in September

What did I say about cherry-picking one outlier poll, Ben? Look at the white circles, those are the averages. McCain never enjoyed a lead.

Iowa in September

No, the best he managed was one tie. Otherwise, it's an Obama blowout.

Ohio, still favoured to McCain

Huh?? The OH charts clearly show Obama currently favored:

http://www.electoral-vote.com/evp2008/Pres/Graphs/ohio.html

Currently O is up about 48-44.

New Mexico, McCain has lead at times during May, August, and September

The chart shows Obama averaging about a 48-44 lead.

Nevada, McCain lead throughout September. Missouri, McCain has lead throughout the campaign

No one said these weren't battlegrounds. Obama currently holds leads in both, however.

Virginia has been a dogfight throughout.

If it's neck-and-neck in Virginia, the Republican is in deep trouble.

We'll see this trend back in November as the pollsters will want to reflect closer to reality and not their wet dreams.

Right, because all pollsters are in the tank for Obama, including the Republican Rasmussen and Zogby, who is running his polls for Free Republic.

Electoral vote.com. It's great that Obama is leading by 10 now according to the polls, In September McCain was leading. You'd better hope these ridiculous polls keep pushing for Obama.

Why are they "ridiculous," Ben? You've yet to establish any rationale as to why ALL the polls are so wrong. McCain led very briefly because of a post-convention bounce. Since then, his numbers has nosedived for a very real and obvious reason: the economy and his erratic response to the crisis.

Obviously, since Gore won Florida in 2000, Bush won Wisconsin in 2004.

Huh?

Florida is decidedly a swing state--there's no consistency as to who has won it over the past 4 elections. Wisconsin is much less so--it's gone for the Democrats 4 of 4 times since 1992, and currently Obama is running ahead of Kerry and Gore.

No, he isn't, I'm sorry. He's not carrying the rust belt.

Because you stamp your feet and declare it so?

If Gore and Kerry could win them, with less-favorable polling numbers leading up to the election, why can't Obama, who is enjoying excellent polling numbers?

Did I say those states were unwinnable? I said clearly that Florida is unwinnable for Obama, despite the fact that it's right there in your predictions.

Uh... you said:

I mean you have Florida going for Obama, and I think any state in which Hillary whupped Obama, is unwinnable for Obama. That includes PA, and, surprise, surprise, NH.

Obviously, it pertains only to purple states, and not deep blue states like Massachusetts.

And again, just because you deem a state purple doesn't make it so.

Where the EC vote didn't match the proximity of the electoral vote?

Kennedy-Nixon. Kennedy squeaked through in big states which lead to a deceptively large EC margin.

Kennedy had over 303 EC votes, and yet Nixon lost by a popular vote margin of only 100k people. Kennedy's margin was 1,000 votes per EC college seat, which is extremely tiny.

No, Ben. The task was to support your claim that there had been elections where a candidate lost every single state that had been deemed toss-up or very close prior to the election in a close election. Of course, this isn't a close election, as Obama is winning handily at the moment.

Considering the vast differences in polling between 1960 and 2008, it's a pretty stupid comparison, anyway.

So the fact that Zogby had McCain up by 5 nationally is wrong while the Obama polls now are the gospel truth?

Ben, if all polls show a consistent lead for one candidate, except for ONE which shows the opposite, doesn't that logically tell you that the flaw was with the one outlier, not all the rest?

This is why, as I said before, you can't cherrypick one or two favorable polls, you have to look at composites. Only composites will offset the margins of error and statistical irregularities found in individual polls.

Beyond that, Zogby is not a reliable pollster. His interactive polls are easily the worst out there, because they don't rely on a random sample of phone voters, but rather internet voting.

Consider this: Zogby predicted Kerry winning in 2004. How did that work out for them?

If you're going to be stupid and put all your eggs in the basket of one pollster, at least pick one who historically has done a good job of accurately predicting the actual election results. Gallup did a far better job than Zogby, and for state-by-state polls, SUSA has the best track record. While it's still dumb to rely on a single poll for anything, at least then you'd be an inch or too closer to reality.

[Darius871]

Originally posted by Zkribbler
The only thing that can now save McCain's candidacy is a major terrorist attack on the U.S., say if al Qaeda nukes New York or something.

By now, wouldn't even that work in Obama's favor?

[Darius871]

oh and Ben

Why does anyone bother?

[Boris Godunov]

Originally posted by Ben Kenobi
Ok.

I'll take bets from all the polytubbies now, that Obama will win by 13 points come election day?

Too chicken to back up your puffery with just a straight Obama vs. McCain, who-will-win bet?

[Boris Godunov]

Originally posted by Ben Kenobi
I don't have faith in their national predictions. I do have faith in their poll archive and their trend lines, which isn't as subject to manipulation.

Except you're ignoring the trend lines, Ben. All of the charts show that the trend right now is that Obama is leading, in many cases comfortably. What you're doing is desperately clinging to outlier polls a month old or more to make a ridiculous case.

BTW, here's a real trend line graph for PA. McCain hasn't enjoyed a lead there since March, and even then it was very slim: