• Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.

PolliGaf 2012 |OT5| Big Bird, Binders, Bayonets, Bad News and Benghazi

Status
Not open for further replies.

Jackson50

Member
I cannot stomach this chicken littling about a storm impacting an election. I'm taking the weekend off from PoliGAF. See you guys on the other side of the Greatest Storm in Human History.
It's not chicken littling, Tim. I looked at the other elections when major storms coincided with the election. And the evidence shows Republicans always benefit. Check out those three links. And this storm is even WORSE. Maybe early voting save Obama. But I'm not resting my hat on it.

*edit* On the maps, blue=Republicans.
 

Cheebo

Banned
The Gallup and Ras polls really do bug me. Hopefully they're the ones that are wrong and not the others.

It is worrisome because they got the 2008 election pretty dead on, same with 2004. I know they were pretty far off in 2010 but national congressional popular vote is a lot different (and they were right who won, just the wrong spread). It still leaves me on edge. Can both, the two longest running tracking polls be this far off?
 

giga

Member
So now that Obama can't win New York, despite having a 25+ margin in the polls there, I wonder what he's going to do after the election. Maybe return to the Senate?
 

gkryhewy

Member
It is worrisome because they got the 2008 election pretty dead on, same with 2004. I know they were pretty far off in 2010 but national congressional popular vote is a lot different (and they were right who won, just the wrong spread). It still leaves me on edge. Can both, the two longest running tracking polls be this far off?

I would be surprised if they both don't converge toward a tie by election day. But we'll see- those two polls are definitely keeping things interesting.
 
So now that Obama can't win New York, despite having a 25+ margin in the polls there, I wonder what he's going to do after the election. Maybe return to the Senate?

Whether he leaves office in 2013 or 2017, Obama is going to spend the rest of his life rolling around in the gigantic mountains of cash he'll make speaking on the private circuit.
 

Amir0x

Banned
I cannot stomach this chicken littling about a storm impacting an election. I'm taking the weekend off from PoliGAF. See you guys on the other side of the Greatest Storm in Human History.

I find it hilarious. This is the same freaking out that happened after the first debate, and in the two weeks following it with the poll watchers

certain people have no idea how to actually view data :p
 

Jackson50

Member
So now that Obama can't win New York, despite having a 25+ margin in the polls there, I wonder what he's going to do after the election. Maybe return to the Senate?
He might become Chief Justice like Taft. But with Romney poised to win because of Sandy, the Republicans will take credit for the recovery. We won't see a Democratic president until AT LEAST 2024. But more likely 2028, and I don't think a Republican appoints Obama Chief Justice.
The blues and reds swapped there hurt my brain at first...
Good catch.
 

Slime

Banned
I could see the cell phone thing throwing off pollsters now more than ever before. The iPhone was a pretty big deal in '08, but smartphones in general have exploded so much more in the 4 years since, so I could see that messing with the methodology a bit. Just anecdotally, my family members had cell phones back then, but mostly stuck with the landline when at home. Since then they've completely replaced it.

I don't know if that's the answer, but I know it's been speculated that it's having some effect. I just think it's possible that, 4 years later, it might be having a bigger effect than expected.
 

AlteredBeast

Fork 'em, Sparky!
He might become Chief Justice like Taft. But with Romney poised to win because of Sandy, the Republicans will take credit for the recovery. We won't see a Democratic president until AT LEAST 2024. But more likely 2028, and I don't think a Republican appoints Obama Chief Justice.

Why not? Obama has been more right wing on so many positions than typical Republican presidents. :p
 
I could see the cell phone thing throwing off pollsters now more than ever before. The iPhone was a pretty big deal in '08, but smartphones in general have exploded so much more in the 4 years since, so I could see that messing with the methodology a bit. Just anecdotally, my family members had cell phones back then, but mostly stuck with the landline when at home. Since then they've completely replaced it.

I don't know if that's the answer, but I know it's been speculated that it's having some effect. I just think it's possible that, 4 years later, it might be having a bigger effect than expected.

Pretty sure rasmussen polls cell phones now.

And they're STILL wildly off compared to everyone else.
 

Slime

Banned
Pretty sure rasmussen polls cell phones now.

And they're STILL wildly off compared to everyone else.

I just Googled to make sure, and Rasmussen does just poll landlines. However, it seems that Gallup polls everyone, so that doesn't help my theory much regardless. Weird.
 
Most of the early voting that will matter to the election is occurring far enough inland that I don't think the hurricane will have a significant effect on turnout. It would be unfortunate if it depressed early voting in North Carolina, but Obama only has an outside shot to win that state anyway. One important factor, as the article indicates, is probably how the administration responds to the hurricane. In order to actually lose NC and FL due to the hurricane, however, those states would have to be remarkably close anyway, which would actually be an improvement on Obama's position now.


Agreed.

I don't think it will effect the Presidential election all that much. It seems like the effect on North Carolina is going to be nil, and the effect on Florida is definitely going to be nil. I don't see northwest VA getting hit that hard, and the Northeast is going to go for Obama (impacts there will cancel each other out.)

ManmadeMan said:
"rain" isn't going to knock out power for a week. It's high winds and downed trees that do that.
Floyd dumped a ton of rain in the area (about 15 inches in a day) and was completely a non issue.

The high winds will be around for a lot longer than other storms because of the slow speed of the storm. This storm is projected to have a worse impact than the other storms.
 

AlteredBeast

Fork 'em, Sparky!
I just Googled to make sure, and Rasmussen does just poll landlines. However, it seems that Gallup polls everyone, so that doesn't help my theory much regardless. Weird.

At the end of the day, neither can really change what they have already done. The damage is done and they pretty much have to follow their methodology through the election and then meter how far off they actually were, then adjust for demographics changes and so on. If they changed their methodology tomorrow, it would look like they were bending to pressure to match PPP, Pew, Marist, etc.
 
If lines are already huge at some places. How big are they come election day and how many people simply choose not to vote because they don't want to stand in line for hours?
 

pigeon

Banned
Given the tenor of this thread currently I thought I'd note that if Obama's lead in Ohio is really 2 points we should probably expect to see Romney ahead in at least one poll before the election.
 

Mgoblue201

Won't stop picking the right nation
Gallup actually overestimated Obama's winning margin by 4 points in 2008, and they were off, on a national level, in 2010 by a larger margin. That's the danger of using any one poll, or even a couple of polls, to predict a race, because they often disagree. Fortunately, the fundamentals of the race still favor Obama to win the electoral vote, and I still think he wins the popular vote by a small margin. The aggregate of the national polls basically show a tied race, and I believe they're underestimating Obama's support by a small amount.
Agreed.

I don't think it will effect the Presidential election all that much. It seems like the effect on North Carolina is going to be nil, and the effect on Florida is definitely going to be nil. I don't see northwest VA getting hit that hard, and the Northeast is going to go for Obama (impacts there will cancel each other out.)
Does VA even have early voting, outside of absentee voting? If not, then the storm probably shouldn't matter there, and Romney has to win VA before he even thinks about winning OH.
 

ido

Member
Romney is just running up totals hardcore in the South.

Not that my small anecdotal slice of the south is any sort of proof at all, but almost every single person that I work around firmly hates Obama, and cites the various lies that we have all heard for their reason(communist, muslim, gay, baby killer, black[not that being black is a lie, but they are quite prejudice against him for this], etc). Further more, they all TRULY believe that Obama has no chance at all to win. During our breaks, the only station they watch on our TV in the break room is FOX, and of course FOX only really shows favorable polls to Romney. All of them quite literally think this will be more of a landslide victory than the election of 1980.

Thankfully, at least so far, the numbers favor Obama. With a little luck, I predict my day at work on November 7th will be amazing. My only other coworker who is voting for Obama is going to have a cake made to look like Obama and bring it to the rest of the crew.
 

Amir0x

Banned
If that's what's happening, and it certainly seems plausible, Obama's going to lose the popular vote.

he very well might

but in the end, what would change? Republicans are going to blockade with or without the popular vote.

Obama's second term is for Supreme Court Justices, ending any challenges to the ACA and, if we're lucky, some form of immigration reform. And of course common sense social policies, vis-a-vis gay equality and women's reproductive rights. Not much of this is going to change no matter what the outcome of the popular vote is
 

ido

Member
Please be right. Please, please be right.

From Sam Wang at Princeton:

Seeing as how predictions are so useful, Andrew Ferguson and I have decided to put the Obama re-elect probability in the topline. We give two probabilities, which are built on the same assumptions that went into calculating the “strike zones” in the history graph. The “Random Drift” number is a minimum (conservative) probability, and the “Prediction” is a maximum probability. In the coming 10 days, the two numbers will converge.

Probability of Obama re-election: Random Drift Model 89%, Bayesian Prediction 97%

These motherfuckers are pretty damn certain, huh.
 
Agreed.

I don't think it will effect the Presidential election all that much. It seems like the effect on North Carolina is going to be nil, and the effect on Florida is definitely going to be nil. I don't see northwest VA getting hit that hard, and the Northeast is going to go for Obama (impacts there will cancel each other out.)



The high winds will be around for a lot longer than other storms because of the slow speed of the storm. This storm is projected to have a worse impact than the other storms.

So, let me get this straight.

This storm is about half the wind speed of Gloria, which had no impact on the area a week out.

This storm is dumping half the rain of Floyd, which had no impact on the area a week out.

So "slow speed" hurricanes with no rain are now the new apocalyptic scenario? who knew?

Gallup actually overestimated Obama's winning margin by 4 points in 2008, and they were off, on a national level, in 2010 by a larger margin. That's the danger of using any one poll, or even a couple of polls, to predict a race, because they often disagree. Fortunately, the fundamentals of the race still favor Obama to win the electoral vote, and I still think he wins the popular vote by a small margin. The aggregate of the national polls basically show a tied race, and I believe they're underestimating Obama's support by a small amount.

I thought I recalled seeing something about how Gallup's likely voter polls were heavily influenced by enthusiasm, which would explain being wildly off in 2008, and the ridiculous Romney lead post 1st debate.
 

Clevinger

Member
So, let me get this straight.

This storm is about half the wind speed of Gloria, which had no impact on the area a week out.

This storm is dumping half the rain of Floyd, which had no impact on the area a week out.

So "slow speed" hurricanes with no rain are now the new apocalyptic scenario? who knew?

Being in Arizona, I know nothing of these storms, but I keep hearing that its pressure is dropping and that's part of why it will be so bad.
 
Just noticed Obama is almost at 75% on pinko Nate Silver's skewed stats blog

Someone should tell him about that 80k person Romney rally in Minnesota today
 
Romney said "please", though. That Cocky Obama and his sidekick were guffawing, pleading for more time when they already had talked more and made everybody (except the moderator's, who they obviously had in their pocket) feel dumb!

While I don't talk politics with my patients I did have one "undecided" who said she was probably going the Romney-Ryan route because during his closing comments, Ryan actually asked for her vote while Biden apparently did not. Face palms all day.
 
Just noticed Obama is almost at 75% on pinko Nate Silver's skewed stats blog

Someone should tell him about that 80k person Romney rally in Minnesota today

don't worry. my algorithm takes into account the respective rally #'s of each candidate. and at this juncture romney has a measured advantage over obama.
 

Drek

Member
If lines are already huge at some places. How big are they come election day and how many people simply choose not to vote because they don't want to stand in line for hours?

Voting is typically in the mid 50's to barely 60% of the voting age populace because of this. If voting was quick, easy, and took little effort more people would vote.

Even '08, that big "swing" year only beat out '04 by 1.7%. There was a bigger jump from 2000 to 2004 (54.2 to 60.1).

So we're actually only working with a shade over half the electorate anyhow, unless this year is some massive freak event.

Chances are it's not and we'll see about 60% voter turnout again.

This is also a questionable part of Gallup's model. Different states have different turnout trends, so their heavy sampling of the southern states ignores that those are also the worst states for turnout. Meanwhile the northeast and midwest have much better turnout as a whole.

Red states historically have poor turnout v. population compared to the national average across regions actually. Indiana consistently under performs against it's midwest peers for example.

Makes you wonder what the political map would be like if ever state made voting as painless as Oregon where they mail ballots to everyone. Oregon has a younger than average population which would normally suggest poor voter turnout, but they have consistently outperformed the national average quite healthily (little over 67% on '08) since going to mail ballots in '98.

If that ever became standard I think we'd see a wave of blue wash over the south as the obviously very disenfranchised non-voters in those regions would have a far easier time making their voices heard. Those disenfranchised non-voters are, historically, minorities and women in the south.
 

mckmas8808

Mckmaster uses MasterCard to buy Slave drives
Please be right. Please, please be right.

From Sam Wang at Princeton:



Probability of Obama re-election: Random Drift Model 89%, Bayesian Prediction 97%

These motherfuckers are pretty damn certain, huh.

Can you imagine what it would be like if both Nate and Sam were both wrong?
 

Mgoblue201

Won't stop picking the right nation
Voting is typically in the mid 50's to barely 60% of the voting age populace because of this. If voting was quick, easy, and took little effort more people would vote.

Even '08, that big "swing" year only beat out '04 by 1.7%. There was a bigger jump from 2000 to 2004 (54.2 to 60.1).

So we're actually only working with a shade over half the electorate anyhow, unless this year is some massive freak event.

Chances are it's not and we'll see about 60% voter turnout again.

This is also a questionable part of Gallup's model. Different states have different turnout trends, so their heavy sampling of the southern states ignores that those are also the worst states for turnout. Meanwhile the northeast and midwest have much better turnout as a whole.

Red states historically have poor turnout v. population compared to the national average across regions actually. Indiana consistently under performs against it's midwest peers for example.

Makes you wonder what the political map would be like if ever state made voting as painless as Oregon where they mail ballots to everyone. Oregon has a younger than average population which would normally suggest poor voter turnout, but they have consistently outperformed the national average quite healthily (little over 67% on '08) since going to mail ballots in '98.

If that ever became standard I think we'd see a wave of blue wash over the south as the obviously very disenfranchised non-voters in those regions would have a far easier time making their voices heard. Those disenfranchised non-voters are, historically, minorities and women in the south.
The one caveat is that more people than normal seem to think that this election matters and are aware of its historical nature. When combined with the greater emphasis on early voting, turnout could be up this year, if only slightly.
 

TCKaos

Member
This whole North East situation reminds me of how the GOP said Obama hired a wizard to ruin the RNC, or some such.

We need to figure out of Romney has hired a witch doctor or shaman to give himself a winning edge.
 

Cheebo

Banned
I imagine that I would be fucking pissed.

WHY WOULD YOU SAY THIS.

I think we will be too busy freaking out about Obama losing and Romney being president to start lashing out at Nate Silver. At least in the near term. But man his blog will lose all credibility if Obama loses.
 

Drek

Member
probably would be head scratching, but not as earth shattering as if both ras and gallup were wrong.

Really?

Because anyone with a modecum of statistical knowledge can see weaknesses in how their numbers add up.

Ras has a huge error in being a cheapo land line robocaller.

Gallup is using a false equivalency on geographical voter turnout.

538 and Princeton both being wrong means that all polling has major errors and those errors are extreme enough to fail proper off-sets.

For a real world example, Ras and Gallup being wrong are like you getting an appraisal on your house from two different sources, one says it's worth 250K, the other 270K. You go to sell it and the best offer you get is 230K. Clearly they both misjudged the market.

538 and Princeton being wrong is like what Wall St. did with their up and down bets on the Eurozone, where somehow most of the major banks had themselves tied into investments that were supposed to statistically off-set one another but somehow a fiscal doomsday scenario occurred where they could lose all their bets across the board, despite half being intentionally aligned to offset the other half. In short, the smarest guys in the room using the largest collection of data sets somehow wound up with a metadata analysis where errors weren't balanced out by competing data, but instead amplified by it, without anyone noticing.

Now obviously the later is possible, as it's happened recently in the financial sector, but it happening in the financial sector is a doomsday scenario everyone involved has been telling tales of woe and carnage about for months. The former scenario happens daily all across the country.
 

Drek

Member
The one caveat is that more people than normal seem to think that this election matters and are aware of its historical nature. When combined with the greater emphasis on early voting, turnout could be up this year, if only slightly.

And it wasn't in '08? Everyone claims '08 as a massive swing election and we saw a 1.7% bounce over '04.

If we hold at '08 numbers most likely voter models will have underestimated voter turnout and Obama is going to win states no one expected him to win (like NC).
 

HylianTom

Banned
If that's what's happening, and it certainly seems plausible, Obama's going to lose the popular vote.
Yep.

And Romney "running up the score" in the South has been demonstated in several pollsters' crosstabs, so the evidence is there; is isn't some weird, funny, let's-laugh-at-the-rednecks-again theory. Obama winning his blue states and swing states by narrower margins has also been demonstrated, so the math is indeed pointing in this direction.

In a normal election year where regional vote differences aren't very dramatic, election viewers could tell the status of the true race with a quick glimpse at the national numbers, The national numbers were a pretty good reflection of the candidates' chances in the Electoral College, and if a candidate was ahead by 1 or 2 percent, he'd almost certainly win electorally.

But with the South skewing the national figures, the national popular vote percentages are no longer a valid indicator of the shape of the race. Obama's campaign managers know this, which is why they aren't running any national polls, as these polls are glamour contests and nothing else.

In a really pervese way, the Election of 2000 may have helped Obama with this big issue. The last time before 2000 where the winner won the electorally but not popularly was in 1888, so the American voting public wasn't going to have such a possibility placed as prominently in their minds. But with Gore's loss, American voters are now more resistant to looking at only those national beauty contest numbers to get a bead on the state of the race.

Voters now know very well it's all about the key states, and media outlets are reinforcing this more than ever, That's why we'll see NBC teasing all day, "hey! We've got new Ohio and Virginia numbers on the NBC Nightly News tonight! Tune in!" If Gore hadn't lost, we'd all still be much more nervous and demoralized from lackluster national numbers. Instead, seeing that Obama is leading in Ohio and other key states has the Democratic base fired-up to get the job done.

I've seen in multiple places claims that "surely the national and state numbers will adjust by Election Day to fit each other, and the nature of those adjustments will determine the winner." But I say that no adjustment is needed. The numbers aren't closing, and if they continue not to, we may very well see it happen cleanly (i.e., without election/ballot controversy) for the first time in 124 years.

Do I fear some of the repercussions? Sure, especially for the President's life, and for the lives of others.

But as far as legitimacy is concerned, Obama will never be "legit" in the GOP's eyes, even though he will have beaten them fairly two times. So fuck 'em. The American people accepted Bush and Hayes as fair winners, and they'll accept Obama for four more years.
 

Kettch

Member
I think we will be too busy freaking out about Obama losing and Romney being president to start lashing out at Nate Silver. At least in the near term. But man his blog will lose all credibility if Obama loses.

Not sure if you're joking or not, but 75% to win means Obama loses 1 out of 4 times. If Nate's 75% predictions were always right, that would cause him to lose credibility.
 
Really?

Because anyone with a modecum of statistical knowledge can see weaknesses in how their numbers add up.

Ras has a huge error in being a cheapo land line robocaller.

Gallup is using a false equivalency on geographical voter turnout.

538 and Princeton both being wrong means that all polling has major errors and those errors are extreme enough to fail proper off-sets.

For a real world example, Ras and Gallup being wrong are like you getting an appraisal on your house from two different sources, one says it's worth 250K, the other 270K. You go to sell it and the best offer you get is 230K. Clearly they both misjudged the market.

538 and Princeton being wrong is like what Wall St. did with their up and down bets on the Eurozone, where somehow most of the major banks had themselves tied into investments that were supposed to statistically off-set one another but somehow a fiscal doomsday scenario occurred where they could lose all their bets across the board, despite half being intentionally aligned to offset the other half. In short, the smarest guys in the room using the largest collection of data sets somehow wound up with a metadata analysis where errors weren't balanced out by competing data, but instead amplified by it, without anyone noticing.

Now obviously the later is possible, as it's happened recently in the financial sector, but it happening in the financial sector is a doomsday scenario everyone involved has been telling tales of woe and carnage about for months. The former scenario happens daily all across the country.

If 538 is wrong, it also means lot of the state polls conducted were wrong. PPP, Quinnipac, Marist, etc, etc.

In the end both models are driven by polls, if the state polls were like Ras National or even some of their state ones, Romney would be favorite in both models. If the underlying data is wrong, the models will be wrong too.
 
Really?

Because anyone with a modecum of statistical knowledge can see weaknesses in how their numbers add up.

Ras has a huge error in being a cheapo land line robocaller.

Gallup is using a false equivalency on geographical voter turnout.

538 and Princeton both being wrong means that all polling has major errors and those errors are extreme enough to fail proper off-sets.

For a real world example, Ras and Gallup being wrong are like you getting an appraisal on your house from two different sources, one says it's worth 250K, the other 270K. You go to sell it and the best offer you get is 230K. Clearly they both misjudged the market.

538 and Princeton being wrong is like what Wall St. did with their up and down bets on the Eurozone, where somehow most of the major banks had themselves tied into investments that were supposed to statistically off-set one another but somehow a fiscal doomsday scenario occurred where they could lose all their bets across the board, despite half being intentionally aligned to offset the other half. In short, the smarest guys in the room using the largest collection of data sets somehow wound up with a metadata analysis where errors weren't balanced out by competing data, but instead amplified by it, without anyone noticing.

Now obviously the later is possible, as it's happened recently in the financial sector, but it happening in the financial sector is a doomsday scenario everyone involved has been telling tales of woe and carnage about for months. The former scenario happens daily all across the country.

i've taken enough stats classes to know that stats can lie and be manipulated to produce any result. now i'm not saying wang and his friend at the ivory tower down the street, silver, are cooking the books but they do have very skeptical methodologies.
 

Clevinger

Member
Yep.

And Romney "running up the score" in the South has been demonstated in several pollsters' crosstabs, so the evidence is there; is isn't some weird, funny, let's-laugh-at-the-rednecks-again theory. Obama winning his blue states and swing states by narrower margins has also been demonstrated, so the math is indeed pointing in this direction.

It's interesting though that Nate doesn't seem to buy it. Last night on Maher he said he thinks there's only a 7% chance of Obama losing the popular but winning the electoral.

Sam Wang said something like a 25% chance a handful of days ago.
 

Owzers

Member
biden/obama seem to be slacking today, only one stop each. Ryan is on a bus tour of ohio with 5 stops, Romney has 3 stops in florida. Must have a lot of spare time by not giving a single interview to anyone :p
 
Status
Not open for further replies.
Top Bottom