There's no reason for a subjective model to get so much attention when it was right once

I already gave you two reasons. If you want to ignore them, that's your call.

And, I don't see "3-1 favorite" as some over-the-top-Romney-is-done prediction. Silver's numbers are telling him that:

1. The popular vote will be very close.
2. Obama has the edge electorally, and will therefore probably win.

But saying Romney has a 25-30% shot at it is not saying that Romney is toast, or that the election is over. Joe C was doing baseball post-season series comps earlier. I think at the moment Nate's model is more or less telling us that we are late in Game 6 in a best-of-7 and Obama is up 3-2 in games and is leading 4-3 in the 8th inning of Game 6, but Romney has a couple of guys on base. Game 7 if necessary will be played Nov 6, so to speak.

You appear to think that we are in Game 7, and Romney is up 5-1 in the 8th inning, and anyone who thinks otherwise is either blind or biased.

One of the challenges with many polls tilting Obama's way is they offer up a party sample breakdown very close to or even exceeding the '08 electorate, which was +7D.

There is quite the lengthy discussion of this about ten pages back.

Funny thing on Nate's site. He has as his most likely electoral outcome as Obama ~ 330. And yet, if you give Obama all the states Nate has in blue, that gets him only 303. To get over 330, Obama has to win Florida as well (Nate has it Romney 63%). The next most likely is 340 something. 302 is only the third most likely outcome. I guess he figures that if Obama carries CO, IA, WI, OH, VA, and NH, he is likely to also carry FL and NC I guess (the only 2 Nate red states that aren't 95% Romney).

I'm not sure why any of us should have great confidence about any polls or aggregation thereof when they are within the margin of error. Published reports indicate that they get something like an 8% response rate, down from 35% a decade or two ago. I haven't seen any evidence showing that non-responders don't tilt one way or another. Not sure it has even been studied extensively. If the polls are off this year, they should probably take a look at that.

Campaigns say they always trust their internal data more than public polls because they are willing to pay for the best data but the public polls have to cut corners to keep costs down. Could be true, and if so, would Obama still be running almost entirely negative ads (at least in Virginia, and from what I hear elsewhere, too) if his data showed him to be ahead?

I'm not really heavily invested in changing anyone's mind, but I do think the Democratic partisans are overconfident, given the state of the race.

RCP doesn't use some national polls that either have the race tied or Obama slightly ahead (Ipsos/Reuters, YouGov, Zogby, Rand Corp, UPI). If those polls were in the average, Romney's +1.0 edge in the RCP average certainly shrinks or disappears entirely.

But if Romney wins Wisconsin, he's almost certainly winning Ohio as well.

It's quite possible that because of all the focus on early voting in Ohio that Obama could win Ohio and not win Wisconsin. I kind of doubt that will happen but I can see it happening. Wisconsin is a funny state. It's a state that you would think would go blue but there are a lot of people that vote red and do it consistently.

There is quite the lengthy discussion of this about ten pages back.

Did it include the mention of Gallup's 35D/36R/29I projected electorate slash line?

Published reports indicate that they get something like an 8% response rate, down from 35% a decade or two ago. I

Are we positive that 8% today doesn't equal or surpass 35% a decade or two ago? I would think it is vastly easier nowadays to dial thousands of phones at once thus you can reach more people and don't need the same % of people responding to build your pool of people.

I'm not really heavily invested in changing anyone's mind, but I do think the Democratic partisans are overconfident, given the state of the race.

Yeah, that makes Dems unique in this regard.

Could be true, and if so, would Obama still be running almost entirely negative ads (at least in Virginia, and from what I hear elsewhere, too) if his data showed him to be ahead?

Both campaigns are and have been running a massive percentage of negative ads, plus I don't see why running negative ads instead of positive ones means his campaigns thinks they're behind.

I enjoy this. Ohio isn't getting harder to poll--it's getting easier with early voting. There's less uncertainty in the polling because rather than having to rely on a likely voter screen, you can just ask people if they've already voted.

In theory, but apparently not in reality. As of a couple days ago, there was an almost 10-point difference between the percentage of people claiming to have voted in the polls of Ohio and the percentage of early voters as reported by the Ohio secretary of state. A lot of people are claiming to have voted who apparently haven't done so.

But if Romney wins Wisconsin, he's almost certainly winning Ohio as well.

Normally, but there might be a bit of a favorite son effect from Ryan being on the ticket, or even some residue from the Recall when Democrats went all-in but came up short. There's also that secret plan for Obama to deliver a generator and plow the driveway for every household in Ohio if Hurricane Sandy hits them - which I assume must have been what Andy was referring to a page or two back.

Out of the 8 national polls published on Friday, the average has Obama +0.2.

Graph of Friday National Polls from 538

RCP doesn't use 4 of those polls (PPP, Reuters, Rand, UPI). Those 4 polls are the first, third, fourth, and fifth most favorable polls for Obama (from the 8 total).

RCP doesn't use 4 of those polls (PPP, Reuters, Rand, UPI). Those 4 polls are the first, third, fourth, and fifth most favorable polls for Obama.

PPP is basically the in-house pollster for unions, while RAND didn't exist before July 2012, UPI had no tracker in 2008, and Reuters is new for 2012 after partnering with Zogby in 2008. They might end up being right, but I don't blame RCP for not using them, especially the three with no pre-2012 track records.

Nate is telling us that Obama remains a ~75 percent favorite because he holds a 2-point lead in a single swing state — and not only a single swing state, but a single swing state that's increasingly difficult to poll due to early voting

You realize that Nate thinks Obama is going to win Ohio (or has a very high percentage to win Ohio) because it's not the fact that he holds a 2-point lead, but because he holds a 2-point lead with less than 2 weeks to go, and he's held a lead in Ohio since the beginning, and there hasn't been any indication of Romney ever really leading in Ohio.

Using just the Ohio polls from RCP, there have been 38 polls of Ohio since September 1st.

Obama leads in 31 of them, tied in 3 of them, and Romney ahead in 4 of them.

The last poll that had Romney ahead by more than a single point in Ohio was Gravis Marketing on September 2nd.
The last polls that had Romney ahead at ALL was by a single point, more than 2 weeks ago.

RCP doesn't use some national polls that either have the race tied or Obama slightly ahead (Ipsos/Reuters, YouGov, Zogby, Rand Corp, UPI). If those polls were in the average, Romney's +1.0 edge in the RCP average certainly shrinks or disappears entirely.

Yep. The repeated claim that RCP is "objective" because it's a straight average ignores the fact that RCP subjectively chooses which polls to include in its average. And the widely held belief that Romney has a solid lead in the national polls is directly tied to RCP's choice of which polls to include in its national average.

But despite all their whining about Silver, the righties just can't escape the fact that 1) Obama leads in Ohio, and 2) Romney is very unlikely to win if he doesn't win Ohio. That's not some out-on-a-limb Silver creation; every aggregator agrees.

Are we positive that 8% today doesn't equal or surpass 35% a decade or two ago? I would think it is vastly easier nowadays to dial thousands of phones at once thus you can reach more people and don't need the same % of people responding to build your pool of people.

Oh, they still get whatever they want for the sample size, but if 92% are non-responders there may be an issue of how representative that sample is even if it's still the same size.

You realize that Nate thinks Obama is going to win Ohio (or has a very high percentage to win Ohio) because it's not the fact that he holds a 2-point lead, but because he holds a 2-point lead with less than 2 weeks to go, and he's held a lead in Ohio since the beginning, and there hasn't been any indication of Romney ever really leading in Ohio.

I thought Nate's model was supposed to regress poll results to the mean? Over the past eight or ten elections, the GOP candidate has never underperformed in Ohio relative to the national popular vote, and the average has been a +2.0 gain. Nate currently has Romney performing at -4.0 in Ohio relative to the national trackers. Might happen, but I'm doubtful.

Per one of the political blogs, in 2004, Kerry was leading in Ohio by almost a point at RCP but ended up losing Ohio by 2 points.

EDIT: I believe this was a date-to-date comparison — e.g., Oct. 25, 2004, to Oct. 25, 2012.

3-1 or 4-1 longshots come in all the time

Actually, they come in about 20-25% of the time. ;-p

Yep. The repeated claim that RCP is "objective" because it's a straight average ignores the fact that RCP subjectively chooses which polls to include in its average. And the widely held belief that Romney has a solid lead in the national polls is directly tied to RCP's choice of which polls to include in its national average.

Wait, so right-wingers are Luddites for trusting the national trackers and for occasionally mentioning an outlier poll in discussion, but now you're complaining that three trackers that didn't exist in 2008 are being excluded from RCP?

Nate's U.S. track record in 2010 was mediocre
...
There's no reason for a subjective model to get so much attention when it was right once and substantially wrong the second time around.

Er, what? Silver's formula called 34 of 36 Senate races correctly in 2010. For the two he missed, he'd incorrectly predicted that the Republican candidate would win. Silver called 36 of 37 Governor races correctly in 2010; again, for the one he missed, he'd incorrectly predicted that the Republican candidate would win.

He estimated a 54-seat GOP pickup in the House, while the real gain was 63. But Silver had explicitly cautioned that three-fourths of the House races had never been publicly polled by anyone, and therefore the range of possibility was more fluid and uncertain.

Did it include the mention of Gallup's 35D/36R/29I projected electorate slash line?

Can't remember - but I think the unskewed polls guy uses this to re-weight things. You could check into it on his website.

Oh, they still get whatever they want for the sample size, but if 92% are non-responders there may be an issue of how representative that sample is even if it's still the same size.

So then the flaw always existed and was probably worse back then. If you called 10,000 people and got 3,500 people to give you a response how is that better or more accurate than calling 100,000 people and getting 8,000 people to give you a response?

Re 5220.

Are you really going to let facts get in the way of a good narrative?

Nate over the last 4 years has earned his reputation and the benefit of the doubt.

I thought Nate's model was supposed to regress poll results to the mean? Over the past eight or ten elections, the GOP candidate has never underperformed in Ohio relative to the national popular vote, and the average has been a +2.0 gain. Nate currently has Romney performing at -4.0 relative to the national trackers. Might happen, but I'm doubtful.

I don't know whether (or how) Nate's model incorporates this, but the popularity of the auto bailout (and Ohio's lower-than-the-national-average unemployment rate) seems like a good reason Ohio might buck that trend this year, particularly given the GOP candidate's outspoken opposition to the bailout.

It's funny how Nate's rooting interest in Obama is a frequent topic of discussion, whereas the rooting interest of the founder of RCP seldom gets mentioned.

RealClearPolitics is a political news and polling data aggregator[2] based in Chicago, Illinois. The site's founders say their goal is to give readers "ideological diversity".[3] They have described themselves as frustrated with what they perceive as anti-conservative, anti-Christian media bias,[4] and while Patrick Stack of Time magazine has described the commentary as conservative-leaning,[5] the site includes columns and commentary from both sides of the political spectrum. Politico Executive Editor Jim VandeHei has called the site "an essential stop for anyone interested in politics."[6]

The site was founded in 2000 by former options trader John McIntyre and former advertising agency account executive Tom Bevan.[4][7][8] Forbes Media LLC bought a 51% equity interest in the site in 2007.[9] RCP has expanded to include a number of sister sites....

Philosophy

In an interview with the Chicago Tribune, McIntyre said,"We're trying to pull together the best political stories, op-eds, news analyses, editorials out there. The proliferation of content is enormous. Part of what we're trying to do is distill it in a clear, simple way for people who don't have hours to spend searching the Net."[6]

He told the Chicago Sun-Times that RealClearPolitics strives to feature "serious intellectual pieces" and that they're "not looking for the over-the-top, vitriolic, red-meat craziness on either side."[12]

In an interview with the conservative magazine Human Events, McIntyre described the philosophy behind the Web site as based on "freedom" and "common-sense values." Said Bevan, "We think debate on the issues is a very important thing. We post a variety of opinions." He further stated, "we have a frustration all conservatives have", which is "the bias in media against conservatives, religious conservatives, [and] Christian conservatives."[4]

Obviously this doesn't mean that RCP's slant is "biased". There's absolutely no reason to see RCP as anything more than an honest attempt to make sense out of the deluge of polling information that's out there.

BUT..... if we're going to be harping on Silver, then the silence concerning McIntrye's "Christian conservative" leanings is rather strange to say the least. Or perhaps "that's different".

So then the flaw always existed and was probably worse back then. If you called 10,000 people and got 3,500 people to give you a response how is that better or more accurate than calling 100,000 people and getting 8,000 people to give you a response?

Well, I don't think pollsters are using larger samples now. It's the same size sample but produced with a much lower response rate. If there is any tilt to the non-responders, the polls are likely to be off.

Er, what? Silver's formula called 34 of 36 Senate races correctly in 2010. For the two he missed, he'd incorrectly predicted that the Republican candidate would win. Silver called 36 of 37 Governor races correctly in 2010; again, for the one he missed, he'd incorrectly predicted that the Republican candidate would win.

He estimated a 54-seat GOP pickup in the House, while the real gain was 63. Silver had explicitly cautioned that three-fourths of the House races had never been publicly polled by anyone, and therefore the estimates were more fluid and uncertain.

This is great stuff. When Nate's model predicts essentially the same thing as the RCP average, it's proof of the genius of Nate's model. But when Nate is wrong, it's because he didn't get enough polling data.

***

So then the flaw always existed and was probably worse back then. If you called 10,000 people and got 3,500 people to give you a response how is that better or more accurate than calling 100,000 people and getting 8,000 people to give you a response?

Yikes.

Obviously this doesn't mean that RCP's slant is "biased". There's absolutely no reason to see RCP as anything more than an honest attempt to make sense out of the deluge of polling information that's out there.

BUT..... if we're going to be harping on Silver, then the silence concerning McIntrye's "Christian conservative" leanings is rather strange to say the least. Or perhaps "that's different".

Yes, it's different because an average is an average is an average, while Nate uses subjective factors in his modeling.

If you want to claim that RCP isn't using RAND or Reuters because of bias rather than because of their lack of any pre-2012 track record, you're free to do so, but I doubt it will get much traction.

Did it include the mention of Gallup's 35D/36R/29I projected electorate slash line?

So, everyone else is crazy for predicting a 2008 electorate, but Gallup is on the money with it's 2010 electorate.

Hi-larious.

Well, I don't think pollsters are using larger samples now. It's the same size sample but produced with a much lower response rate. If there is any tilt to the non-responders, the polls are likely to be off.

But why wouldn't they be calling more people nowadays?

This is great stuff. When Nate's model predicts essentially the same thing as the RCP average, it's proof of the genius of Nate's model. But when Nate is wrong, it's because he didn't get enough polling data.

You wrote that his overall track record in 2010 was "mediocre," and "substantially wrong." He went 496-for-508. What would an excellent performance have been, 540-for-508?

So, everyone else is crazy for predicting a 2008 electorate, but Gallup is on the money with it's 2010 electorate.

If Gallup's right, Romney wins in a landslide. But I don't believe many Romney supporters are really expecting that; they just know that if the electorate is even halfway between 2008 and 2004 (or 2008 and 2010), Romney wins. It's the "poll truther" thing turning into reality.

You wrote that his overall track record (496-for-508) was "mediocre," and "substantially wrong." What would an excellent performance have been, 540-for-508?

Nate's analysis of the GOP's prospects in 2009 and early '10 was, to be polite, a clown show. (Who can forget "The Republican Death Spiral" and its ~18 months of follow-ups?) His model swung wildly toward the GOP during the course of 2010, but he still ended up underestimating the GOP's House gains by over 20 percent. It was a mediocre effort.

But why wouldn't they be calling more people nowadays?

That costs money, too. I'm sure they'd like to be accurate but they don't want to go broke, either. Most pollsters aren't very forthcoming about response rates, so I'm not sure if there is much variance in the industry. Interestingly, there is a recent poll from a Democartic polling firm I had never heard of, showing Michigan tied. More relevant to this discussion, the poll actually disclosed that 35,000 calls were placed, and 1,122 respondents fully participated in the survey - that's a 3% response rate.

This is an interesting read about an incumbent on the ropes from a guy who's been there, done that.

I'm really confused about the polls right now. I keep waiting for someone like Silver to really get into the problem, but so far I haven't seen it adequately addressed anywhere.

The disagreement between the state and national polls is systematic and it is super weird.

The national polling suggest a race that is leaning Romney by a point or so. The state polling suggests that Obama has a clear but small lead of three points or so. This is not an "Ohio problem" - the state polling in pretty much all the swing states (NV, WI, NH, IA, VA, CO, NC, etc) are consistent with an O+3 race, not with a tied or R+1 race. I don't know if this has ever happened before, but I don't know of any parallels.

I can think of four explanations.

(1) Random variation. The state polls just happen to be giving us more prObama data, and the national polls just happen to be giving us better data for Romney. We can average the two together, which gives us a projection of a smaller Obama lead (about +2, since there is a lot more data at the state level than at the national level). This is what the 538 model does - it takes the national and state polls as all equivalent data and averages them all together. It shows Obama as a reasonable favorite.

I think this explanation is probably still the best guess, but given the massive, massive n's involved, the thousands upon thousands of people called by national and state pollsters, it's really weird that they could diverge by 3-4 points.

(2) A massive electoral college / popular vote divergence. There is little enough polling data from states like Texas, Alabama, California, and Maryland that perhaps both the national polls and the swing state polls are correct. Perhaps the bluest states are closer than usual, and Romney is running up huge leads in the safest red states. Then Obama is a pretty big favoriate.

On the plus side, this would explain the gap between the state and national polling without turning to silly shouts of "polling bias" in either direction. I would really rather have an explanation that explains the gap. However, I'd think that this sort of gap should show up in the polling - there is non-zero polling of safe states - and I'd think someone would have tested by now to see if it's the case. I'm skeptical.

(3) and (4) are "polling bias". Either the state polls are systematically biased to Obama or the national polls are systematically biased to Romney. Both are extremely hard for me to accept. It's possible, as public opinion polling is really difficult, and "bias" could be as simple as mediocre survey design. But mediocre surveys should tend to be wrong in both directions equally, not in one direction at the state level (in all swing states) and in the other direction at the national level (among all polling firms).

So I think that Nate's explanation (1), it's just random variation, is probably the best explanation. But I don't find it satisfying. I'd like to see more work done on this.

PPP is basically the in-house pollster for unions, while RAND didn't exist before July 2012, UPI had no tracker in 2008, and Reuters is new for 2012 after partnering with Zogby in 2008. They might end up being right, but I don't blame RCP for not using them, especially the three with no pre-2012 track records.

This would be a reasonable argument if RCP had some objective measure of quality and bias that they screened polls for. But they don't; they include Zogby (who is just awful) and Rasmussen (who was awful and biased in 2010). Does RCP exclude any pollster for having a bad track record? Or do they not care how good or bad a pollster's track record is?

And Reuters uses Ipsos for their polling. RCP included Ipsos in their 2008 averages when Ipsos was polling for McClatchy.

Perhaps the bluest states are closer than usual, and Romney is running up huge leads in the safest red states.

Meh. Not that we should place too much emphasis on Rasmussen's swing state-average polling results, but IIRC the most recent one had Romney seven points ahead, as opposed to four points overall.

Here's another reason to burn Frank Newport at the stake: Obama's favorability rating, which stood at 53% three days ago, is down to 46%.

Well, I was never calling a sure thing, and I'm certainly nervous now.

The post-election polling recap will be more interesting than the actual election.

The post-election polling recap will be more interesting than the actual election.

After tomorrow night might we say the same thing about the World Series?

Over the past eight or ten elections, the GOP candidate has never underperformed in Ohio relative to the national popular vote, and the average has been a +2.0 gain. Nate currently has Romney performing at -4.0 in Ohio relative to the national trackers. Might happen, but I'm doubtful.

Interesting.

(2) A massive electoral college / popular vote divergence

This would be my guess, actually. I can definitely see a scenario in which Obama loses the pop but wins the EC.

Over the past eight or ten elections, the GOP candidate has never underperformed in Ohio relative to the national popular vote, and the average has been a +2.0 gain. Nate currently has Romney performing at -4.0 in Ohio relative to the national trackers. Might happen, but I'm doubtful.


Interesting.

Interesting, indeed. But one of those "precedents" was actually wrong. Anyone know which one? Matt? Jason?

1992, Andy?

re #5244: 1952 is wrong. Eisenhower brought in a GOP House.

And, going into the final stretch, the two campaigns appear to have precisely opposite strategies.

The Romney campaign is emphasizing momentum. Confidence. Even inevitability. You see it in their post-debate spin: “Mitt Romney did well enough that for the first time in six years, Romney folks emailed, ‘We’re going to win.’” You see it in their electoral college spin: “Seriously, 305 electoral votes,” an anonymous Romney adviser told Politico. You see it in their theory of the race — that we’re seeing a final break of unhappy independents toward the challenger, and that having permitted Romney to pass this commander-in-chief threshold, there’s really nothing Obama can do to salvage the election.

The Obama campaign is emphasizing how tight it is. How hard they’re going to have to fight this one out. How possible it is that the president might lose. Their latest fundraising e-mail, for instance, reads as desperate. It’s supposedly from Obama himself, and the headline is, “Stick with me.”  The first line is even graver: “I don’t want to lose this election.” Remember that this is coming from the campaign that won last night’s debate and clearly leads in the electoral college. They could just as easily have written an e-mail entitled, “We’re going to win this thing!”

What you’re seeing here are two very different views of base psychology. The Romney campaign believes — and polling confirms — that Republicans are fired up to fire the president. They don’t need to worry about voter enthusiasm. But they do worry about voter confidence. If Republicans don’t believe they can win, they may not turn out to the polls. They don’t like the former Massachusetts governor enough to turn out on his behalf. So as they see it, confidence is their friend: Every Republicans wants to say they helped turn Obama out of office.

The Obama campaign believes — and polling confirms — that Democrats aren’t particularly fired up about the president. But they’re very fired up by the idea that Romney might become president. For the Obama campaign, then, voter enthusiasm can be squelched by voter confidence: If Democrats don’t think Romney can win, they may not be motivated to vote.

Until recently, this had Chicago pretty worried. Democrats have long seemed to believe they’ve got this election in the bag. That changed with the first debate (in part because the first debate really did make it likelier that Romney would win the election). So rather than comfort worried Democrats or strut about their debate win, the Obama campaign, in the final weeks of the election, is trying to scare its base, to persuade them that Republicans really could retake the White House. That gives their people a reason to go vote. And if their people have a reason to vote, the Obama campaign’s ground game will do the rest.

The bottom line is that Boston fears scared Republicans won’t vote and Chicago fears confident Democrats won’t vote. And so, in this final stretch, Boston wants Republicans confident and Chicago wants Democrats scared. Keep that in mind as you read the spin.

http://www.washingtonpost.com/blogs/ezra-klein/wp/2012/10/23/mitt-romney-wants-republicans-confident-president-obama-wants-democrats-scared/

Bad things are afoot at the Circle K.

re #5244: 1952 is wrong. Eisenhower brought in a GOP House.

Very good, fellow Bob Kuzava fan.

And that Ezra Klein take on the two campaign's respective motivation strategies is spot on, at least if the dozen e-mails a day I've been getting lately is any indication. They've sure scared me into coughing up more money than I thought I ever would. The thought of four more years of the Bush crowd running things again is more than enough to horrify me.

The bottom line is that Boston fears scared Republicans won’t vote and Chicago fears confident Democrats won’t vote. And so, in this final stretch, Boston wants Republicans confident and Chicago wants Democrats scared.

Makes sense. Presidential elections rely on voter turnout more than changing anybody's mind. And you get the most turnout when it's closest to a coinflip. It probably helps both sides that the media has its own reasons to want campaigns to appear as close as possible.

That costs money, too. I'm sure they'd like to be accurate but they don't want to go broke, either. Most pollsters aren't very forthcoming about response rates, so I'm not sure if there is much variance in the industry. Interestingly, there is a recent poll from a Democartic polling firm I had never heard of, showing Michigan tied. More relevant to this discussion, the poll actually disclosed that 35,000 calls were placed, and 1,122 respondents fully participated in the survey - that's a 3% response rate.

They're autodialing and if someone picks up and wishes to partake they a routed to a human being or they might not even get a human being. 20 years ago it was a human dialing each every person.

This would be my guess, actually. I can definitely see a scenario in which Obama loses the pop but wins the EC.

Obama is getting absolutely crushed in the South so it is quite possible for Romney to get a ton of useless votes in the South, win the popular vote, and still lose the election in the electoral college.

If Obama's name was Barry Harrison he probably wins over McCain in 2008 and probably wins the 2012 election by 5 or so points.

Or if those Kenyan Mooslim schemers had at least thought to name him Jefferson Davis Obama instead of naming him after that terrorist guy.

The Des Moines Register has endorsed Romney. They hadn't made a GOP endorsement in a Presidential Election in 40 years.

Orlando and Houston papers went Obama 2008 and now Romney 2012, iirc.
Of course, Salt Lake City paper went ... Obama.

Any other interesting ones, granting the unlikelihood that they actually mean anything?

Nate's analysis of the GOP's prospects in 2009 and early '10 was, to be polite, a clown show. (Who can forget "The Republican Death Spiral" and its ~18 months of follow-ups?)

I'm not sure how a January 2009 column, even one with a provocative title, renders Silver's grade-A projections 22 months later "mediocre" and "substantially wrong." But as long as you're satisfied, that's all that matters.

His model swung wildly toward the GOP during the course of 2010, but he still ended up underestimating the GOP's House gains by over 20 percent. It was a mediocre effort.

Over 20 percent, eh? Or to put it another way, Silver scored 426-for-435. Who can say which is the more sensible way to look at it?

Somehow it seems like missing 9 of 63 seat pickups isn't a failure rate of over 20 percent, but then, any math associated with Nate Silver is automatically suspect.

Bad things are afoot at the Circle K.

That is some scary #### there. That's a good reminder for people on both sides of the aisle here what real extremism looks like.

You're killing the narrative!

The Des Moines Register has endorsed Romney. They hadn't made a GOP endorsement in a Presidential Election in 40 years.

From the endorsement:

The president’s prescription upon entering office was a dose of government stimulus, which was the right call because it put cash in the pockets of consumers, made investments in vital infrastructure and kept millions of teachers and police officers on the job.

That stimulus was necessary to bridge the nation from recession to recovery, but the time is past for more government stimulus.

I don't recall Obama saying that more stimulus is necessary; if he was right with the first moves he made then why not have a little confidence that he'll make the correct moves in the future?

There is not a lot of difference between the two candidates’ short-term economic plans, as both are heavy on a promise of tax cuts for the middle class but short on details. Romney’s plan, however, goes beyond helping the middle class with tax breaks.

Yes, it includes giving the rich significant tax breaks while providing nebulous promises on how to pay for it without creating an impossible budget or one that hurts the lower and middle classes.

Romney has a strategy for job growth through tax and regulatory relief for small businesses, encouraging all forms of domestic energy production, education that prepares graduates with job skills, expanding foreign trade and reducing the burden of federal deficits.

He SAYS he has a strategy for all this but all we actually know about are tax cuts and eliminating regulations on banks and Wall Street. The implied criticism earlier in the article was that Obama hasn't shown a real plan for improving the economy - like anything better can be said about Romney?

Since then, he has recalibrated his campaign to focus on his concern for the middle class, and that is believable if the real Mitt Romney is the one on display as governor of Massachusetts who passed a health care reform plan that became the model for the one passed by Congress.

A health care reform plan he promised to repeal most of.

Romney should not squander an opportunity to build consensus in Washington by wasting time on issues that animate many in his party. We cannot rewind the clock on progress for minorities, women, gays and lesbians. We must make it easier for immigrants to come here to live and work legally and stop making criminals of those who are living here lawfully, paying taxes and raising families. The federal government must continue to insist on clean air and water and encourage clean and renewable energy.

The problem is that he isn't planning to do the things you're advising him to do.

Seriously, that's the worst endorsement of a presidential candidate I've ever read. It's like they gave the job of writing it to one of the board members that wanted to endorse Obama.

That is some scary #### there. That's a good reminder for people on both sides of the aisle here what real extremism looks like.

People who wander around in a haze of luxury and privilege talking about the elimination of "government" and "the state" should pay close attention to what happens when the state collapses.

I thought the reading of "The Protocols of the Elders of Zion" during a session of the Greek government was a nice touch.

People who wander around in a haze of luxury and privilege talking about the elimination of "government" and "the state" should pay close attention to what happens when the state collapses.

The Greeks shrank their government and then drowned it in Agamemnon's bathtub.

People who wander around in a haze of luxury and privilege talking about the elimination of "government" and "the state" should pay close attention to what happens when the state collapses.

Except the people who do that are generally not the ones that get the attention by the hate groups that's is part of the reason why they advocate less government.

I thought the reading of "The Protocols of the Elders of Zion" during a session of the Greek government was a nice touch.

Fascism is a simple, brutally elegant theory of politics. Identify the other. Then kill them. When you pull the veneer back, this is what the monkey does. It does not prance merrily about, shaking hands and congratulating you on Junior's lovely showing in the equestrian events while quietly assuming "natural law" is in charge.

What you're seeing bubble to the surface in Greece, as the Greek state collapses (under the demands of global financial "laws" that have no humanity in them at all) _is natural law._

as the Greek state collapses (under the demands of global financial "laws" that have no humanity in them at all)

Oh, the humanity! Should one consider Greece to be the metaphorical Hindenburg, an empty shell puffed up with little more than gas, now crashing and burning?

All that money should never have been forced on Greece at advantageous rates in the first place.

What you're seeing bubble to the surface in Greece, as the Greek state collapses (under the demands of global financial "laws" that have no humanity in them at all) _is natural law._

All you're seeing is the failure of the mainstream parties. If the any mainstream party was willing to default on their debt, leave the Euro, and impose fiscal reform, the wackos would be nowhere to be seen.

The problem in Greece is the elites would rather sacrifice the Greek people on the altar of the Euro and European unity, than do what's best for their citizens. So, the people have no other choice besides the fascists.

But that is what the elites will always do, if they are allowed to. They have to be made that they have something to lose. And that will come about through political process, not the economic one. Their position, economically and otherwise, has to be at play, and that has to be through a supervening political system.

But that is what the elites will always do, if they are allowed to.

Yes. The elites create a process by which they gain all of the "winnings" but socialize all of the "losings," then tut-tut about those terrible untermensch and their animalistic ways when the wheels come off. (The undermensch, for their part, being ####### idiots, attack some hated, powerless minority rather than the elites, because water flows downhill, and sewers are just ####-filled streams.)

Just wanted to say that MCoA's #5236 expresses very well several puzzlements I have vaguely had about the polls and the campaign, and many others that hadn't even occurred to me. Excellent analysis.

Just wanted to say that MCoA's #5236 expresses very well several puzzlements I have vaguely had about the polls and the campaign, and many others that hadn't even occurred to me. Excellent analysis.

concur; thanks for that well thought-out post

But that is what the elites will always do, if they are allowed to. They have to be made that they have something to lose. And that will come about through political process, not the economic one. Their position, economically and otherwise, has to be at play, and that has to be through a supervening political system.

That depends. It is absolutely what they will do when they no longer view themselves as primarily Greek, or French, or American, but rather "European", or "Citizens of the world".

It's a major flaw in the globalization that everyone is so gung ho for. The elites in NY, Washington, London or Paris no longer feel any kinship with the schlub working on the assembly line in Akron, Sheffield or Lille. They feel most connected to each other.

Just wanted to say that MCoA's #5236 expresses very well several puzzlements I have vaguely had about the polls and the campaign, and many others that hadn't even occurred to me. Excellent analysis.

Contra Nate and the quants, I personally think this is the primary disconnect between national/popular vote and state/electoral college polls:

(2) A massive electoral college / popular vote divergence. There is little enough polling data from states like Texas, Alabama, California, and Maryland that perhaps both the national polls and the swing state polls are correct. Perhaps the bluest states are closer than usual, and Romney is running up huge leads in the safest red states.

I think this, and the quants' aversion to this possibility, bleeds into the only real "polling bias" that exists. The bias that suggests that the numbers should align and if they don't, there's something wrong with the sample. I would call this the homo economicus bias.

I think Nate lives in a safe blue state. I don't know that he travels extensively in the great red wastelands of Flyover and The South. Mitt Romney is going to win the deep red states by massive margins. He will lose the blue states by much smaller margins. He seems to be on track to win the swing states by a few points.

When this happens, do not underestimate the rage you will see from the red wastes.

The elites in NY, Washington, London or Paris no longer feel any kinship with the schlub working on the assembly line in Akron, Sheffield or Lille.

Yeah, the way they used to, back in the good old days of 1835. Oh, wait --

Biden is headed to Pennslyvania on Thursday.

Andy, JFK didn't win a majority of the Catholic vote?

It's a major flaw in the globalization that everyone is so gung ho for. The elites in NY, Washington, London or Paris no longer feel any kinship with the schlub working on the assembly line in Akron, Sheffield or Lille. They feel most connected to each other.

In the world where the business elites used to "feel kinship" with the guys on the assembly line, is that another part of the Republican 1950s fantasy world they always go on about?

Perhaps the bluest states are closer than usual, and Romney is running up huge leads in the safest red states.

As of 10/16, East, Obama +4; Midwest, Obama +4; West, Obama +6; South, Obama -22.

In the world where the business elites used to "feel kinship" with the guys on the assembly line, is that another part of the Republican 1950s fantasy world they always go on about?

Well, you didn't see GM or IBM moving plants to Mexico in the 1950's and 1960's.

All you're seeing is the failure of the mainstream parties. If the any mainstream party was willing to default on their debt, leave the Euro, and impose fiscal reform, the wackos would be nowhere to be seen.

I'm not sure how Greece gets food, medicine, and oil in this scenario, but it's a pleasant narrative, I guess.

Biden is headed to Pennslyvania on Thursday.

Is this supposed to be predictive, Jason?

That depends. It is absolutely what they will do when they no longer view themselves as primarily Greek, or French, or American, but rather "European", or "Citizens of the world".

There's nothing immutable about most any classification you can think of. The same thing was said by Spartans and Athenians or any other sub-groups of a developing supervening group, all the way down to tribes, clans, families and marriages. Someone and some group is always dissatisfied. You gotta pay for what you get. There is no free lunch--and that includes the Edenic state.

This outta get Ray's goat:

New York subway to shut down as Sandy nears

Stupid thing I had never heard before this week: even in states that Romney has no chance of winning (California, New York) Democrats should still vote, and vote for Obama, because if he loses the popular vote while winning the EC, then Republicans will engage in shenanigans of some kind, and... well, OK, at that point I couldn't really follow it. We're not switching to election by popular vote sometime in the next week, are we?

Republicans will engage in shennanigans of some kind regardless of what form an Obama win takes.

All you're seeing is the failure of the mainstream parties. If the any mainstream party was willing to default on their debt, leave the Euro, and impose fiscal reform, the wackos would be nowhere to be seen.

I'm not sure how Greece gets food, medicine, and oil in this scenario, but it's a pleasant narrative, I guess.

See, here's the thing, though: you're both actually right. Greece is going to continue getting worse until they hit the trough. Unemployment in Greece is up over 20%. Since there's no tolerance in Brussels or Frankfurt for the kind of long-term wealth transfers that would enable Greece to stabilize, the only rational decision is to default and devalue. Greece has a great tourist and shipping industry. The nature of Greece's tourism doesn't really depend on the Euro.

But you're right in that there will be real, and serious issues with leaving the Euro. Goods will be disrupted for ordinary citizens. lives will be ruined. It's a really terrible example of what happens when you outsource monetary policy.

I'm not sure how a January 2009 column, even one with a provocative title, renders Silver's grade-A projections 22 months later "mediocre" and "substantially wrong." But as long as you're satisfied, that's all that matters.

Nate's 2009 and early '10 analysis didn't render his 2010 projections to be "mediocre," but it was evidence that Nate is much better with numbers than he is with reading political tea leaves, and to the extent the latter helps the former, it brings the subjective parts of his methodology into question.

Beyond that, if you're running a projection system and you're not only the last to get to the right position but you still underestimate the final plus/minus result by roughly 20 percent, your projection system had a mediocre year.

Over 20 percent, eh? Or to put it another way, Silver scored 426-for-435. Who can say which is the more sensible way to look at it?

Somehow it seems like missing 9 of 63 seat pickups isn't a failure rate of over 20 percent, but then, any math associated with Nate Silver is automatically suspect.

To each his own, but it seems strange to give Nate credit for "predicting" 250 or 300 House races that weren't really competitive in the first place. The point of Nate's model for the House was to predict which party would control it and then to project the net gain or loss for that party. A week or two before Election Day 2010, I believe Nate only had the GOP as a ~70 percent favorite to take the House, and even when his model converged with the consensus, he still underestimated the GOP's gains by ~12 seats.

This is why conservatives are much less likely to see Nate's model as anything close to infallible. Around this point in 2010, Nate had the GOP as a lesser favorite to retake the House (~70 percent) than he currently has Obama winning reelection (73.6 percent as of right now). If Nate was behind the curve during a wave election, it's unclear why people are so confident in his projection abilities during one of the closest elections in decades.

This outta get Ray's goat:

Ray says they bought their ticket. Let them die.

[EDITED for clarity.]

Stupid thing I had never heard before this week: even in states that Romney has no chance of winning (California, New York) Democrats should still vote, and vote for Obama, because if he loses the popular vote while winning the EC, then Republicans will engage in shenanigans of some kind, and... well, OK, at that point I couldn't really follow it. We're not switching to election by popular vote sometime in the next week, are we?

Never underestimate the GOP's ability to do the stupid. The fact that Bill Clinton only won a plurality in '92 was the initial pivot that eventually begat impeachment. And I assure you that if Obama wins via EC but loses the popular vote you will hear nothing but "he has no real mandate" for at least two years.

I love the smell of feudalism in the morning...

This would be my guess, actually. I can definitely see a scenario in which Obama loses the pop but wins the EC.

I can, too, but not by more than 0.5 to 1 point. Does anyone here believe Obama could lose the popular vote by more than that and still win the Electoral College? If so, what's the number -- 1.5? 2?

***

Stupid thing I had never heard before this week: even in states that Romney has no chance of winning (California, New York) Democrats should still vote, and vote for Obama, because if he loses the popular vote while winning the EC, then Republicans will engage in shenanigans of some kind, and... well, OK, at that point I couldn't really follow it. We're not switching to election by popular vote sometime in the next week, are we?

The argument is probably that a president who loses the popular vote but wins the Electoral College is seen in some corners as illegitimate. I didn't buy into this with Bush in 2000 and I won't buy into it if Obama wins this way in 2012, but the position got widespread play in 2000.

Is this supposed to be predictive, Jason?

I don't know why the Veep is scheduled to spend time in Pannsylvania a mere 96 hours before the election, Sam. We'll need to ask Chicago, I suppose.

Stupid thing I had never heard before this week: even in states that Romney has no chance of winning (California, New York) Democrats should still vote, and vote for Obama, because if he loses the popular vote while winning the EC, then Republicans will engage in shenanigans of some kind, and... well, OK, at that point I couldn't really follow it. We're not switching to election by popular vote sometime in the next week, are we?

The point of that argument isn't about who would win the 2012 election. The point of it is the entirely reasonable proposition that a split decision (so to speak) would make the last four years seem like a bipartisan lovefest compared to what would ensue between now and 2016. And of course it also raises the prospect of a 5-4 Supreme Court possibly deciding yet another election.

OTOH the upside would be that such a split might enable a constitutional amendment to elect the president by popular vote pick up support from Republicans who otherwise would never support it. It's all pure speculation, but at least that's what the argument is.

And I assure you that if Obama wins via EC but loses the popular vote you will hear nothing but "he has no real mandate" for at least two years.

If Obama squeaks back in with a 281-257 win in the Electoral College while the GOP holds the House and perhaps gains a seat in the Senate, Obama won't have anything resembling a mandate, whether he wins the popular vote or not.

(And if Obama allegedly would have a mandate, then a mandate for what? Obama's entire campaign has been about bashing Romney rather than laying out proposals for the next four years. Even the Washington Post laughed at that agenda booklet he put out last week.)

I don't know why the Veep is scheduled to spend time in Pannsylvania a mere 96 hours before the election, Sam. We'll need to ask Chicago, I suppose.

Do you really think PA is in play? He's probably going to a location near the Ohio border.

I'm happy to wager a $20.00 BBREF sponsorship on the outcome of PA, if you really think that it means anything, that is.

edit:

If Obama squeaks back in with a 281-257 win in the Electoral College while the GOP holds the House and perhaps gains a seat in the Senate, Obama won't have anything resembling a mandate.

Bush in 2004: 286 EC Votes. Bush in 2000: 271. Loud claims of a mandate? Each year.

Bush in January, 2005: "Let me put it to you this way: I earned capital in the campaign, political capital, and now I intend to spend it."

I can, too, but not by more than 0.5 to 1 point. Does anyone here believe Obama could lose the popular vote by more than that and still win the Electoral College?

I don't see either side winning the popular vote by more than 2% (like, 50-48-2).

I don't know why the Veep is scheduled to spend time in Pannsylvania a mere 96 hours before the election, Sam.

Are you sure you didn't mean Jill Biden? I can't find anything about Joe going. Jill's is now cancelled.

elect the president by popular vote pick up support from Republicans who otherwise would never support it

That's a nightmare waiting to happen.

Imagine the delay when you have to have recounts in 50 states because the final popular vote was within the margin that requires one (say, 0.5%).
You need some sort of compartmentalization to stop the recount flood from overwhelming the country.
With the EC, you can limit it to one state, maybe two at most (if R wins state A OR state B, the result flips, and both are within the recall margin).

Without it, you're forced to recount at EVERY SINGLE VOTING STATION.

Bush in 2004: 286 EC Votes. Bush in 2000: 271. Loud claims of a mandate? Each year.

Bush in January, 2005: "Let me put it to you this way: I earned capital in the campaign, political capital, and now I intend to spend it."

And he was absolutely right. Not only did Bush win in 2004, but the GOP held the House (and gained three seats) and held the Senate (and gained four seats). When the voting public gives your party full control of the federal government, that's a mandate.

And he was absolutely right. Not only did Bush win in 2004, but the GOP held the House (and gained three seats) and held the Senate (and gained four seats). When the voting public gives your party full control of the federal government, that's a mandate.

So Obama had a clear mandate for Health Care Reform in 2008, right?

flip