Baseball for the Thinking Fan

Login | Register | Feedback

btf_logo
You are here > Home > Baseball Newsstand > Baseball Primer Newsblog > Discussion
Baseball Primer Newsblog
— The Best News Links from the Baseball Newsstand

Wednesday, October 31, 2012

OTP November 2012 - Moneypoll! The Pundits vs. The Election-Data Nerds

Come next Tuesday night, we’ll get a resolution (let’s hope) to a great ongoing battle of 2012: not just the Presidential election between Barack Obama and Mitt Romney, but the one between the pundits trying to analyze that race with their guts and a new breed of statistics gurus trying to forecast it with data.

In Election 2012 as seen by the pundits–political journalists on the trail, commentators in cable-news studios–the campaign is a jump ball. There’s a slight lead for Mitt Romney in national polls and slight leads for Barack Obama in swing-state polls, and no good way of predicting next Tuesday’s outcome beyond flipping a coin. ...

Bonus link: Esquire - The Enemies of Nate Silver

Joe Kehoskie Posted: October 31, 2012 at 11:42 PM | 11298 comment(s) Login to Bookmark
  Tags: mr president, off-topic, politics, sabermetrics, usa

Reader Comments and Retorts

Go to end of page

Statements posted here are those of our readers and do not represent the BaseballThinkFactory. Names are provided by the poster and are not verified. We ask that posters follow our submission policy. Please report any inappropriate comments.

Page 10 of 114 pages ‹ First  < 5 6 7 8 9 10 11 12 13 14 15 >  Last ›
   901. Tripon Posted: November 03, 2012 at 11:41 AM (#4291597)

Not snark, question: $25 per gallon gas solves which problem in NJ at the moment?


It would probably lead to more gun fights. The issue with Stossel's position is that pretending that a market created by natural disasters is 'normal' is incredibly asinine.
   902. Gold Star - just Gold Star Posted: November 03, 2012 at 11:42 AM (#4291598)
Ryan camp already talking (off the record) about what he does after Romney loses:
A return also would make Ryan a leading target for Democrats. For the next few years, Democrats would lay traps in legislation, forcing him to take sides on measures that could come back to haunt him during a presidential bid.

That is why some of Ryan's biggest boosters are considering whether it wouldn't be better for Ryan to resign from the House. He could write a book — "saving America" is a theme often bandied about — or teach at a university.
   903. McCoy Posted: November 03, 2012 at 11:44 AM (#4291599)
Not snark, question: $25 per gallon gas solves which problem in NJ at the moment?



Gets a ton of gas to NJ doesn't it? It also creates an incentive for shop owners to get there stores open one way or another. Lowers demand for gas.

It would probably lead to more gun fights. The issue with Stossel's position is that pretending that a market created by natural disasters is 'normal' is incredibly asinine.


Charging $5 a gallon for a limited supply of gas is not the answer nor does putting a cap on gas create a desire for suppliers to spend money to get the gas to NJ.
   904. Lassus Posted: November 03, 2012 at 11:51 AM (#4291601)
As the kids say: That awkward moment you keep refreshing the page multiple times looking for responses before you realize your entry is the last one on the page.

Anyhow, I've heard the arguments for the free market plenty of times, I guess. I'm more curious out of morbid fascination than anything else to see if a $30 per gallon price helps anyone or anything in NJ. Not as convinced, obviously.

   905. Jolly Old St. Nick Still Gags in October Posted: November 03, 2012 at 11:55 AM (#4291606)
Not snark, question: $25 per gallon gas solves which problem in NJ at the moment?


Gets a ton of gas to NJ doesn't it? It also creates an incentive for shop owners to get there stores open one way or another.

My heart's with Christie, but my head's with Stossel. It's hard to imagine how anyone would want to transport gas from faraway places to New Jersey if they thought they couldn't make a profit on it, especially considering that under normal circumstances New Jersey's gas prices are usually about the lowest anywhere in the East.


   906. Jack Keefe Posted: November 03, 2012 at 11:58 AM (#4291608)
Well Al it is all most all over but the counting and today I voted in Terre Haute using Early Voting but I had to leaf town at once and so I send you this post Al to let you know I was there and my Mam is well she is into Yarn Art these days Al you should buy some. I went to the Poles and you know what I had left my Wallit at Ma's. So no Voter ID no Vote and tho Indiana is redder than A.J. Pierogi's hindquarters I wanted to cask my vote for Barack Obrella all the same Al it is my Rite. So they set up an Indentity Test for me. Judge Richter made me set up 60'6" from the booth and Mrs. Fanger from the Pubic Library grabbed a Oven Mitt and the Judge said smoke 1 in here Jacko and well know if its you or not. Al I did not want to knock nice Mrs Fanger over with high Heat so I threw my drop curve and it caught the plate but Judge Richter needs new Mutlifocals Al he called it Ball One and said that's the Keefe I know step right up to the old Diebold my boy he is quite the Jokester Al it made me sore but I voted any way and that is Dimocracy at work.
   907. Random Transaction Generator Posted: November 03, 2012 at 12:00 PM (#4291609)
All this New Jersey gas talk made me realize they just recently passed the law to allow people to pump their own gasoline.
   908. Tripon Posted: November 03, 2012 at 12:04 PM (#4291611)
My heart's with Christie, but my head's with Stossel. It's hard to imagine how anyone would want to transport gas from faraway places to New Jersey if they thought they couldn't make a profit on it, especially considering that under normal circumstances New Jersey's gas prices are usually about the lowest anywhere in the East.


The issue is one of supply. Not just of gas, but of gas stations.

Sal Risalvato, executive director of the New Jersey Gasoline, Convenience Store and Automotive Association, told FoxNews.com that virtually “nothing [has] been restored,” adding that up to 80 percent of gas stations in the northern part of the state either don’t have gas to sell or the power to dispense it.
Damage from the storm that killed at least 90 people has forced many gas stations to close and disrupted fuel deliveries to those that remained open. The American Automobile Association estimates that 60 percent of gas stations in New Jersey are shuttered, compared to 70 percent with those on New York's Long Island.


Read more: http://www.foxnews.com/us/2012/11/02/authorities-work-restore-power-gas-stations-in-sandy-aftermath/#ixzz2BAyoFECH


Another thing Stossel is saying that people are waiting in line to top off their gas. Like anyone would wait two hours to pull off that ########.
   909. JL Posted: November 03, 2012 at 12:06 PM (#4291612)
Stossel loves his free markets but has no problem milking the government. He was on O'Reilly the other day bemoaning government flood insurance. He then admitted that he took over $250k some years ago from FEMA. He pledged that he would not do it again, but then told O'Reilly he would not return it because government is wasteful. So you only should pay back ill gotten money if the person you took it from is not wasteful? That is a great philosophy.
   910. Shredder Posted: November 03, 2012 at 12:12 PM (#4291614)
He's right. Price fixing is stupid and detrimental to solving a problem.

Not snark, question: $25 per gallon gas solves which problem in NJ at the moment?
It solves the problem of rich people having to wait in line for too long because too many not-rich people are buying gas.
   911. Morty Causa Posted: November 03, 2012 at 12:12 PM (#4291615)
   912. AROM Posted: November 03, 2012 at 12:13 PM (#4291616)
Took a look at Silver's site and Real Clear Politics, put them in a spreadsheet state by state.

The easy states to project give Obama 201, Romney 191. That leaves these as battlegrounds:

CO, FL, IA, MI, NC, NH, NV, OH, PA, VA, WI

Of those, Silver has Romney ahead in only North Carolina and Florida. Real Clear Politics agrees in every case (in verdict if not in margin) except Virginia.

Nate's count shows Obama winning 303-235 (his site projects 305-233, close enough, the difference being the relative odds of some states ending up differently than the current poll leader).

RCP would have Obama winning 290-248. Romney's chance pretty much means he needs:

1. Close Florida/Virginia poll advantages to be real and hold up
2. Win Ohio
3. Win another state that Obama leads - one of New Hampshire, Iowa, or Colorado most likely.

Or the polls to be so f***ed that they are meaningless.

I don't like expressing a once in 4 years event in percentages like that, and the idea that the outcome of that single event can be used to validate or discredit the system. But the math does work out. The odds of Romney winning are probably equivalent to the odds of Obama winning Florida, holding his poll leads in the other battlegrounds, and topping 330 electoral votes.
   913. Jolly Old St. Nick Still Gags in October Posted: November 03, 2012 at 12:14 PM (#4291618)
My heart's with Christie, but my head's with Stossel. It's hard to imagine how anyone would want to transport gas from faraway places to New Jersey if they thought they couldn't make a profit on it, especially considering that under normal circumstances New Jersey's gas prices are usually about the lowest anywhere in the East.

The issue is one of supply. Not just of gas, but of gas stations.

Sal Risalvato, executive director of the New Jersey Gasoline, Convenience Store and Automotive Association, told FoxNews.com that virtually “nothing [has] been restored,” adding that up to 80 percent of gas stations in the northern part of the state either don’t have gas to sell or the power to dispense it.
Damage from the storm that killed at least 90 people has forced many gas stations to close and disrupted fuel deliveries to those that remained open. The American Automobile Association estimates that 60 percent of gas stations in New Jersey are shuttered, compared to 70 percent with those on New York's Long Island.


That's a fair point in this case, at least as it stands today.

But two related questions: Did Christie emphasize that point in his statement condemning "price gouging"? And once the gas stations' power is restored, will there still be a problem of these stations receiving a normal gas supply at "normal" wholesale prices?
   914. Shredder Posted: November 03, 2012 at 12:17 PM (#4291620)
And after all of this I probably won't even end up voting. I'm stuck in California with my dad in ICU (hopefully) recovering from a very serious surgery. I was supposed to fly back today, but now won't head back before Wednesday. Barack will have to find a way to carry Illinois without me.
   915. McCoy Posted: November 03, 2012 at 12:19 PM (#4291621)
The issue is one of supply. Not just of gas, but of gas stations.

A gas station that can charge more for gas can find ways to get their power on so that they can pump and sell their gas. Hell, if you could sell gas for 25 to 30 dollars a gallon you could set up a home delivery system where you go to people's homes or designated areas and pump gas out of drums or tankers. That isn't an option when you can only charge $5 a gallon.

Another thing Stossel is saying that people are waiting in line to top off their gas. Like anyone would wait two hours to pull off that ########.


I doubt people are waiting 2 hours or more to put a gallon or two in their cars but I'm also sure that people are getting in line with cars that are half full or better and I'm also sure that people with their gas guzzlers are making sure they are filled to the brim when they do go to get gas and probably fill up a few gas containers as well. Charging $5 a gallon causes people to horde because there is no disincentive not to while there is an incentive to horde.
   916. formerly dp Posted: November 03, 2012 at 12:27 PM (#4291622)
I'm also sure that people are getting in line with cars that are half full or better and I'm also sure that people with their gas guzzlers are making sure they filled to the brim when they do go to get gas and probably fill up a few gas containers as well.


Gouging solves this problem.
   917. Shredder Posted: November 03, 2012 at 12:33 PM (#4291624)
Charging $5 a gallon causes people to horde because there is no disincentive not to while there is an incentive to horde
There's a very simple solution to this: Limit X gallons per customer. Set the pumps to shut off after 10 gallons, or something. I buy a lot of limited release beers and this system seems to work fine.
   918. Gold Star - just Gold Star Posted: November 03, 2012 at 12:34 PM (#4291625)
My Florida tweeps are reporting polling-place lines of up to 5 hours.
   919. OCF Posted: November 03, 2012 at 12:41 PM (#4291629)
As Nate wrote this morning, Obama wins unless the polls are systematically wrong - his 16% or so chance for Romney is pretty much all in the "polls are biased" basket.

So how could the polls be systematically wrong?

There's general response bias. I'll admit that I'm unpollable - I have Caller ID and I won't pick up the phone. I suspect that that's extremely common across many segments of the population. I don't know which way it skews. It might be near neutral, but who can be sure?

There are two more specific issues with response bias. One is the issue of cell phones versus landlines. Some polls don't call cell phones at all; some do try to call cell phones (although I suspect that cell phone users are on average less likely to pick up than land line users.) So there's probably an underrepresentation of cell phone users in the sampling, and that skews younger, and skewing younger should be correlated with being an Obama supporter.

The other issue is reaching Latino voters, especially Latinos would would be more comfortable speaking Spanish. (And not just Latinos - possibly also Chinese, Korean, etc. voters who may not be entirely comfortable speaking with an English-speaking pollster even if they do speak English.) This is not equally distributed across states, and much of the Latino (and Asian) population is concentrated in safe-D states like California. But it might matter in Nevada and Florida (yes, there are a lot of non-Cuban Latinos in Florida). And if it does matter, it probably means that Obama could do better than the polls.

Then there's people being less than totally honest with pollsters. This also comes in more than one flavor.

One would be people lying about who they actually support - this is the Bradley effect, or Shy Tory effect. Frankly, I expect that there's none of it at the Presidential level, that it will be a non-issue. People are going to say who they're going to vote for. Maybe there's some rationalization built into their reasoning that obscures some factors - but they won't lie about the vote itself. If there's an issue here, it might be with the Missouri and Indiana Senate races - maybe people will be a little shy about admitting that they support Akin or Mourdock.

More important is whether people are giving misleading answers about whether they're likely to vote. I can see a lot people expressing absolute certainty about voting who wind up not doing so, and I can see people expressing hedges and uncertainties who will wind up voting. This could be sex-linked: I can see males being more likely to express certainty. A big story has been that the "likely voter" screens tend to give different results than the "registered voter" tabulations. And the likely voter screens could be off. And if they're off, Obama could do better than he polls in the LV models.

Put it all together, and where is the possible source of bias that would cause Romney to do better than he polls? The only place I see that is in the very first item: if 10% or less even answer polls, what do we really know about the other 90%? But other than that, and even there, there's no strong reason to think that it would skew for Romney. All of the other possible sources of bias would seem to either be neutral, or to potentially favor Obama.
   920. tshipman Posted: November 03, 2012 at 12:52 PM (#4291632)
A big story has been that the "likely voter" screens tend to give different results than the "registered voter" tabulations.


Huh? This is the case every year. If every registered voter actually voted, Obama would win by 10 points. It's that way every year.


Re: price fixing:

So, there are arguments on both sides. The argument for allowing any price is that it increases supply. People aren't willing to move heaven and earth to get gasoline up to New Jersey for $5.00 per gallon when they make almost zero profit on gas. If you can make windfall profits, then you'll have more supply quicker. In addition, by setting the price levels higher, you disincentivize hoarding. There are almost certainly large numbers of people who are waiting in line with jerry cans who will still have gasoline in those cans when service is back to normal. In addition, Gasoline is not necessary to survival, so letting prices float freely will result in greater efficiency and a greater likelihood that someone who really does need that gasoline has quicker access to it.

The arguments against mostly boil down to free-price floating in emergencies extends the impact of emergencies on the poor. In addition, you have a greater risk for breakdown of law and order and violent riots. No one ever discusses rioting in the context of price gouging, but it's a serious concern.
   921. McCoy Posted: November 03, 2012 at 12:54 PM (#4291634)
There's a very simple solution to this: Limit X gallons per customer. Set the pumps to shut off after 10 gallons, or something. I buy a lot of limited release beers and this system seems to work fine.


Well, that still doesn't address the no incentive to get the gas to the customer that price controls create. It also increase costs for the buyer. Plus a limit simply means that people have to get in line more often for gas or have more people committed to buying gas instead of it being one person's task.
   922. The Yankee Clapper Posted: November 03, 2012 at 12:54 PM (#4291636)
So has anyone in this thread gone on record as predicting a Romney win?

In due time . . still more than 48 hours until the polls open in Dixville Notch. However, I would highly recommend a separate prediction thread, without all the clutter that plagues this one.
   923. Joe Kehoskie Posted: November 03, 2012 at 12:55 PM (#4291638)
Ryan camp already talking (off the record) about what he does after Romney loses:

Total misrepresentation of a standard pre-election "what if?" story.
   924. Tilden Katz Posted: November 03, 2012 at 12:55 PM (#4291639)
The other issue is reaching Latino voters, especially Latinos would would be more comfortable speaking Spanish.


This is believed to be the reason Harry Reid did much better on Election Day then polls had him doing.
   925. McCoy Posted: November 03, 2012 at 12:58 PM (#4291642)
In addition, you have a greater risk for breakdown of law and order and violent riots. No one ever discusses rioting in the context of price gouging, but it's a serious concern.

If New Jersey motorists riot over the price of gasoline then the biggest problem you got going isn't the price of gasoline.

I'm not sure how $5 a gallon gas during disasters helps the poor when the poor are the least likely to be able to get the limited supply of gasoline or need it.
   926. Jim Wisinski Posted: November 03, 2012 at 12:58 PM (#4291643)
My Florida tweeps are reporting polling-place lines of up to 5 hours.


Absentee ballots, people! Seriously, I asked for mine, got it in the mail, then walked into the lobby of the local library to drop it off at a desk with nobody in line.
   927. snapper (history's 42nd greatest monster) Posted: November 03, 2012 at 12:58 PM (#4291644)
So has anyone in this thread gone on record as predicting a Romney win?


I think it's more likely than not. The polls are likely missing the enthusiasm/turnout gap, many by a substantial margin. The fundamentals are just too unfavorable for Obama.

Couple that with the fact that Obama has done nothing to explain why he deserves a second term, and Romney looking like a perfectly credible and safe alternative, I expect undecideds to break heavily for Romney, and a lot of soft Obama supporters to just stay home.

I'll say Romney 51-48, he takes OH, FL, VA, NH, CO, and one of IA or WI.
   928. BDC Posted: November 03, 2012 at 01:01 PM (#4291645)
I don't like expressing a once in 4 years event in percentages like that

Agreed. I think that a considerable amount of the #### that Silver takes comes from people who log in for a real quick look, see Obama "winning" 83.7 to 16.3 (like this morning), and think, "son, you're high."

   929. McCoy Posted: November 03, 2012 at 01:09 PM (#4291646)

Agreed. I think that a considerable amount of the #### that Silver takes comes from people who log in for a real quick look, see Obama "winning" 83.7 to 16.3 (like this morning), and think, "son, you're high."


So then what is the point of his site? Should he lie? Should he say the race is 50/50 to make people happy? Should he say the race is 60/40 because people can't understand odds?

Sure the decimals are stupid but that should be a minor quibble.
   930. Joe Kehoskie Posted: November 03, 2012 at 01:11 PM (#4291647)
A big story has been that the "likely voter" screens tend to give different results than the "registered voter" tabulations. And the likely voter screens could be off. And if they're off, Obama could do better than he polls in the LV models.

Put it all together, and where is the possible source of bias that would cause Romney to do better than he polls?

You listed it in the preceding paragraph, but from the wrong angle. Many if not most of the LV screens being used in recent polls are passing as many as 96 percent of respondents through as LV. Now, this might be an accurate poll of those respondents, but, given that the 2008 turnout was only ~72 percent, it can't be representative of the electorate as a whole.
   931. Howie Menckel Posted: November 03, 2012 at 01:12 PM (#4291648)
The NJ problem supposedly is NOT supply - the gas is here, but the stations don't have the electricity to pump it.

I filled up on Sunday with no waiting, and still have half a tank left in an area where 2- to 4-hour waits are common. I drive past them on my work commute each morning and night. miles and miles of cars, literally.

I think Silver's literal-mindedness is leading him to logical conclusions, but I don't believe that he or anyone else has such a strong grip on how accurate the state polls are, given these circumstances.

Also, props to him for pointing out the fact that Obama will win unless the polls are wrong. But that point seems to be lost on many: I wonder, in fact, if Nate would be a bit less shocked should they be wrong than more emotional Obama supporters would be.

I don't think it's disputable that if the state polls are right, Obama wins.

   932. snapper (history's 42nd greatest monster) Posted: November 03, 2012 at 01:14 PM (#4291649)
So then what is the point of his site? Should he lie? Should he say the race is 50/50 to make people happy? Should he say the race is 60/40 because people can't understand odds?

He should avoid precise percentages when talking about events that are too infrequent to estimates percentage odds reliably. It's like quantifying a teams chance of winning the World Series in percentages in April, it's stupid. Stick to toss-up, leans-O/R, likely O/R, very-likely O/R.
   933. McCoy Posted: November 03, 2012 at 01:14 PM (#4291650)
Couple that with the fact that Obama has done nothing to explain why he deserves a second term, and Romney looking like a perfectly credible and safe alternative, I expect undecideds to break heavily for Romney, and a lot of soft Obama supporters to just stay home.

I don't think Obama has done a great job explaining to the nation what he will do in his second term but I think a lof of that is because the nation doesn't really want to hear the truth. Nobody wants to be told they have to make sacrifices, that their taxes will go up and they will get less services. So politicians, all politicians, have to bribe people with their own money while also telling them that the government will take less money from them. If the electorate was actually an informed electorate that was capable of understanding nuanced positions and willing to let the information come out in an orderly fashion we would all be better off. But that isn't reality and so consequently we get sound bites and vague promises.

Having said what I said about Obama I don't think that means we vote for the "other guy" simply because OBama hasn't spelled out exactly what he is going to do in the next four years. The "other guy" hasn't spelled out what he is going to do either and his base is full of nutters and he has surrounded himself with a bunch of nutters. So that isn't promising.
   934. Random Transaction Generator Posted: November 03, 2012 at 01:15 PM (#4291651)
I'll say Romney 51-48, he takes OH, FL, VA, NH, CO, and one of IA or WI.


I respect a man who sticks by his convictions and goes all-in, even if it's not what I believe.
   935. McCoy Posted: November 03, 2012 at 01:16 PM (#4291653)
He should avoid precise percentages when talking about events that are too infrequent to estimates percentage odds reliably. It's like quantifying a teams chance of winning the World Series in percentages in April, it's stupid. Stick to toss-up, leans-O/R, likely O/R, very-likely O/R.

Because some human beings are incapable of understanding odds?

Likely this or lean that is meaningless gobblygook. What does it mean when someone says X is likely? It will happen 55% of the time? 58%? 90? What?
   936. Random Transaction Generator Posted: November 03, 2012 at 01:18 PM (#4291654)
However, I would highly recommend a separate prediction thread, without all the clutter that plagues this one.


I said the same thing a couple of pages ago, but people think ANOTHER OT thread would be too much on the side-bar.

Of course, it is the off-season so there really isn't THAT much else to talk about for the next few days...
   937. Swedish Chef Posted: November 03, 2012 at 01:18 PM (#4291655)
Also, props to him for pointing out the fact that Obama will win unless the polls are wrong.

I haven't followed the Nate Silver saga, but what added value does he claim to bring over just looking at the polls themselves?
   938. snapper (history's 42nd greatest monster) Posted: November 03, 2012 at 01:18 PM (#4291656)
Having said what I said about Obama I don't think that means we vote for the "other guy" simply because OBama hasn't spelled out exactly what he is going to do in the next four years. The "other guy" hasn't spelled out what he is going to do either and his base is full of nutters and he has surrounded himself with a bunch of nutters. So that isn't promising.

You don't, because you have strong ideological viewpoints. The middle 10% of the electorate doesn't have a strong ideology, or has mixed positions that cause them to waver (e.g. socially conservative union worker, or socially liberal rich suburbanite).

That type of voter will tend to "throw the bum out" if the fundamentals are poor, and the incumbent hasn't made a strong case for his re-election.
   939. McCoy Posted: November 03, 2012 at 01:19 PM (#4291657)
The NJ problem supposedly is NOT supply - the gas is here, but the stations don't have the electricity to pump it.

I filled up on Sunday with no waiting, and still have half a tank left in an area where 2- to 4-hour waits are common. I drive past them on my work commute each morning and night. miles and miles of cars, literally.


And again if a gas owner could charge X amount in order to make a profit he would have incentive to get his pumps working. As of right now it would cost more for a gas owner to pump gas than to sit idle and sell nothing. So the gas owner sits idle.
   940. Joe Kehoskie Posted: November 03, 2012 at 01:21 PM (#4291658)
As Nate wrote this morning, Obama wins unless the polls are systematically wrong - his 16% or so chance for Romney is pretty much all in the "polls are biased" basket.

So if Nate is right, it shows the genius of his model. But if Nate is wrong, it's because he was provided with faulty polling data. I have to give him credit: That's a hell of a racket he's crafted.
   941. McCoy Posted: November 03, 2012 at 01:22 PM (#4291659)
I haven't followed the Nate Silver saga, but what added value does he claim to bring over just looking at the polls themselves?

Earlier in the year he uses other factors besides just the polls to come up with his numbers while closer to the election it becomes more and more about the polls. By using state polls, nationals polls, and weighting them theoretically he can make them more accurate.

Think of it like sabermetrics. Palmer's linear weights might explain 93% of run scoring and BaseRuns can explain 95% of run scoring. It is pretty much the same thing with Nate and aggregators.
   942. snapper (history's 42nd greatest monster) Posted: November 03, 2012 at 01:22 PM (#4291660)
Because some human beings are incapable of understanding odds?

Likely this or lean that is meaningless gobblygook. What does it mean when someone says X is likely? It will happen 55% of the time? 58%? 90? What?


No. Because nobody knows the odds. We have no idea about the relationship between registered voters, likely voters, turnout by R/D/I because they are constantly changing (especially with cell-phone/voice-mail/call screening), and we can't measure enthusiasm well, b/c everyone tells the pollsters they plan to vote.

If MLB played 16 games a year, like football, a ZiPs type forceast would be assinine. We would literally have zero idea based on the stats who was good and who was bad.
   943. Howie Menckel Posted: November 03, 2012 at 01:23 PM (#4291661)

"As of right now it would cost more for a gas owner to pump gas than to sit idle and sell nothing. So the gas owner sits idle."

interesting.
can you provide a link? (apologies if you have already done so upthread and I missed it)
   944. snapper (history's 42nd greatest monster) Posted: November 03, 2012 at 01:23 PM (#4291662)
Think of it like sabermetrics. Palmer's linear weights might explain 93% of run scoring and BaseRuns can explain 95% of run scoring. It is pretty much the same thing with Nate and aggregators.

Just beat my comment. It's sabermetrics with a 10-game season; a woefully inadequate observation sample.
   945. Random Transaction Generator Posted: November 03, 2012 at 01:24 PM (#4291663)
But if Nate is wrong, it's because he was provided with faulty polling data. I have to give him credit: That's a hell of a racket he's crafted.


Well, not just Nate would be wrong.
RCP would be wrong.
Princeton election guys would be wrong.

All of the polls he used would be wrong.
   946. McCoy Posted: November 03, 2012 at 01:24 PM (#4291664)
You don't, because you have strong ideological viewpoints. The middle 10% of the electorate doesn't have a strong ideology, or has mixed positions that cause them to waver (e.g. socially conservative union worker, or socially liberal rich suburbanite).

That type of voter will tend to "throw the bum out" if the fundamentals are poor, and the incumbent hasn't made a strong case for his re-election.


They don't because the middle 10% are a bunch of morons. Again, if we had a patient and educated electorate almost all of the stupidity of the election season would go away.
   947. snapper (history's 42nd greatest monster) Posted: November 03, 2012 at 01:25 PM (#4291665)
"As of right now it would cost more for a gas owner to pump gas than to sit idle and sell nothing. So the gas owner sits idle."

interesting.
can you provide a link? (apologies if you have already done so upthread and I missed it)


Not my quote, but I have heard "talking-heads" on TV discussing the fact that for a gas station owner to bring in generators to get open would cost a lot, both in the generator rental, and the gasoline to power the generators.
   948. Random Transaction Generator Posted: November 03, 2012 at 01:25 PM (#4291666)
Just beat my comment. It's sabermetrics with a 10-game season; a woefully inadequate observation sample.


Except, it's not 10 games. It's hundreds of thousands of random games.
   949. Swedish Chef Posted: November 03, 2012 at 01:26 PM (#4291667)
All of the other possible sources of bias would seem to either be neutral, or to potentially favor Obama.

It doesn't work like that, the pollers process their numbers to account for biases in the raw data. If any fudge factor they're using is wrong it could swing towards either side.
   950. Joe Kehoskie Posted: November 03, 2012 at 01:26 PM (#4291668)
No. Because nobody knows the odds. ...

Exactly right. If we flip a coin, the odds of getting "heads" are 50-50, for reasons we all understand. But for Nate's odds to be correct down to the tenth of a percentage point (or ten-thousandths of a point, as he posted to Twitter last night), he needs to be absolutely right about a long list of assumptions about all sorts of things of which he can't possibly have irrefutable data, knowledge, or foresight.
   951. Random Transaction Generator Posted: November 03, 2012 at 01:27 PM (#4291669)
Not my quote, but I have heard "talking-heads" on TV discussing the fact that for a gas station owner to bring in generators to get open would cost a lot, both in the generator rental, and the gasoline to power the generators.


The important question is: "Would the station owner gouge himself for the gas to power those generators?"

;)
   952. Steve Treder Posted: November 03, 2012 at 01:27 PM (#4291670)
No. Because nobody knows the odds.

There's knowing, and there's knowing. The odds Silver is projecting are based on a transparent and rigorous methodology. He has invested vastly more time and energy and knowledge on this question than the rest of us here combined.

But go ahead, pretend otherwise.
   953. The Yankee Clapper Posted: November 03, 2012 at 01:27 PM (#4291671)
re: #919, I think that is mostly right, and curmudgeonly conservatives may well be more likely to hang up on pollsters. AFAIK, there isn't any research on whether non-responders break evenly along party or ideological lines - at least none has been cited here.

I'm not sure if the National Polls do a better job addressing the non-responders. Perhaps their methodology is better at replacing non-responders with similarly situated voters rather than the next random phone number? Or perhaps there is less resistance to the long-established, well known pollsters like Gallup? Is Rasmussn's supposed tilt to the GOP so commonly known that GOP leaning voters are more likely to respond? Or does its use of automated technology make it more likely to get a response from those who don't want to chat with strangers about their political preferences? Do polls by organizations with a well-known ideology (e.g. Washington Post, New York Times) get more non-responses from those opposed to that ideology? There are a lot of things we don't know about polling that could make a difference in a close election, and "poll bias" includes a lot of things other than pollsters manipulating their results.
   954. cercopithecus aethiops Posted: November 03, 2012 at 01:28 PM (#4291672)
It's hard to imagine how anyone would want to transport gas from faraway places to New Jersey if they thought they couldn't make a profit on it, especially considering that under normal circumstances New Jersey's gas prices are usually about the lowest anywhere in the East.


This is why we have a thing called "government."

That type of voter will tend to "throw the bum out" if the fundamentals are poor, and the incumbent hasn't made a strong case for his re-election.


What is the historical evidence for this? Relative to polling data that is. 1980? Reagan overperformed his polling, but there was a relative dearth of published polling data back then, especially at the state level. Also seems to me that there actually may have been a lot of high quality state level polling data back then that was not available to the news media or the general public, and was possibly quite accurate if anecdotal accounts are to be believed. Pat Caddell supposedly told Carter it was over with more than a week to go.
   955. McCoy Posted: November 03, 2012 at 01:28 PM (#4291673)
interesting.
can you provide a link? (apologies if you have already done so upthread and I missed it)


A closed gas station with no power for pumps would have to either bring in equipment that can generate power (thus guzzling gas) or run lines to areas that do have power or buy equipment that can pump gas via non electric means. All of that costs an incredible amount of money per gallon and gas station owners are already operating under razor thin margins. They simply cannot absorb those costs and make money. They have to increase their prices but they cannot because they'd get hammered by the government.
   956. JL Posted: November 03, 2012 at 01:28 PM (#4291674)
930 - but that does not matter. The only issue is if you don't think that sample of LV is not representative of LV as a whole.
   957. snapper (history's 42nd greatest monster) Posted: November 03, 2012 at 01:29 PM (#4291675)
Again, if we had a patient and educated electorate almost all of the stupidity of the election season would go away.

And now you see why the founders wanted to limit gov't power so much. They knew that a huge chunk of the population is not educated, or patient. Rather, they are self-interested, venal, lazy and tribal.
   958. Monty Predicts a Padres-Mariners WS in 2016 Posted: November 03, 2012 at 01:30 PM (#4291676)
Again, if we had a patient and educated electorate almost all of the stupidity of the election season would go away.


I'd settle for patient and educated candidates.
   959. Random Transaction Generator Posted: November 03, 2012 at 01:30 PM (#4291677)
They knew that a huge chunk of the population is not educated, or patient. Rather, they are self-interested, venal, lazy and tribal.


It's a good thing they get to have guns!
   960. McCoy Posted: November 03, 2012 at 01:31 PM (#4291678)
And now you see why the founders wanted to limit gov't power so much. They knew that a huge chunk of the population is not educated, or patient. Rather, they are self-interested, venal, lazy and tribal.

I don't just see it now. I've been for limited government for a very long time.
   961. snapper (history's 42nd greatest monster) Posted: November 03, 2012 at 01:32 PM (#4291682)
What is the historical evidence for this? Relative to polling data that is. 1980? Reagan overperformed his polling, but there was a relative dearth of published polling data back then, especially at the state level. Also seems to me that there actually may have been a lot of high quality state level polling data back then that was not available to the news media or the general public, and was possibly quite accurate if anecdotal accounts are to be believed. Pat Caddell supposedly told Carter it was over with more than a week to go.

We don't have the data. The issue is, the number of incumbent Presidents who have run for re-election in poor economies is not large enough to make a statistical case. Elections like Carter-Reagan, and Bush-Clinton show the phenomenon.

You're trying to force "sabermetrics" into a situation where you don't have enough data, and should rely on "scouting" instead.

   962. cercopithecus aethiops Posted: November 03, 2012 at 01:34 PM (#4291685)
So if Nate is right, it shows the genius of his model. But if Nate is wrong, it's because he was provided with faulty polling data. I have to give him credit: That's a hell of a racket he's crafted.


I think I sort of half agree with this criticism. There's not really all that much to the model beyond the basic concept of aggregating all the available polling data. If the data is crap, the model can't fix it. So if Nate is right, it mostly means that the polls were right, and not all that much more than that.
   963. McCoy Posted: November 03, 2012 at 01:34 PM (#4291686)
No political party with power (i.e. Republicans/Democrats) wants less power. The only parties advocating for a less powerful goverment are parties with no power. Thus if they get elected they will increase their power even if the government overall would then have less power. Afterall 1 is more than 0 but having gained power they would do little to lessen their power and would over time take steps to increase their power.
   964. snapper (history's 42nd greatest monster) Posted: November 03, 2012 at 01:35 PM (#4291687)
It's a good thing they get to have guns!

Hey, you don't forfeit your natural rights by being a jackass.
   965. tshipman Posted: November 03, 2012 at 01:36 PM (#4291689)
And now you see why the founders wanted to limit gov't power so much. They knew that a huge chunk of the population is not educated, or patient. Rather, they are self-interested, venal, lazy and tribal.


Yes, but we have to let Republicans vote anyways.

But if Nate is wrong, it's because he was provided with faulty polling data. I have to give him credit: That's a hell of a racket he's crafted.


It's the truth, though. If Romney wins, it will be because of systemic and wide-spread inaccuracies with polling data to an extent not seen since the Dewey/Truman election. If Nate was 100% confident of polling data, then his estimate of the election would look more like Sam Wang's.


***

Also, I think it was funny that Szymborski was randomly accusing people of having sock puppets (I do not have multiple accounts, for the record). Does he care to acknowledge the "He's bought a bat like Prince Fielder" sock puppet?
   966. Joe Kehoskie Posted: November 03, 2012 at 01:37 PM (#4291691)
930 - but that does not matter. The only issue is if you don't think that sample of LV is not representative of LV as a whole.

How could it be representative of LV as a whole? If a pollster talks to 1,000 RV and then passes 960 of them through as LV, then the pollster is either projecting 96 percent turnout or substantially over-counting voters for one party or the other.
   967. cercopithecus aethiops Posted: November 03, 2012 at 01:38 PM (#4291693)
You're trying to force "sabermetrics" into a situation where you don't have enough data, and should rely on "scouting" instead.


First, I'm not trying to do anything. Second, the scouts are mostly either overtly partisan or vested in a "too close to call" narrative. So you're basically saying that there's really nothing to rely on.
   968. McCoy Posted: November 03, 2012 at 01:38 PM (#4291694)
We don't have the data. The issue is, the number of incumbent Presidents who have run for re-election in poor economies is not large enough to make a statistical case. Elections like Carter-Reagan, and Bush-Clinton show the phenomenon.

You're trying to force "sabermetrics" into a situation where you don't have enough data, and should rely on "scouting" instead.


Not having the precedent is why it is possible the polls can be wrong but that doesn't mean one cannot come up with a probability based on the polling numbers.

Right now it seems like you're advocating we can't do linear weights because we don't know what happened in the 1874 season despite the fact that we have millions and millions of other data points to use. The applies to this election. We have hundreds of thousands of data points. We can't simply toss those aside because we are in unfamiliar territory or because we think are in unfamiliar territory.
   969. snapper (history's 42nd greatest monster) Posted: November 03, 2012 at 01:38 PM (#4291695)
No political party with power (i.e. Republicans/Democrats) wants less power. The only parties advocating for a less powerful goverment are parties with no power. Thus if they get elected they will increase their power even if the government overall would then have less power. Afterall 1 is more than 0 but having gained power they would do little to lessen their power and would over time take steps to increase their power.

I think you have to separate the "power" of the politicians, from the "power" of the government. Lots of gov't power resides elsewhere than the elected representatives. The Federal agencies, and programs on "auto-pilot" may well have more actual "power" than Congress.

No politician wants less personal political power, but they may well want the gov't as a whole to have less influence over the nation.
   970. McCoy Posted: November 03, 2012 at 01:41 PM (#4291697)
No politician wants less personal political power, but they may well want the gov't as a whole to have less influence over the nation

If it means they get power or get to maintain it. Furthermore by lessening power elsewhere he/they are in fact gaining power for themselves. Think of Hoover or McCarthyism as examples of this. If you can take away power then that means you gain power because of that ability. A party with power will never intentionally decrease their power, ever. It just won't happen.
   971. snapper (history's 42nd greatest monster) Posted: November 03, 2012 at 01:41 PM (#4291699)
It's the truth, though. If Romney wins, it will be because of systemic and wide-spread inaccuracies with polling data to an extent not seen since the Dewey/Truman election.

That's not true. Given an electorate so closely split, deviations in R/D/I turnout of 2-3% from the poll models in one or two states can swing the whole election.

That's not a systematic failure. It's just the inherent margin of error b/c we can't know who's going to actually bother to vote.


First, I'm not trying to do anything. Second, the scouts are mostly either overtly partisan or vested in a "too close to call" narrative. So you're basically saying that there's really nothing to rely on.


When the race is this close? Correct.
   972. spike Posted: November 03, 2012 at 01:41 PM (#4291700)
the generator rental

As if a, there are generators lying around unused big enough to power a gas station to be rented, b, that there is a way to convert a gas station to run on Generator rather than city power fairly quickly, and c, that there is a way to get said generator to the site before normal power is resumed. This is really why you shouldn't listen to talking-heads. Do people think there a bunch Jim's Jiant Jenerator Rental franchises somewhere in Jersey? Christ.
   973. Random Transaction Generator Posted: November 03, 2012 at 01:41 PM (#4291701)
It's a good thing they get to have guns!

Hey, you don't forfeit your natural rights by being a jackass.


Guns are "natural" but voting isn't?
   974. Joe Kehoskie Posted: November 03, 2012 at 01:42 PM (#4291702)
It's the truth, though. If Romney wins, it will be because of systemic and wide-spread inaccuracies with polling data to an extent not seen since the Dewey/Truman election. If Nate was 100% confident of polling data, then his estimate of the election would look more like Sam Wang's.

I've seen MCoA make this claim, too, but it's simply not true. Right now, RCP has Obama winning at the state level 290-248. If Romney wins, that means the polling will have been off in as few as two states (Ohio plus New Hampshire or Ohio plus Colorado). That's hardly a "systemic and widespread" breakdown in polling. Frankly, it's a narrow enough breakdown that a supposedly sophisticated model should be able to overcome it, especially if the model is using all sorts of non-poll data such as economic indicators, historical voting trends, etc., etc.
   975. cercopithecus aethiops Posted: November 03, 2012 at 01:43 PM (#4291703)
Last Ipsos/Reuters passed 82% of RV through the LV screen. Obama led 46-44 among RV and 47-46 among LV. Of course, it's web-based polling, so not worth a damn, right?
   976. tshipman Posted: November 03, 2012 at 01:43 PM (#4291704)
That's not true. Given an electorate so closely split, deviations in R/D/I turnout of 2-3% from the poll models in one or two states can swing the whole election.

That's not a systematic failure. It's just the inherent margin of error b/c we can't know who's going to actually bother to vote.


Romney winning Ohio + IA or WI would be a systematic failure in polling.
   977. snapper (history's 42nd greatest monster) Posted: November 03, 2012 at 01:44 PM (#4291705)
We have hundreds of thousands of data points.

Pardon? We've had <25 Presidential elections since nationwide polling was even attempted.
   978. snapper (history's 42nd greatest monster) Posted: November 03, 2012 at 01:45 PM (#4291707)
Guns are "natural" but voting isn't?

Self-defense is a natural right, democracy isn't. But, I've never advocated limiting the franchise more than it currently is.
   979. McCoy Posted: November 03, 2012 at 01:45 PM (#4291708)
That's not true. Given an electorate so closely split, deviations in R/D/I turnout of 2-3% from the poll models in one or two states can swing the whole election.

Which would make it true not untrue. Many, many, many polls are saying one thing. If the electorate swings 2-3% in several states to swing the election that means that all of those polls missed what was really going on. Thus aggregator sites and Nate's model will be off and it will be off because the polls dropped the ball.

If the polls are wrong or badly off then the models/aggregators will be wrong/badly off. If the polls are right or largely correct then the models will work.

If all of the errors go in one direction then this simply isn't a matter of being within the margin of error. That isn't how it works.
   980. 'Spos lost the handle trying to make the transfer Posted: November 03, 2012 at 01:46 PM (#4291709)
They knew that a huge chunk of the population is not educated, or patient. Rather, they are self-interested, venal, lazy and tribal.

People have often used this as a rationale for more government too.
   981. McCoy Posted: November 03, 2012 at 01:47 PM (#4291710)
As if a, there are generators lying around unused big enough to power a gas station to be rented, b, that there is a way to convert a gas station to run on Generator rather than city power fairly quickly, and c, that there is a way to get said generator to the site before normal power is resumed. This is really why you shouldn't listen to talking-heads. Do people think there a bunch Jim's Jiant Jenerator Rental franchises somewhere in Jersey? Christ.

Yes it possible to do it quickly. All it really requires is money and if you can't charge more then you don't have the money to do it.

Does that mean every single gas station could do it? Of course not. If they were all capable of doing it it would crash the gas market and cause them all to take a huge loss but more gas stations could get back online more quickly if you allowed them to charge market price for gas.
   982. snapper (history's 42nd greatest monster) Posted: November 03, 2012 at 01:47 PM (#4291711)
Romney winning Ohio + IA or WI would be a systematic failure in polling.

No, it would just mean more R's & I's, and fewer D's showed up than they estimated.

These polls all report margins of error of 3-5 pts. Do you just ignore that? A 48-47 poll is a coin flip.

To repeat, there's no way to tell who's actually going to show up on Tues.
   983. Joe Kehoskie Posted: November 03, 2012 at 01:49 PM (#4291712)
Which would make it true not untrue. Many, many, many polls are saying one thing. If the electorate swings 2-3% in several states to swing the election that means that all of those polls missed what was really going on. Thus aggregator sites and Nate's model will be off and it will be off because the polls dropped the ball.

If Nate's model is so sophisticated, it should know that an LV screen that passes 96 percent of respondents through is probably unreliable as a data point.
   984. snapper (history's 42nd greatest monster) Posted: November 03, 2012 at 01:50 PM (#4291714)
People have often used this as a rationale for more government too.

Very true.

It all depends on your relative tolerance for individual failure, vs. collective failure.
   985. McCoy Posted: November 03, 2012 at 01:51 PM (#4291715)
Last Ipsos/Reuters passed 82% of RV through the LV screen. Obama led 46-44 among RV and 47-46 among LV. Of course, it's web-based polling, so not worth a damn, right?

As I said before at this point in the game anyone who is picking up the phone to answer a poll is going to vote or has already voted thus the high % of LV per poll is meaningless. Nor does it mean the polls aren't accurate because only 60% of eligible voters actually do vote. An accurate poll doesn't need to have 60% LV screen to be accurate in fact the screen % is meaningless. Get your 1500 to 2000 people who do say they will vote and you'll get numners you can work with. It doesn't matter if you have to ask 2010 people to get 2000 LV or 5000 people to get that 2000.

Why do you need to have a large amount of people who say they aren't going to vote in your poll to get an accurate poll? It's an absurd statement and only people who want to knock a poll because they don't like what it says would use such a stupid argument.
   986. snapper (history's 42nd greatest monster) Posted: November 03, 2012 at 01:52 PM (#4291716)
As I said before at this point in the game anyone who is picking up the phone to answer a poll is going to vote or has already voted

Where's the evidence for that?
   987. Swedish Chef Posted: November 03, 2012 at 01:52 PM (#4291717)
So, I just looked at Silver's blog. HE'S USING THREE ####### SIGNIFICANT DIGITS!!! The man is a fraud and a charlatan. And using four digits for the elector seats is stupid on several levels.

Sure, I know Obama is going to win, I agree with Silver there. But if you believe that you can put the chance of Obama winning at 83.7%, and believe it means something, you deserve to get beaten up in a dark ally by Nassim N. Taleb.
   988. Joe Kehoskie Posted: November 03, 2012 at 01:54 PM (#4291718)
As I said before at this point in the game anyone who is picking up the phone to answer a poll is going to vote or has already voted thus the high % of LV per poll is meaningless. Nor does it mean the polls aren't accurate because only 60% of eligible voters actually do vote. An accurate poll doesn't need to have 60% LV screen to be accurate in fact the screen % is meaningless. Get your 1500 to 2000 people who do say they will vote and you'll get numners you can work with. It doesn't matter if you have to ask 2010 people to get 2000 LV or 5000 people to get that 2000.

Incorrect. If you talk to 1,000 RV and 96 percent pass through as LV, then you might end up with an accurate poll of those 960 LV, but you're sure as hell not getting a representative sample of the entire electorate — unless, of course, you're projecting 96 percent turnout nationwide.
   989. Jose Can Still Seabiscuit Posted: November 03, 2012 at 01:54 PM (#4291719)
These polls all report margins of error of 3-5 pts. Do you just ignore that? A 48-47 poll is a coin flip.


You don't ignore it but if every state flips 3-5 points in one direction, that seems like a systematic polling error to me. If 6 states show Obama +2 and they all come in at Romney +2, that's a fundamental problem, if the 6 states show Obama +6 in 3 of them and Romney +2 in the other 3, that seems more like a true "Margin of error" issue.
   990. JL Posted: November 03, 2012 at 01:54 PM (#4291720)
966 - so are all polls suspect that don't have a LV screen of the 70% that you deem appropriate? Or is it possible that there will be reasons why the number is so high, such as more LV answering, or just random distribution?

Besides , the goal is not to get a accurate poll of the entire electorate, it is to get an accurate poll of those people who are likely to vote.
   991. phredbird Posted: November 03, 2012 at 01:56 PM (#4291722)
Of course, it is the off-season so there really isn't THAT much else to talk about for the next few days...


ARE YOU NUTS? did you not see that mark mcgwire is going to take the hitting coach position for the dodgers!!!111!!!!??!!

also, jack keefe's post is not getting the respect it deserves.
   992. McCoy Posted: November 03, 2012 at 01:56 PM (#4291723)
Where's the evidence for that?

The LV screens this late in the game. As we get closer and closer to the election people firm up on whether or not they will vote and whether or not they are going to take a poll. If you go back and look the further away from election day you get the smaller the LV % is per poll.

If you look at a lot of the polls that give you data you'll find that occasionally the pollers ask them if they were contacted by the parties recently and a good chunk of them say yes. The people still picking up the phone this late in the game are the people who are likely to vote or have voted while people who at this point in the game who are not going to vote have no incentive to take the poll thus they do not take the poll.
   993. Joe Kehoskie Posted: November 03, 2012 at 01:57 PM (#4291724)
966 - so are all polls suspect that don't have a LV screen of the 70% that you deem appropriate? Or is it possible that there will be reasons why the number is so high, such as more LV answering, or just random distribution?

I touched on this in #988. If 96 percent of your RV sample is passed through as LV, then the only way it's a representative sample is if you're projecting 96 percent turnout nationwide.
   994. McCoy Posted: November 03, 2012 at 01:58 PM (#4291726)
So, I just looked at Silver's blog. HE'S USING THREE ####### SIGNIFICANT DIGITS!!! The man is a fraud and a charlatan. And using four digits for the elector seats is stupid on several levels.

Sure, I know Obama is going to win, I agree with Silver there. But if you believe that you can put the chance of Obama winning at 83.7%, and believe it means something, you deserve to get beaten up in a dark ally by Nassim N. Taleb.


So if Nate said something like OBama has roughly a 75 to 85% chance of winning the election based on the polls he would deserve a pat on the back instead of deserving a beating in a back alley?
   995. Jolly Old St. Nick Still Gags in October Posted: November 03, 2012 at 01:59 PM (#4291727)
Many, many, many polls are saying one thing. If the electorate swings 2-3% in several states to swing the election that means that all of those polls missed what was really going on. Thus aggregator sites and Nate's model will be off and it will be off because the polls dropped the ball.

If the polls are wrong or badly off then the models/aggregators will be wrong/badly off. If the polls are right or largely correct then the models will work.

If all of the errors go in one direction then this simply isn't a matter of being within the margin of error. That isn't how it works.


Well, it's "how it works" if you think that there's some sort of a fiendish conspiracy among pollsters to tilt their polls in the same direction, even though this would result in them all looking like a bunch of jackasses after Tuesday.
   996. snapper (history's 42nd greatest monster) Posted: November 03, 2012 at 02:00 PM (#4291730)
You don't ignore it but if every state flips 3-5 points in one direction, that seems like a systematic polling error to me. If 6 states show Obama +2 and they all come in at Romney +2, that's a fundamental problem, if the 6 states show Obama +6 in 3 of them and Romney +2 in the other 3, that seems more like a true "Margin of error" issue.

Not if all or most of the polls are using the same projected R/D/I turnout.

If 5 of 7 polls show Obama winning OH by 2-3 pts., and those 5 are all using turnout assumptions based on 2008, and the actual turnout is 2-3 pts. more Republican than projection, then they'll all be wrong.

Yet, none of them "failed", they just made the wrong assumption about turnout. At the end of the day, the turnout breakdown is nothing but an educated guess. We know people lie about their intentions to vote, and if they voted last-time.
   997. McCoy Posted: November 03, 2012 at 02:00 PM (#4291731)
Besides , the goal is not to get a accurate poll of the entire electorate, it is to get an accurate poll of those people who are likely to vote.

Bingo. As I said before requiring that 20 to 30% of your respondents saying they aren't voting for you to have an accurate poll is beyond stupid and is only being argued for because some people don't like what the polls are saying. What matters is what the LV will do and it doesn't matter how many non voting pollees you go through to get your n sample.
   998. BDC Posted: November 03, 2012 at 02:01 PM (#4291733)
To clarify #928, I don't think Silver should do anything differently, nor do I care about percentage points in his odds. My point is basically that most people don't understand math: I'm one of them, so I speak with some authority :) Obama has, as of this morning, a very strong chance to win Tuesday's election, given that he only has to win it for that to happen, not win it by very much. I guarantee you some people will wake up Wednesday, if Obama wins with 50.1% of the vote to 49.9%, and say "Hah! That fanboy Nate Silver said it would be 83 to 17!"

   999. McCoy Posted: November 03, 2012 at 02:02 PM (#4291734)
Not if all or most of the polls are using the same projected R/D/I turnout.

If 5 of 7 polls show Obama winning OH by 2-3 pts., and those 5 are all using turnout assumptions based on 2008, and the actual turnout is 2-3 pts. more Republican than projection, then they'll all be wrong.


Missing the turnout ratio is a systemic error and not an error on the part of aggregators.
   1000. Joe Kehoskie Posted: November 03, 2012 at 02:03 PM (#4291736)
[moved to next page]
Page 10 of 114 pages ‹ First  < 5 6 7 8 9 10 11 12 13 14 15 >  Last ›

You must be Registered and Logged In to post comments.

 

 

<< Back to main

BBTF Partner

Support BBTF

donate

Thanks to
James Kannengieser
for his generous support.

Bookmarks

You must be logged in to view your Bookmarks.

Hot Topics

NewsblogJosh Lueke Is A Rapist, You Say? Keep Saying It.
(217 - 8:05am, Apr 24)
Last: Lassus

NewsblogMatt Williams: No problem with Harper's two-strike bunting
(1 - 7:57am, Apr 24)
Last: Rough Carrigan

NewsblogOMNICHATTER for 4-24-2014
(2 - 7:56am, Apr 24)
Last: Rickey! In a van on 95 south...

NewsblogFull Count » Red Sox to call up right-hander Alex Wilson, option Daniel Nava
(1 - 7:52am, Apr 24)
Last: villageidiom

NewsblogMichael Pineda ejected from Red Sox game after pine tar discovered on neck
(34 - 7:44am, Apr 24)
Last: villageidiom

NewsblogKeri: Slump City: Why Does the 2014 MLB Season Suddenly Feel Like 1968?
(18 - 7:32am, Apr 24)
Last: Jorge Luis Bourjos (Walewander)

NewsblogOTP April 2014: BurstNET Sued for Not Making Equipment Lease Payments
(2399 - 7:19am, Apr 24)
Last: formerly dp

NewsblogOMNICHATTER for 4/23/2014
(180 - 7:18am, Apr 24)
Last: Bunny Vincennes

NewsblogDaniel Bryan's 'YES!' chant has spread to the Pirates' dugout
(188 - 4:32am, Apr 24)
Last: SouthSideRyan

NewsblogDoyel: How was Gerrit Cole not suspended? He basically started the brawl
(28 - 2:44am, Apr 24)
Last: the Hugh Jorgan returns

NewsblogJonah Keri Extended Interview | Video | Late Night with Seth Meyers | NBC
(8 - 1:54am, Apr 24)
Last: Jim (jimmuscomp)

NewsblogOT: The NHL is finally back thread, part 2
(226 - 1:39am, Apr 24)
Last: Mess with the Meat, you get the Wad!

NewsblogColiseum Authority accuses Athletics of not paying rent
(12 - 1:12am, Apr 24)
Last: greenback likes millwall, they don't care

NewsblogMorosi: Diamondbacks' growing gloom might mean doom for manager or GM
(12 - 1:11am, Apr 24)
Last: KT's Pot Arb

NewsblogPrimer Dugout (and link of the day) 4-23-2014
(14 - 12:44am, Apr 24)
Last: vortex of dissipation

Demarini, Easton and TPX Baseball Bats

 

 

 

 

Page rendered in 1.0582 seconds
52 querie(s) executed