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 23 of 114 pages ‹ First  < 18 19 20 21 22 23 24 25 26 27 28 >  Last ›
   2201. Jolly Old St. Nick Is A Jolly Old St. Crip Posted: November 05, 2012 at 05:40 PM (#4293659)
Not to mention that if the actual goal is to reduce abortion's numbers, the surest way to do that would be to provide free contraceptives to the girls and women who are most likely to get pregnant unintentionally. Obviously that may send a different message than we'd ideally like, all other things being equal, but in terms of reducing abortions, it'd sure be more humane and more effective than the alternative of criminalization.

There's no evidence that knowledge of, or availability of, contraception has any effect on "unwanted pregnancy." (That is, women who get pregnant unintentionally do not list inability to obtain contraception as a reason.)


Really? From the Chicago Tribune, October 12, 2012....

Free birth control tied to drop in abortions

NEW YORK (Reuters Health) - During a four-year experiment that offered the women of St. Louis, Missouri, free contraceptives - including expensive long-term implants - rates of teen pregnancy and repeat abortions in the area dropped dramatically, according to a new study.

Researchers say the project, begun in 2007, was intended to model the likely benefits of mandatory insurance coverage for contraception under the Affordable Care Act.

"We believe that this is a major step to reducing unintended pregnancy in the United States," said lead author Dr. Jeffrey Peipert, a professor at Washington University School of Medicine in St. Louis....
   2202. Rickey! trades in sheep and threats Posted: November 05, 2012 at 05:41 PM (#4293660)
Nate DOES have a lot riding on the outcome of this election. What I don't understand is why nobody else seems to. Gallup, RCP, PPP, Rasmussen, all those talking heads... you know those guys are all still going to be on top of the mountain no matter how wrong they get things. Only Nate seems to be in any danger.


I'm as big of a Nate proponent as you're likely to find, but let's not pretend that he hasn't put himself out there as an individual, and made much stronger claims to objective truth, than any of the individual pollsters you mention above.
   2203. Jolly Old St. Nick Is A Jolly Old St. Crip Posted: November 05, 2012 at 05:41 PM (#4293662)
As for your nutrition label, Andy - I take issue with a few things, principally with how they define welfare and I imagine it'd be impossible to come up with a definition that makes everybody happy. Well designed, though.

Yeah, the nomenclature for "welfare" and "foodness" are kind of jargonistic, but they're getting at the right thing.
   2204. Rickey! trades in sheep and threats Posted: November 05, 2012 at 05:42 PM (#4293664)
NEW YORK (Reuters Health) - During a four-year experiment that offered the women of St. Louis, Missouri, free contraceptives - including expensive long-term implants - rates of teen pregnancy and repeat abortions in the area dropped dramatically, according to a new study.


David is not interested in reducing abortions. He is interested in slut-shaming women for daring to have sex without his moral permission.
   2205. zonk Posted: November 05, 2012 at 05:43 PM (#4293665)
This creates the problems you specify without solving the problem you want it to. If you increase the House to 2,000 from 435, then you decrease the number of people represented by a Representative from about 700,000 to 150,000. I don't think that a guy representing 150K people has any "fighting chance to actually know" his constituents.

To actually create such a "fighting chance," you'd probably need to make it, say, 10,000. Which is... silly.


Well, the original constitutional requirement set the lower limit at 30K per district -- and the size of the House regularly increased up until the 1911 apportionment act... 10K may be silly, but if not for the apportionment act, normal growth would definitely have us up into the thousands by now.

Plenty of state legislatures make it work -- NH, I think, has a rather large lower chamber with pretty small districts.

I personally wouldn't mind seeing a larger House... but I don't think it really matters. Hard for me to see how anything changes so far as legislation goes -- the Senate will still be the cooling saucer/where legislation goes to die regardless of how large/responsive/representative of the 'will of the people' the House becomes.
   2206. Jolly Old St. Nick Is A Jolly Old St. Crip Posted: November 05, 2012 at 05:43 PM (#4293666)
Nate DOES have a lot riding on the outcome of this election. What I don't understand is why nobody else seems to. Gallup, RCP, PPP, Rasmussen, all those talking heads... you know those guys are all still going to be on top of the mountain no matter how wrong they get things. Only Nate seems to be in any danger.

I've been one among several people here who've made that same point, but when it comes to tunnel vision, Ray's got a degree of glaucoma that no eye drops can cure.
   2207. Bring Me the Head of Alfredo Griffin (Vlad) Posted: November 05, 2012 at 05:43 PM (#4293667)
What will be real interesting is that there's increasing buzz that Pelosi may step aside...


That'd be a shame. She's a good wartime consigliere.
   2208. Morty Causa Posted: November 05, 2012 at 05:43 PM (#4293668)
   2209. Jolly Old St. Nick Is A Jolly Old St. Crip Posted: November 05, 2012 at 05:45 PM (#4293669)
David is not interested in reducing abortions. He is interested in slut-shaming women for daring to have sex without his moral permission.

I don't even think it's even that complicated. Anytime he sees a perceived "liberal" post something, he simply reaches for his teleprompter and fires away.
   2210. David Nieporent (now, with children) Posted: November 05, 2012 at 05:45 PM (#4293670)
Nate DOES have a lot riding on the outcome of this election. What I don't understand is why nobody else seems to. Gallup, RCP, PPP, Rasmussen, all those talking heads... you know those guys are all still going to be on top of the mountain no matter how wrong they get things. Only Nate seems to be in any danger.
Nate is not a pollster. Those companies provide a service which produces data. Nate provides analysis of that data. If his analysis is wrong, he's not providing anything at all.
   2211. Bring Me the Head of Alfredo Griffin (Vlad) Posted: November 05, 2012 at 05:45 PM (#4293671)
What's a train? You are so last millennium?


I believe it's some type of ceremonial structure constructed from Reardon Steel.
   2212. tshipman Posted: November 05, 2012 at 05:46 PM (#4293672)
Re: generic congressional ballot:

There's pretty significant evidence that the ballot is a very good indicator of house races, once you adjust for incumbency. It performs better than a combo of individual polls, simply due to the scarcity of polling. Generic ballot is something like D+3 or +2 (I can't remember), and once you adjust for incumbency, it looks like a gain of 1-10 seats at best or a loss of 5 at worst.


***

I agree that Nate has more at stake than pundit hacks because his whole brand is built on being more accurate than hacks. If he's not right, what good is he?


   2213. Bring Me the Head of Alfredo Griffin (Vlad) Posted: November 05, 2012 at 05:47 PM (#4293673)
Nate is not a pollster. Those companies provide a service which produces data. Nate provides analysis of that data. If his analysis is wrong, he's not providing anything at all.


As far as generated value is concerned, is there a significant gap between a provider of inaccurate data and a provider of nothing at all?
   2214. Jolly Old St. Nick Is A Jolly Old St. Crip Posted: November 05, 2012 at 05:48 PM (#4293674)
Good link to The Onion, Morty, and it's just as good as it was several hundred posts back. (Time kind of flies around here.) And the sub-link about the Times is even better.
   2215. David Nieporent (now, with children) Posted: November 05, 2012 at 05:49 PM (#4293675)
There's no evidence that knowledge of, or availability of, contraception has any effect on "unwanted pregnancy." (That is, women who get pregnant unintentionally do not list inability to obtain contraception as a reason.)
I slightly misreported this study, from memory. Knowledge may have helped, but availability did not. It was a CDC study, and found that only 13% of teens who got pregnant listed trouble getting contraception as the reason -- even though that would have been the most socially acceptable answer.
   2216. Jack Carter, calling Beleaguered Castle Posted: November 05, 2012 at 05:50 PM (#4293676)
Virginia is the only state that figures to be interesting for President - an Obama win puts Romney in really bad shape and makes Ohio an absolute necessity. Neither VT or SC figure to have much of interest to tell us.
You bet. If Obama takes Virginia (and assuming NV and WI), Romney has to run the table.

And if it's not clear, this is what I think will happen when they frack into northern New York's exemplary aquifer.
The Nation article I linked to talks about the pressure on Cuomo to halt fracking in NY. How it plays out will be fascinating. The most relevant paragraph is,

In June, the New York Times reported that Cuomo was prepared to allow fracking in a few counties along the Pennsylvania border where the gas is richest and the resistance weakest. His administration denies that it has reached any decision.

Nationally, the anti-fracking movement’s gains have been incremental—a major environmental study and a series of more restrictive rules on the federal level and regulatory overhauls in several states (those that have banned the practice do not have extensive gas reserves). A decision by Cuomo to ban fracking would be momentous. A decision to maintain a moratorium pending yet further review, less so—although an increasingly restive anti-fracking movement is unlikely to let any review process pass quietly. For those living above the shale, the stakes could not be higher. “It’s one of those decisions,” McKibben said, “that will be recorded in geologic time.”
   2217. Bring Me the Head of Alfredo Griffin (Vlad) Posted: November 05, 2012 at 05:51 PM (#4293677)
It was a CDC study, and found that only 13% of teens who got pregnant listed trouble getting contraception as the reason -- even though that would have been the most socially acceptable answer.


Assuming that the results of that study are representative, in what way is preventing 13% of the country's unwanted teenage pregnancies (and the numerous resulting abortions) an unworthy goal?
   2218. RoyalsRetro (AG#1F) Posted: November 05, 2012 at 05:52 PM (#4293678)
Wait--your individual candidates are going to have to make their positions known, right? We'll call you the Good Government Party for the moment. Once the GGP candidate in Missouri mentions he's pro-choice, he's running for what are now Democratic votes. How does he get elected?

I like your idea, but I just don't see how your candidate is fighting for anything other than half the votes. I don't see paths to victory.


I wouldn't assume that all the Republican votes are pro-life (my GOP state Rep in red Kansas is pro-choice), or that abortion is a very important topic for even a majority of voters. Actually she's a great example of how you win. She came to my door the other day and tried to convince me how liberal she was (its a moderate district, I think Obama carried it in '08, but Bush in '04). I'm completely convinced that if we were in Massachusetts rather than Kansas, she'd run as a Democrat. They had a forum between her and her Democratic opponent and I can't tell you a single thing they disagreed about. With how gerry-mandered districts are now (another good reform platform plank), you're going to get candidates from both parties that are quite similar on the issues. But because she's a Republican, I associate her with the more radical aspects of the GOP, and I won't vote for her. The same is true for Democrats running in red districts, even if they're quite conservative.

Now imagine a candidate that substantively takes many of the same positions, but presents a bold new procedural reform to "fix" the system. Wouldn't that be a lot more appealing to some voters, particularly independent-minded voters?
   2219. zonk Posted: November 05, 2012 at 05:54 PM (#4293679)
I'd actually have a tough time responding to such a poll question myself... I think federal income tax rates are too low, but I think an awful lot of 'fees' that nickel and dime us in the masses are out of hand.



Now, that's a smart observation. And what's the cost for administering and enforcing all that picayune stuff?


Well - I don't expect this to be a particularly popular opinion - but frankly, there are certain state level aspects that I think could use some federalization/standardization... Economies of scale and such when it comes to things from license plates to licenses.

Of course, there's an absolute ton of constitutional muster that most of my plans wouldn't pass, so it's probably moot for the most part, even if it were the sort of thing that had public support (and I have no illusions about that, either -- my position is almost certainly a very much minority one).
   2220. Jolly Old St. Nick Is A Jolly Old St. Crip Posted: November 05, 2012 at 05:54 PM (#4293680)
Nate DOES have a lot riding on the outcome of this election. What I don't understand is why nobody else seems to. Gallup, RCP, PPP, Rasmussen, all those talking heads... you know those guys are all still going to be on top of the mountain no matter how wrong they get things. Only Nate seems to be in any danger.


Nate is not a pollster. Those companies provide a service which produces data. Nate provides analysis of that data. If his analysis is wrong, he's not providing anything at all.

If Nate's analysis is "wrong", and if he's basing his analysis in great part on those polls, while it's true that he'd have egg on his face for relying on their faulty data, I'm not sure just how the pollsters can at the same time walk away unscathed themselves. Seems to me that if Romney wins, all of those who projected an Obama win would be reaching for the soap and water, whether you call them "data providers" or "data analysts".
   2221. Jack Carter, calling Beleaguered Castle Posted: November 05, 2012 at 05:54 PM (#4293681)
@2218: It would indeed appeal to some voters, including myself. I've been digging around for stats on single issue voters. I'll report in if I can find them. My suspicion that, particularly in the south, the number re abortion is very high.
   2222. Slivers of Maranville descends into chaos (SdeB) Posted: November 05, 2012 at 05:56 PM (#4293682)

To actually create such a "fighting chance," you'd probably need to make it, say, 10,000. Which is... silly.


Not silly at all! It's the heart of my ideal reform. You divide every state into districts of 10,000 people, each elects a representative every two (or four) years, and then you select 435 (or however many), from that number ... by lot. Poof, you have your House of Representatives.

The rest are assigned to their state legislature and possibly other elective offices (county boards, parole boards, etc) by lot as well.
   2223. Jolly Old St. Nick Is A Jolly Old St. Crip Posted: November 05, 2012 at 05:57 PM (#4293683)
It was a CDC study, and found that only 13% of teens who got pregnant listed trouble getting contraception as the reason -- even though that would have been the most socially acceptable answer.


Assuming that the results of that study are representative, in what way is preventing 13% of the country's unwanted teenage pregnancies (and the numerous resulting abortions) an unworthy goal?

Because it might involve raising David's taxes, what else?
   2224. DA Baracus Posted: November 05, 2012 at 05:57 PM (#4293684)
Nate is not a pollster. Those companies provide a service which produces data. Nate provides analysis of that data. If his analysis is wrong, he's not providing anything at all.


Neither is Real Clear Politics. Yet their reputation is somehow not at stake.
   2225. Rickey! trades in sheep and threats Posted: November 05, 2012 at 05:59 PM (#4293686)
I don't even think it's even that complicated. Anytime he sees a perceived "liberal" post something, he simply reaches for his teleprompter and fires away.


I actually think this undersells David a little. He really is a reactionary conservative nutter on social issues. He just doesn't like to admit as much in public.
   2226. spike Posted: November 05, 2012 at 06:01 PM (#4293687)
You bet. If Obama takes Virginia (and assuming NV and WI) Romney has to run the table

That's a tough table - a VA win would be the end of it. Obama is going to win one of CO, NH, FL, IA, or OH.
   2227. Ray (RDP) Posted: November 05, 2012 at 06:01 PM (#4293688)
If Nate's analysis is "wrong", and if he's basing his analysis in great part on those polls, while it's true that he'd have egg on his face for relying on their faulty data, I'm not sure just how the pollsters can at the same time walk away unscathed themselves.


Because the pollsters are just saying "Look, here is our data." Nate is going a step further, to say: "Based on their data, I can give you an accurate prediction." But if his prediction is not accurate, then his prediction is worth no more than anyone else's - including mine.

He is telling us that despite whatever flaws there are in the polling, he can nevertheless rely on the data to a large extent - and can model the data to come up with an accurate prediction.
   2228. Jolly Old St. Nick Is A Jolly Old St. Crip Posted: November 05, 2012 at 06:03 PM (#4293690)
From that Chicago Tribune link in #2101:

"Many doctors and nurse practitioners are reluctant to use IUDs in women who are at risk for infection. They also fear it can cause infertility. Many of these are myths we need to get past," Peipert told Reuters Health.

To address such barriers to access, and to promote long-acting contraceptive methods, Peipert and his colleagues set up the Contraceptive CHOICE Project, with funding from the Susan Thompson Buffett Foundation, in St. Louis.

Between 2007 and 2011, a total of 9,256 women enrolled and were offered any FDA-approved contraceptive they wanted after counseling about the various choices, their effectiveness, risks and benefits.

The researchers also educated doctors and nurses in the program about long-acting birth control, and urged them to encourage women to opt for these methods.

Among the women who participated in the project, 63 percent had previously had an unintended pregnancy and 41 percent had had an abortion.

Most women - 75 percent - chose a long-acting method of contraception.

Overall, abortions, repeat abortions and teen pregnancies all dropped in St. Louis while rates elsewhere in Missouri remained stable, Peipert's team found.

From 2008 to 2010, for instance, the number of abortions in the St. Louis metropolitan area declined by 20 percent, while the rate of abortions in the rest of Missouri remained unchanged.

Similarly, in St. Louis the proportion of abortions requested by women who had had one previously fell from 47 percent in 2006 to 39 percent in 2010.

In comparison, repeat abortions in Kansas City, Missouri, rose from 46 percent to 51 percent during the same time period.


Peipert's group also found that rates of teen pregnancy were substantially lower among girls participating in the project. Among teen girls aged 15 to 19 enrolled in Contraceptive CHOICE, the number of pregnancies translated to a rate of 6.3 per 1,000. That compares to the national average for 15-19 year-olds of 34 out of every 1,000 girls.

The study cannot prove that improved access to birth control caused the drop in abortions and teen pregnancies.

But Peipert said he hopes that with the insurance coverage changes for birth control under the Affordable Care Act, long-acting contraceptives will become more accessible to women.

The healthcare law requires insurance companies to offer birth control without co-pays. Some women may have already seen this change to their benefits, but others will have to wait until their new insurance plan year.

"It is possible for us, with our current medical knowledge and evidence and clinical services, to address the unintended pregnancy rate in the U.S.," said Cynthia Harper, a professor at the University of California, San Francisco, who was not involved in the study.

Harper said that the program's focus not only on cost barriers but information barriers among health care providers and women made it successful.

In particular, women were counseled on the effectiveness of each birth control method, an approach Harper said isn't always taken in the doctor's office.

"Patient education is important, and women don't have a very good sense of the effectiveness of contraception," she told Reuters Health.

"The unintended pregnancy rate has been high for decades, and this shows we can make progress and it's not beyond our grasp," said Harper.
   2229. tshipman Posted: November 05, 2012 at 06:06 PM (#4293692)
Neither is Real Clear Politics. Yet their reputation is somehow not at stake.


RCP is not claiming to know anything, but Nate is, and that's the real difference. Anyone can look at the polls and say, "Well, Obama's the favorite." Nate is looking at the polls, and saying that Obama's 85% to win.

It's the specificity and the newness that makes Nate out on a ledge.

Edit: Coke to Ray.
   2230. Jolly Old St. Nick Is A Jolly Old St. Crip Posted: November 05, 2012 at 06:06 PM (#4293693)
If Nate's analysis is "wrong", and if he's basing his analysis in great part on those polls, while it's true that he'd have egg on his face for relying on their faulty data, I'm not sure just how the pollsters can at the same time walk away unscathed themselves.

Because the pollsters are just saying "Look, here is our data."


Only a ####### lawyer could come up with a spin like that. As if the pollsters aren't telling us that the candidate with the biggest percentage is likely to win the election. If they weren't implying that, and if people weren't inferring that, the market for their service would evaporate.
   2231. zonk Posted: November 05, 2012 at 06:06 PM (#4293694)
Nate DOES have a lot riding on the outcome of this election. What I don't understand is why nobody else seems to. Gallup, RCP, PPP, Rasmussen, all those talking heads... you know those guys are all still going to be on top of the mountain no matter how wrong they get things. Only Nate seems to be in any danger.


Nate is not a pollster. Those companies provide a service which produces data. Nate provides analysis of that data. If his analysis is wrong, he's not providing anything at all.


Pollsters do tend to come and go, though... beyond a few of the long-time, big players -- Pew and Gallup come to mind as pollsters that never seem to lose business no matter how badly they do in a given cycle -- others, not so much. Nobody pays attention to Zogby anymore, for example (as much I'd dearly like to -- he had some really, really good state numbers for Obama today) because he's had a couple out on a limb cycles that were comically bad. R2K - the pollster where Nate and others found some crosstabs that just made no sense and looked fabricated and had a very messy breakup with Dailykos - is another that's gone...

I agree that Nate has a lot riding on the results... if he misses, his mainstreet cred is done -- no one's going to cite him any more.

It's the punditocracy that I wish would take a big hit.... maybe not so much the acknowledged partisans - I mean, they're gonna spin no matter what - but more the supposed "one the level" folks, and especially those "on the level" folks who claim to be making clear-eyed predictions (I'm looking at you Barone, Todd, Gerson, et al). Some of them - Chuck Todd, for one, maybe - I do occasionally find insightful... but they all seem to get distracted by shiny objects. I suppose it's just a natural part of the turf, but I get tired of day after day of pretending that they've found some new groundshaking piece of insight.

   2232. formerly dp Posted: November 05, 2012 at 06:07 PM (#4293695)
He is telling us that despite whatever flaws there are in the polling, he can nevertheless rely on the data to a large extent - and can model the data to come up with an accurate prediction.


He's talking about the likeliness of certain outcomes. He's not saying definitely what will or won't happen. I'm not amazed that the mainstream discussion on Silver continues to not appreciate the difference, but I am amazed that some people here are doing the same.
   2233. Harveys Wallbangers Posted: November 05, 2012 at 06:10 PM (#4293697)
zonk

state licensing boards are bogus. the nation would be a lot better if we reduced the licensing burden on individual businesses.

state licensing is a means of creating a barrier to entry. it's counterproductive at this point.

morty

i hear the concern about legalization but again, we have tried all the other stuff and it's just costly and we have lots of folks in prison learning to be better criminals. i think we have to explore the alternative some how some way.



   2234. Jolly Old St. Nick Is A Jolly Old St. Crip Posted: November 05, 2012 at 06:10 PM (#4293698)
BTW Obama's now gone from -1 to +3 in the last two days of the WaPo tracking poll, pushing his RCP advantage to 0.7. I suppose that this isn't supposed to imply anything, either, but if Nate should infer anything from it, then he's the dummy.
   2235. zonk Posted: November 05, 2012 at 06:11 PM (#4293699)
What will be real interesting is that there's increasing buzz that Pelosi may step aside...



That'd be a shame. She's a good wartime consigliere.


Oh, I agree --

She ran a tight ship as speaker - if you're a partisan and expect your majority leader to keep the caucus in line and deliver the votes when needed, she was at least as good as anyone since Gingrich, and I'd say on the D side - she was better than anyone since Tip O'Neil.

However, she's also 72 and given that there's virtually no chance of the gavel going back to the D's this cycle - I have a tough time seeing how it makes sense for Team D not to start grooming the next leadership for... whenever.
   2236. Fernigal McGunnigle has become a merry hat Posted: November 05, 2012 at 06:11 PM (#4293700)
Not silly at all! It's the heart of my ideal reform. You divide every state into districts of 10,000 people, each elects a representative every two (or four) years, and then you select 435 (or however many), from that number ... by lot. Poof, you have your House of Representatives.

The rest are assigned to their state legislature and possibly other elective offices (county boards, parole boards, etc) by lot as well.


So if the guy I vote for gets sent to Congress then I don't have a state legislator? The other thing is that with a district every 10K people, California is going to have close to 3,800 districts and the country as a whole will have about 31,000. That's just way too many. EDIT: There are about 3200 counties in the US. Counties vary wildly in population and size, but still, 10 people elected per county is a bit much.

The lots thing is fun though, and very Italian. This is Wikipedia's simplified version of how they chose a Doge in Venice, from the 13th century until 1797:

Thirty members of the Great Council, chosen by lot, were reduced by lot to nine; the nine chose forty and the forty were reduced by lot to twelve, who chose twenty-five. The twenty-five were reduced by lot to nine and the nine elected forty-five. Then the forty-five were once more reduced by lot to eleven, and the eleven finally chose the forty-one who actually elected the doge. None could be elected but by at least twenty-five votes out of forty-one, nine votes out of eleven or twelve, or seven votes out of nine electors.
   2237. David Nieporent (now, with children) Posted: November 05, 2012 at 06:11 PM (#4293701)
As far as generated value is concerned, is there a significant gap between a provider of inaccurate data and a provider of nothing at all?
That depends what you mean by "inaccurate data." If the data is simply throwing darts at a dartboard such that it's complete garbage, then no, there's no gap. If the data is inaccurate because it makes certain assumptions and we can strip apart those assumptions and analyze it, then yes, there's a significant gap.
   2238. David Nieporent (now, with children) Posted: November 05, 2012 at 06:13 PM (#4293702)
To actually create such a "fighting chance," you'd probably need to make it, say, 10,000. Which is... silly.

Not silly at all! It's the heart of my ideal reform. You divide every state into districts of 10,000 people, each elects a representative every two (or four) years, and then you select 435 (or however many), from that number ... by lot. Poof, you have your House of Representatives.

The rest are assigned to their state legislature and possibly other elective offices (county boards, parole boards, etc) by lot as well.
You misinterpreted me. I wasn't saying that you'd need to make the districts cover only 10,000 people; I was saying that the House would need to have 10,000 people in it.
   2239. Jolly Old St. Nick Is A Jolly Old St. Crip Posted: November 05, 2012 at 06:14 PM (#4293703)
[Nate's] talking about the likeliness of certain outcomes. He's not saying definitely what will or won't happen. I'm not amazed that the mainstream discussion on Silver continues to not appreciate the difference, but I am amazed that some people here are doing the same.

After seeing some of these folks at work over the past few years, nothing about their reading comprehension (or lack of it) surprises me. You'd think he was predicting that Obama was going to get 85% of the vote, rather than the 50.6% that he's actually projecting.
   2240. formerly dp Posted: November 05, 2012 at 06:14 PM (#4293704)
It's the punditocracy that I wish would take a big hit....


This is, IMO, a lot of what's riding on Nate's accuracy. I understand that the right-leaners on this board don't like the predictions Nate's turning out, but I'm not sure what part of his systemic critique of the MSM pundit class they could possibly disagree with. If Nate's right, which I get is bad news for their side, hopefully his critique will gain more traction. Probably not, but we can hope. It'll certainly get more airtime. The SABR stuff steered mainstream conversations about baseball in a better direction; Nate is trying to do something similar here by calling the popular press out for being nothing more than a distraction engine. The downside, of course, is that it feeds this abominable trend of horserace coverage, but that wasn't going away anytime soon, with or without Nate's intervention.
   2241. tshipman Posted: November 05, 2012 at 06:14 PM (#4293705)
state licensing boards are bogus. the nation would be a lot better if we reduced the licensing burden on individual businesses.


This is actually a Democrat position. Why do we require licences for beauticians again?

***

Btw, coming back to this from a couple of pages ago:
If #1947 is Obama's second-term agenda, he's done a fine job of hiding it. For the past six months, Obama's second-term agenda has been, "Mitt Romney is really, really scary! And also, Mitt Romney is really, really scary!"


It's pretty clear what Obama's second term agenda is.

1. Draw down Afghanistan
2. Some form of Grand Bargain, with increased revenue from high net worth individuals and cuts to social programs with the intent of reducing the deficit.
3. Implement ACA.
4. Some kind of immigration reform.
5. Some kind of climate legislation.

This has been pretty clear for some time. Obama is emphasizing Romney's policies because they are seriously out of step with most American's preferred policies.
   2242. Rickey! trades in sheep and threats Posted: November 05, 2012 at 06:15 PM (#4293706)
He's talking about the likeliness of certain outcomes. He's not saying definitely what will or won't happen. I'm not amazed that the mainstream discussion on Silver continues to not appreciate the difference, but I am amazed that some people here are doing the same.


There is a distinction between discussions of what is and discussions of what ought to be. Most everyone outside of the Kehoskieite fringes recognize what Nate is doing and the distinction between his forecasting model and outright claims of certainty. This is an OT thread on BTF, after all. Primates understand probability and predictive modeling as well as any group of internet nerds you're likely to find.

But this *is* the OT thread about *politics,* and politically Nate Silver can't go wide with his model (via the NYT, his increasingly confident-bordering-on-arrogant Twitter posts, national cable television appearances) and then complain that the masses he's courting (in order to push his personal brand and move into the stratosphere of the pundit class himself) don't really understand the nuance of his model. If you want an audience who will stick with statistical nuance, publish at 538.com. If you want fame, publish at NYT.com and go on cable talk shows. But don't try to ask for both. That's not how politics work.
   2243. zonk Posted: November 05, 2012 at 06:16 PM (#4293707)
It's the punditocracy that I wish would take a big hit....



This is, IMO, a lot of what's riding on Nate's accuracy. I understand that the right-leaners on this board don't like the predictions Nate's turning out, but I'm not sure what part of his systemic critique of the MSM pundit class they could possibly disagree with. If Nate's right, which I get is bad news for their side, hopefully his critique will gain more traction. Probably not, but we can hope. It'll certainly get more airtime. The SABR stuff steered mainstream conversations about baseball in a better direction; Nate is trying to do something similar here by calling the popular press out for being nothing more than a distraction engine. The downside, of course, is that it feeds this abominable trend of horserace coverage, but that wasn't going away anytime soon, with or without Nate's intervention.


Of course, then we run into --- what exactly are the cable nets going to do for 10-12 months every couple years? They might actually start, you know, reporting on policy, issues, and such... the horror... the horror...
   2244. tshipman Posted: November 05, 2012 at 06:16 PM (#4293708)
The other thing is that with a district every 10K people, California is going to have close to 3,800 districts and the country as a whole will have about 31,000. That's just way too many.


Why is it way too many? There are that many people who ####### live in California. I'm sick and tired of ######## in New Hampshire and Iowa having more say in our politics than the largest state in the union.
   2245. RoyalsRetro (AG#1F) Posted: November 05, 2012 at 06:17 PM (#4293710)
Have we ever had an election where both Presidential candidates failed to win their home state? Mitt Romney is down some 20 points in Massachusetts, and of course, Kenya is not allowed to vote for our President. *ducks* (I'm an Obama supporter!)
   2246. Morty Causa Posted: November 05, 2012 at 06:18 PM (#4293711)
It doesn't strike me as quite right that someone who gives a probability should then be held to the standard of having to be 100% correct or else. No matter how much Silver's method right now says Obama is heavily favored, Romney still has a chance. Judging whether his system failed or succeeded would require a little more finesse when it comes to parsing. If he relies on data that turns out not to have predicted the reality that came about, how's that his fault?

EDIT: pretty much in line with formerly dp above.
   2247. spike Posted: November 05, 2012 at 06:19 PM (#4293712)
If Nate's right, which I get is bad news for their side,

In fairness to said right leaners, if Nate's right, it's bad news for the pundit class generally, regardless of party. The Chris Cillizas of the world have been pushing the dead heat meme that is just as contradicted by Silver's analysis as Michael Barone's Romney landslide prediction.
   2248. David Nieporent (now, with children) Posted: November 05, 2012 at 06:19 PM (#4293713)
Neither is Real Clear Politics. Yet their reputation is somehow not at stake.
I don't follow these things as closely as many here, but I thought RCP just did a straight average of the polls that came out. They don't attempt to analyze the data; they just say, "Here's all the polls." It's the difference between BPro and BBref. You can look at BBref and make your own prediction for how Albert Pujols will hit next year; BPro purports to tell you. If BPro's projections are garbage, you have no use for them; if dWAR turns out to be useless, that doesn't invalidate BBref.
   2249. Los Angeles El Hombre de Anaheim Posted: November 05, 2012 at 06:20 PM (#4293714)
Nate is not a pollster. Those companies provide a service which produces data. Nate provides analysis of that data. If his analysis is wrong, he's not providing anything at all.
But if the polls are misleading or wrong, that's worse than just bad analysis. If the polling companies are collecting and distributing bad data, that's far more damaging than anything Nate could do on his own.

Yet nobody is uncomfortable with the bulletproof nature of those companies. It's all just Nate.
   2250. Bring Me the Head of Alfredo Griffin (Vlad) Posted: November 05, 2012 at 06:20 PM (#4293715)
RCP is not claiming to know anything


One might think that they are at least implicitly claiming that their opinion has value, by virtue of the fact that they're offering it in the first place.

That depends what you mean by "inaccurate data." If the data is simply throwing darts at a dartboard such that it's complete garbage, then no, there's no gap. If the data is inaccurate because it makes certain assumptions and we can strip apart those assumptions and analyze it, then yes, there's a significant gap.


Under the scenario we were discussing, the data is (implicitly) inaccurate enough that it blew up all the poll aggregators' models. Which would seem to lean more toward the former scenario, rather than the latter, since if it were capable of being stripped, analyzed, and cleaned into a useful form, the poll aggregators would already be doing that.
   2251. tshipman Posted: November 05, 2012 at 06:21 PM (#4293716)
I don't follow these things as closely as many here, but I thought RCP just did a straight average of the polls that came out. They don't attempt to analyze the data; they just say, "Here's all the polls." It's the difference between BPro and BBref.


They exclude some polls (for various reasons that some people consider rational and some people consider slightly partisan). It's not that big of a deal, but it effects things at the margins.
   2252. Bring Me the Head of Alfredo Griffin (Vlad) Posted: November 05, 2012 at 06:21 PM (#4293717)
*ducks* (I'm an Obama supporter!)


Why would the ducks care who you're voting for? It's not like either candidate is planning to get rid of the corn subsidy.
   2253. zonk Posted: November 05, 2012 at 06:21 PM (#4293718)
Have we ever had an election where both Presidential candidates failed to win their home state? Mitt Romney is down some 20 points in Massachusetts, and of course, Kenya is not allowed to vote for our President. *ducks*


Actually - the record for home state loss is (IIRC) Hoover losing Iowa by 18 points in '32 - chances are very good Romney sets a new record for losing his home state.
   2254. Steve Treder Posted: November 05, 2012 at 06:21 PM (#4293719)
It doesn't strike me as quite right that someone who gives a probability should then be held to the standard of having to be 100% correct or else.

It isn't quite right. No one who understands the concept of statistical probability should do so.

But as Sam points out, there's what should be, and what is.
   2255. Rickey! trades in sheep and threats Posted: November 05, 2012 at 06:22 PM (#4293720)
If he relies on data that turns out not to have predicted the reality that came about, how's that his fault?


He's making a truth claim. There is no way around that. He is making a claim that his ability to aggregate data is valuable. If his aggregate results don't predict reality, his claim to value is going to take a hit. It's not like he's some guy sitting quietly behind his spreadsheets asking to be left alone while he maths. He picked a Twitter fight with Joe Scarborough. He's making bold claims about how to properly evaluate political science*. He's saying the famous, entrenched classes of political talkers are stupid and wrong. If he makes a mistake, he will pay for it. The fact that he was the noble, good guy doesn't change the fact that one mistake and Ned Stark loses his head. This ain't Yahtzee!
   2256. zonk Posted: November 05, 2012 at 06:23 PM (#4293722)
I don't follow these things as closely as many here, but I thought RCP just did a straight average of the polls that came out. They don't attempt to analyze the data; they just say, "Here's all the polls." It's the difference between BPro and BBref.



They exclude some polls (for various reasons that some people consider rational and some people consider slightly partisan). It's not that big of a deal, but it effects things at the margins.


Right - they'll list all polls, but their 'RCP Average' does do some picking and choosing (and I'm among the partisans that would say cherrypicking and choosing).
   2257. Slivers of Maranville descends into chaos (SdeB) Posted: November 05, 2012 at 06:24 PM (#4293723)

So if the guy I vote for gets sent to Congress then I don't have a state legislator?


Well, you have a Congressman. That's pretty damn good. Think of it as a kind of promotion.


The other thing is that with a district every 10K people, California is going to have close to 3,800 districts and the country as a whole will have about 31,000. That's just way too many. EDIT: There are about 3200 counties in the US. Counties vary wildly in population and size, but still, 10 people elected per county is a bit much.


I don't think so. If you combined state legislature with county boards and city council's you'd absorb 31,000 reps with plenty of slots left to fill. It wouldn't be hard to fiddle with the size of the districts and the number of slots to get them to match.

I think the best advantage of my plan is that running would be within the means of just about anybody who wanted. No need for multi-million dollar campaigns.
   2258. David Nieporent (now, with children) Posted: November 05, 2012 at 06:24 PM (#4293724)
state licensing boards are bogus. the nation would be a lot better if we reduced the licensing burden on individual businesses.

This is actually a Democrat position. Why do we require licences for beauticians again?
It would be nice if it were a Democrat position, but it isn't. (Look at the hate mail Matt Yglesias gets -- from the left -- when he suggests that perhaps barbers need not be licensed.)
   2259. Rickey! trades in sheep and threats Posted: November 05, 2012 at 06:27 PM (#4293725)
It would be nice if it were a Democrat position, but it isn't.


It's a bipartisan bit of idiocy. Show me a GOP controlled municipality and I'll show you a GOP regime that loves licensing fees. Show me a Dem city, and I'll show you a Dem regime that loves those same fees.
   2260. David Nieporent (now, with children) Posted: November 05, 2012 at 06:28 PM (#4293726)
In fairness to said right leaners, if Nate's right, it's bad news for the pundit class generally, regardless of party. The Chris Cillizas of the world have been pushing the dead heat meme that is just as contradicted by Silver's analysis as Michael Barone's Romney landslide prediction.
No, the dead heat meme is not contradicted by Silver's analysis. To quote Andy in 2239: "After seeing some of these folks at work over the past few years, nothing about their reading comprehension (or lack of it) surprises me. You'd think he was predicting that Obama was going to get 85% of the vote, rather than the 50.6% that he's actually projecting."
   2261. Morty Causa Posted: November 05, 2012 at 06:28 PM (#4293728)
i hear the concern about legalization but again, we have tried all the other stuff and it's just costly and we have lots of folks in prison learning to be better criminals. i think we have to explore the alternative some how some way.


That's true. All I'm saying is that however bad it is now, it might get worse with legalization. For instance, most crime is committed under the influence of drugs and/or alcohol or by addicts/alcoholics. Increasing use/addiction is going to do what?

Will grass be a commercial product? Will we see Marlboro Highs? What are the advertising rules--if any? Are there constitutional implications with regulating commercial speech wrt the grass business? How will protocols change wrt DUI's? Does any state really check for DUIs wrt anything but alcohol? How will that change?

How about lawsuits? Will FDA regulate? How?

This is just off the top of my head. Legalization just may be a Pandora's Box?
   2262. tshipman Posted: November 05, 2012 at 06:29 PM (#4293729)
It would be nice if it were a Democrat position, but it isn't. (Look at the hate mail Matt Yglesias gets -- from the left -- when he suggests that perhaps barbers need not be licensed.)


This is so far down as a substantive issue it's unreal (and that's a shame), but O's made headfakes in that direction at least. See, for example, the SOTU where he talks about "bad regulations."

It's a pretty standard "New Democrat" position that state licensing is overly onerous.
   2263. formerly dp Posted: November 05, 2012 at 06:32 PM (#4293730)
But this *is* the OT thread about *politics,* and politically Nate Silver can't go wide with his model (via the NYT, his increasingly confident-bordering-on-arrogant Twitter posts, national cable television appearances) and then complain that the masses he's courting (in order to push his personal brand and move into the stratosphere of the pundit class himself) don't really understand the nuance of his model.


I agree with your critique of Nate's tone, at least to some extent. His analysis of the pundit class, in voice though not in content, seems a little weird and not as measured as it should be if he really wants to break into those ranks, and most definitely reminds me of the pugilism between B-Pro and the beat writers a decade ago.

That said: I think there's a way for Silver to do an appearance on Daily Show and explain his method well-- he's not the nerdliest nerd with a nerdy book to ever be Stewart's guest, and Stewart does well with those types of guests (as well as can be expected in a 3-minute interview). The Real Time appearance seemed to be the bigger miscalculation. Bill Maher's a pretty terrible interviewer and not prone to bringing out the nuance in a person's argument. I haven't seen his other television appearances-- cut the cord with Comcast, so I miss the joys of CNN/MSN/Fox...
   2264. zonk Posted: November 05, 2012 at 06:32 PM (#4293731)

That's true. All I'm saying is that however bad it is now, it might get worse with legalization. For instance, most crime is committed under the influence of drugs and/or alcohol or by addicts/alcoholics. Increasing addiction is going to do what?

Will grass be a commercial product? Will we see Marlboro Highs? What are the advertising rules--if any? Are there constitutional implications with regulating commercial speech wrt the grass business? How will protocols change wrt DUI's? Does any state really check for DUIs wrt anything but alcohol? How will that change?

How about lawsuits? Will FDA regulate? How?

This is just off the top of my head. Legalization just may be a Pandora's Box?


Frankly, as much as I support legalization -- I have a hard time seeing where this tax windfall comes from...

I mean - let's imagine for a moment that the method of packaging and selling will essentially be the same as a pack of cigarettes... most smokers are plopping down taxes for a pack a day... I have an awfully, awfully hard time seeing how pot sells in the same quantity - I don't see how even the most ardent pothead smokes 20 joints a day.

   2265. Rickey! trades in sheep and threats Posted: November 05, 2012 at 06:32 PM (#4293732)
It's a pretty standard "New Democrat" position that state licensing is overly onerous.


A good distinction between New Dem moderates and more traditional GOP moderates is *where* they want to cut "the size of the state."
   2266. zonk Posted: November 05, 2012 at 06:35 PM (#4293734)
An endorsement you don't want...given that Jay's developed a real reputation as a #########.
   2267. formerly dp Posted: November 05, 2012 at 06:36 PM (#4293735)
In fairness to said right leaners, if Nate's right, it's bad news for the pundit class generally, regardless of party.


Absolutely. I'm speaking specifically to people on this board, who seem to be rooting for Nate to fail because they need Nate to fail to get a Romney victory. Which is understandable, but also doesn't help make media coverage of politics any better. I don't think Nate provides us with much of a solution on this front, as much as he does correctly identify the problem.
   2268. Steve Treder Posted: November 05, 2012 at 06:37 PM (#4293736)
No, the dead heat meme is not contradicted by Silver's analysis. To quote Andy in 2239: "After seeing some of these folks at work over the past few years, nothing about their reading comprehension (or lack of it) surprises me. You'd think he was predicting that Obama was going to get 85% of the vote, rather than the 50.6% that he's actually projecting."

Yes, it is contradicted by Silver's analysis. In the first place, in an electorate the size of the United States, 50.6% to 48.5% (his projection as of this morning) is not a dead heat.

More importantly, however close the national vote is or isn't, it's not what counts in the election. The state-by-state electoral college vote is, and Silver's analysis has consistently projected Obama as clearly winning that. He's never projected Romney as leading there, and the closest they ever were was 28 electoral votes on Oct. 12. That is not a dead heat.
   2269. Fernigal McGunnigle has become a merry hat Posted: November 05, 2012 at 06:38 PM (#4293737)
I think the best advantage of my plan is that running would be within the means of just about anybody who wanted. No need for multi-million dollar campaigns.


This I agree with.

My last little complaint is that (if my math is right) if the 10,000 or whatever electees were evenly split between parties and chosen by lots to a Congress of 435, there would be almost a 50% chance that one party would have 210 or fewer seats. So nearly half of every 50/50 election would result in one side with at least a 25-seat majority. (I expect that a massive electoral system like this wouldn't result in only two parties, but similar results would occur with more).

You can get around this with some system of proportional representation, where a party that wins X% of the elections gets X% of Congressional seats. But that leaves a small party with no chance at ever making Congress, if we assume that you need to win 1/435th of the elections to get into the Congress. So in 32,000 elections, a party needs 73 wins, which means that a party could sweep one of the smallest states and not get a place in Congress. So much for the North Dakota Independence Party.
   2270. Harveys Wallbangers Posted: November 05, 2012 at 06:39 PM (#4293738)
zonk

i am not saying you try and tax it back underground.

but certainly some form of tax will generate some amount of revenue

plus the cost savings of law enforcement, court work, prisons, blah, blah, blah

i say treat it as cigarettes and find out. we had prohibition for a decade or so. social experiment failed. oh well

we can't try?

this not trying stuff just drives me nuts.

trial and error is a perfectly fine methodology. as long as you learn from the errors. we are not doing that. we just keep doing the same pointless stuff.

it's ridiculous
   2271. Lassus Posted: November 05, 2012 at 06:39 PM (#4293739)
Assuming that the results of that study are representative, in what way is preventing 13% of the country's unwanted teenage pregnancies (and the numerous resulting abortions) an unworthy goal?

It gives more American teenage girls safe access to sex and orgasms, which causes wars to be fought in the Middle East, makes the baby jesus cry, and supports the theory of climate change.


I also like that David has been all over this thread but absolutely refuses to simply acknowledge that he's been proven incorrect on his assertion in #2187.]


(Look at the hate mail Matt Yglesias gets -- from the left -- when he suggests that perhaps barbers need not be licensed.)

Really? I'd like to see this hate mail, more out of morbid curiosity.
   2272. Jolly Old St. Nick Is A Jolly Old St. Crip Posted: November 05, 2012 at 06:41 PM (#4293740)
In fairness to said right leaners, if Nate's right, it's bad news for the pundit class generally, regardless of party. The Chris Cillizas of the world have been pushing the dead heat meme that is just as contradicted by Silver's analysis as Michael Barone's Romney landslide prediction.


No, the dead heat meme is not contradicted by Silver's analysis. To quote Andy in 2239: "After seeing some of these folks at work over the past few years, nothing about their reading comprehension (or lack of it) surprises me. You'd think he was predicting that Obama was going to get 85% of the vote, rather than the 50.6% that he's actually projecting."

The problem comes with conflating the probability of an Obama win of any margin (which Silver says is 85%) with the percentage of that projected win in the PV (50.6%) or the number of projected EV's (something north of 300). He's not at all projecting any "dead heat" in the sense that there's much doubt (according to him) that Obama's going to win the election, but he's also not projecting anything remotely resembling a landslide of 2008 proportions.

In fairness to the "horse race" opinionators, it's true that a small shift of votes in the swing states---within the margin of error, but all in the same direction, and that's the catch---could make Romney a winner. But that possibility is factored into that 85% number, and it's why that number isn't 99% instead.
   2273. spike Posted: November 05, 2012 at 06:41 PM (#4293741)
No, the dead heat meme is not contradicted by Silver's analysis. To quote Andy in 2239: "After seeing some of these folks at work over the past few years, nothing about their reading comprehension (or lack of it) surprises me. You'd think he was predicting that Obama was going to get 85% of the vote, rather than the 50.6% that he's actually projecting."


Wha? Conflation of vote percentages has nothing to do with the idea that an 85% chance is a significant, no way a dead heat lead. The two are different propositions entirely. From Friday, when Obama's probability got over 80%

Although the fact that Mr. Obama held the lead in so many polls is partly coincidental — there weren’t any polls of North Carolina on Friday, for instance, which is Mr. Romney’s strongest battleground state — they nevertheless represent powerful evidence against the idea that the race is a “tossup.” A tossup race isn’t likely to produce 19 leads for one candidate and one for the other — any more than a fair coin is likely to come up heads 19 times and tails just once in 20 tosses. (The probability of a fair coin doing so is about 1 chance in 50,000.)
   2274. Harveys Wallbangers Posted: November 05, 2012 at 06:44 PM (#4293743)
just for sam

weather in cleveland tomorrow is mostly sunny while it is rain the rest of the week

despite the dark forces at work we couldn't make the weather work in our favor tomorrow

gadzooks. foiled again
   2275. Harveys Wallbangers Posted: November 05, 2012 at 06:45 PM (#4293745)
wewather is going to stink tomorrow in wisky

the public is fired up though. the senate race in many ways has folks more excited than the presidential stuff
   2276. David Nieporent (now, with children) Posted: November 05, 2012 at 06:45 PM (#4293746)
Under the scenario we were discussing, the data is (implicitly) inaccurate enough that it blew up all the poll aggregators' models. Which would seem to lean more toward the former scenario, rather than the latter, since if it were capable of being stripped, analyzed, and cleaned into a useful form, the poll aggregators would already be doing that.
Let's suppose the polls fail to appropriately adjust for cell phone users. The poll aggregators don't have any way to know that (until after the fact). It would blow up the models, but could be adjusted for.
   2277. formerly dp Posted: November 05, 2012 at 06:47 PM (#4293747)
but certainly some form of tax will generate some amount of revenue


Really quickly: I think the tax gains need to be directly channeled back into dealing with the inevitable negative consequences that will emerge as a result of blanket legalization, so that it'll basically be a wash. Arguing that tax revenue will be a positive consequence of legalization makes for good rhetoric when you're trying to sell the idea, but I don't think it's terribly honest. The real benefits will be in the other areas you mentioned--less law enforcement, fewer prisons, ect. But I'd rather spend the money on substance abuse programs and counseling for the gainfully employed than on substance abuse and counseling for the imprisoned.
   2278. Joe Kehoskie Posted: November 05, 2012 at 06:49 PM (#4293748)
There is a distinction between discussions of what is and discussions of what ought to be. Most everyone outside of the Kehoskieite fringes recognize what Nate is doing and the distinction between his forecasting model and outright claims of certainty.

Oh, I recognize what Nate is doing, too. He's doing a lot of the same things as the pundit class he dislikes, except Nate pretends his non-falsifiable model offers four-decimal-place precision.

***
Of course, then we run into --- what exactly are the cable nets going to do for 10-12 months every couple years? They might actually start, you know, reporting on policy, issues, and such... the horror... the horror...

I agree that the pundit class needs an overhaul — this will be doubly true if Romney wins — but the cable networks and pundits are just giving their viewers what they want. If MSNBC or Fox News ran hourlong shows about the debt limit or inflation policy or any number of other issues, they'd be envious of C-SPAN's ratings.

***
But if the polls are misleading or wrong, that's worse than just bad analysis. If the polling companies are collecting and distributing bad data, that's far more damaging than anything Nate could do on his own.

Yet nobody is uncomfortable with the bulletproof nature of those companies. It's all just Nate.

Nate seems to trust the topline poll numbers more than many of the actual pollsters do. We know and Nate knows there's at least a chance the polls are wrong, so if his model trusts those numbers too much, that's on Nate.

Yesterday, when I said that "If Nate is right, he's a genius, and if he's wrong, it's someone else's fault," I expected a lot of pushback. But it seems the vast majority of Nate fans here believe just that. It's an amazing niche Nate and his fans have carved out for him: He privatizes the glory and socializes the blame. Where have we heard that before?
   2279. tshipman Posted: November 05, 2012 at 06:49 PM (#4293749)
we had prohibition for a decade or so. social experiment failed. oh well


You know, this is repeated a lot, but it's only debatably true at best.

Do you know what the average alcohol consumption was in 1919? 15 Gallons per capita, and most of that hard liquor.

Prohibition was successful at changing drinking habits from hard liquor to beer. Not a great policy, especially when you consider all the smuggling and gang violence, but definitely somewhat effective.
   2280. Steve Treder Posted: November 05, 2012 at 06:51 PM (#4293751)
Not a great policy, especially when you consider all the smuggling and gang violence

Other than that, how was the play, Mrs. Lincoln?

And Ciudad Juarez the past few years has made 1920s Chicago look like Shangri-La.
   2281. formerly dp Posted: November 05, 2012 at 06:53 PM (#4293752)
Prohibition was successful at changing drinking habits from hard liquor to beer.


Have we discussed the PBS documentary here yet? I thought it did a great job at setting up the problem prohibition responded to.

Edit: My understanding was that it did the opposite-- pushed people to drinking hard liquor, because it was a better bang for the risk than beer. But I'm working from memory on this one.
   2282. Jolly Old St. Nick Is A Jolly Old St. Crip Posted: November 05, 2012 at 06:59 PM (#4293753)
Let's suppose the polls fail to appropriately adjust for cell phone users. The poll aggregators don't have any way to know that (until after the fact). It would blow up the models, but could be adjusted for.

I would hope that the individual pollsters, the poll aggregators, and the aggregator analysts would all continue to fine tune their methodology in the light of unforeseen results. As Joe E. Brown would say, "Hey, nobody's perfect." What distinguishes the good ones from the bad ones is how well they adjust their models.

One thing I don't share with Nate, BTW, is his quarrel with the pundits. Most of them are fairly predictable, but when they get outside the realm of trying to make precise predictions and put their partisanship aside, they're more than capable of contributing insight to specific races. AFAIC Barone the opinionator is a complete blowhard, but Barone the walking encyclopedia of congressional districts is a national resource. Those are not contradictory statements.
   2283. Joe Kehoskie Posted: November 05, 2012 at 07:03 PM (#4293755)
Absolutely. I'm speaking specifically to people on this board, who seem to be rooting for Nate to fail because they need Nate to fail to get a Romney victory. Which is understandable, but also doesn't help make media coverage of politics any better. I don't think Nate provides us with much of a solution on this front, as much as he does correctly identify the problem.

As I said in the last thread, this is the one and only reason I'm rooting for Nate to be wrong. I'm certainly not rooting against him personally or professionally. (And as I said about six months ago, I'd love for there to be a mother's-basement math geek who doesn't care at all about politics but studies and sifts through the data like Nate. No matter how much Nate might be a stats wiz, it's tough to trust someone as a down-the-middle analyst when we know he has a rooting interest. There are a lot of ways Nate can be wrong without being biased, but I also don't believe a partisan can craft a truly bias-free model in the first place.)
   2284. tshipman Posted: November 05, 2012 at 07:04 PM (#4293756)
One thing I don't share with Nate, BTW, is his quarrel with the pundits. Most of them are fairly predictable, but when they get outside the realm of trying to make precise predictions and put their partisanship aside, they're more than capable of contributing insight to specific races. AFAIC Barone the opinionator is a complete blowhard, but Barone the walking encyclopedia of congressional districts is a national resource. Those are not contradictory statements.


I couldn't disagree more with this. Pundits have been lying to the public for as long as I can remember. Not being mistaken, but flatly lying. It's good to have the bullshit stripped away from them. There is a 100% bias against making the election outcome seem predetermined or predictable. There's a focus on "inside" information over accurate information. There's a willingness to spin for various campaigns just to keep on getting access.

Win or lose, I hope Nate is successful at blowing up bullshit mountain.
   2285. Monty Posted: November 05, 2012 at 07:06 PM (#4293757)
He privatizes the glory and socializes the blame. Where have we heard that before?


Everywhere?
   2286. Ray (RDP) Posted: November 05, 2012 at 07:07 PM (#4293758)
In fairness to the "horse race" opinionators, it's true that a small shift of votes in the swing states---within the margin of error, but all in the same direction, and that's the catch---could make Romney a winner. But that possibility is factored into that 85% number, and it's why that number isn't 99% instead.


Is this correct? I thought Nate was saying that the only thing that can make Romney a winner at this point is not a "shift" but is purely poll bias. He specifically discounts a shift because it's too late for that. The 15% represents the chance of pure bias at this point.

And Joe is right: According to people here, if Nate wins, he's a genius, and if he's wrong, it was faulty polls. But Nate has told us that the 15% accounts for faulty polls. So Obama really should win this thing; Nate is 85% certain of that. But if Michael Barone pegs the election instead, Nate takes a nuclear-level hit. Because it means that Nate probably seriously underestimated the effect of poll bias.
   2287. spike Posted: November 05, 2012 at 07:09 PM (#4293761)
There is a 100% bias against making the election outcome seem predetermined or predictable. There's a focus on "inside" information over accurate information.

David Axelrod on Dick Morris today - "I've had a foot in my mouth plenty of times, but it's always been my own!"
   2288. Eddo Posted: November 05, 2012 at 07:10 PM (#4293762)
Nate has told us that the 15% accounts for faulty polls. So Obama really should win this thing; Nate is 85% certain of that. But if Michael Barone pegs the election instead, Nate takes a nuclear-level hit. Because it means that Nate probably seriously underestimated the effect of poll bias.

I guess I'm still not seeing this. If I roll a die and claim there's a 16% chance it's NOT a six, and then a six comes up, have I seriously underestimated the fairness of the die?

(late) EDIT: I mean, I see that Nate *will* take a huge hit if this happens. I just don't see that it's rational.
   2289. McCoy Posted: November 05, 2012 at 07:11 PM (#4293763)
RCP at +.7 for Obama.
   2290. Steve Treder Posted: November 05, 2012 at 07:11 PM (#4293764)
So Obama really should win this thing; Nate is 85% certain of that. But if Michael Barone pegs the election instead, Nate takes a nuclear-level hit.

Everyone understands this, undoubtedly including Nate himself.
   2291. McCoy Posted: November 05, 2012 at 07:13 PM (#4293766)
And Joe is right: According to people here, if Nate wins, he's a genius, and if he's wrong, it was faulty polls.

Um, no. If his model is right then it is right. If his model is wrong there will be reasons for it to be wrong. Or are you arguing that if the model is wrong he is a moron?

He's a pretty smart guy regardless of how the election turns out. He's reached the top in at least two fields he's tried his hand in. That takes some intelligence.
   2292. Ray (RDP) Posted: November 05, 2012 at 07:15 PM (#4293767)
I guess I'm still not seeing this. If I roll a die and claim there's a 16% chance it's NOT a six, and then a six comes up, have I seriously underestimated the fairness of the die?


Part of the problem is that we'll never know, because there's only one "roll" here. With a die, you roll it 20 times or whatever and if a six comes up in 10 of them, you know your die is not a fair die.

With a Romney win, Nate has to say "I have a fair six-sided die and it came up 6." But how does he show that he had a fair six-sided die? That's the serious problem he would have. Why should we believe him, just because he says it's so?
   2293. Eddo Posted: November 05, 2012 at 07:16 PM (#4293768)
Part of the problem is that we'll never know, because there's only one "roll" here. With a die, you roll it 20 times or whatever and if a six comes up in 10 of them, you know your die is not a fair die.

With a Romney win, Nate has to say "I have a fair six-sided die and it came up 6." But how does he show that he had a fair six-sided die? That's the serious problem he would have. Why should we believe him, just because he says it's so?

OK, that's reasonable.
   2294. Ray (RDP) Posted: November 05, 2012 at 07:19 PM (#4293769)
No matter how much Nate might be a stats wiz, it's tough to trust someone as a down-the-middle analyst when we know he has a rooting interest. There are a lot of ways Nate can be wrong without being biased, but I also don't believe a partisan can craft a truly bias-free model in the first place.)


I disagree. I have no doubt that Nate's sole goal in devising his model is to come up with the best election predictor he can. It would simply be detrimental to his career to do anything else.

My problem is not so much with Nate as it is with the false certainty he professes (84.3% is an example, but even 84% would be false certainty rather than 80% or 85%), and with the liberals here who have bought into his model hook, line, and sinker - in no small part because he's a liberal.
   2295. McCoy Posted: November 05, 2012 at 07:19 PM (#4293770)
As if Nate's site is going to go dark for two years if he is wrong and he'll not write about the 2012 election ever again. Nor will anyone ever look into the election and the numbers/polls ever.
   2296. McCoy Posted: November 05, 2012 at 07:21 PM (#4293772)
Yeah, liberals on BTF like Nate's stuff because he is a liberal. Deep thinking there. No wait, scratch that. It isn't simply liking his working. No, they have been brainwashed by it.
   2297. tshipman Posted: November 05, 2012 at 07:23 PM (#4293775)

Part of the problem is that we'll never know, because there's only one "roll" here. With a die, you roll it 20 times or whatever and if a six comes up in 10 of them, you know your die is not a fair die.

With a Romney win, Nate has to say "I have a fair six-sided die and it came up 6." But how does he show that he had a fair six-sided die? That's the serious problem he would have. Why should we believe him, just because he says it's so?


First of all, I agree with the larger point that a Romney win would seriously harm Nate Silver's career, which has thus far been somewhat meteoric in its trajectory (fairly or unfairly).

How does he show that he had a fair six-sided die? I think that's the "after the election" stuff comes around to. I think he could demonstrate that a specific voter bloc was under-represented or over-represented in the polls, and dissect how that systematic bias poisoned the analysis. Like, let's say for example that 20% more seniors vote in 2012 than in 2008. That's really hard to predict, and voter screens will make that information be missed. In addition, he can take other pollsters projections and run them through his model and show the percentages. In other words, what prediction would just feeding the RCP average, the Pollster.com average, the HuffPo average generate?

I think those steps would demonstrate that systematic polling bias prevented accurate projections. Basically at this point, all the poll-people are projecting an Obama win, and most of them are doing so quite handily. It would be one thing if Nate was saying one thing and RCP was saying another. But that isn't the case. The only people who are projecting Romney wins are members of the Republican partisan media.
   2298. Steve Treder Posted: November 05, 2012 at 07:23 PM (#4293776)
My problem is not so much with Nate as it is with the false certainty he professes (84.6% is an example, but even 84% would be false certainty),

It isn't false certainty. In the sims he ran, Obama came up winning 84.6% of the time. That's a number. It's a model.

You've got your panties in a bunch about Nate because his model is predicting an Obama victory. There's nothing more to it than that.
   2299. Joe Kehoskie Posted: November 05, 2012 at 07:23 PM (#4293777)
There's a focus on "inside" information over accurate information.

Perhaps, but inside info. can have substantial value, and can radically alter the shape of a race. Does Nate's model account for the possibility of a last-minute revelation, like the Bush DUI story in 2000, which indisputably changed votes and/or affected turnout?

***
Um, no. If his model is right then it is right. If his model is wrong there will be reasons for it to be wrong. Or are you arguing that if the model is wrong he is a moron?

With a sample size of one election, how will anyone know if Nate's model was right? If Obama wins, how will we know he was an 85 percent favorite instead of a 95 percent favorite or a 52 percent favorite? And even if Nate's numbers are exactly correct, how will we know it was the model and not just dumb luck?
   2300. bunyon Posted: November 05, 2012 at 07:27 PM (#4293780)

It gives more American teenage girls safe access to sex and orgasms,


You vastly over-credit the American teenage boy.


Page 23 of 114 pages ‹ First  < 18 19 20 21 22 23 24 25 26 27 28 >  Last ›

You must be Registered and Logged In to post comments.

 

 

<< Back to main

BBTF Partner

Support BBTF

donate

Thanks to
Sheer Tim Foli
for his generous support.

Bookmarks

You must be logged in to view your Bookmarks.

Hot Topics

NewsblogPrimer Dugout (and link of the day) 11-28-2014
(5 - 11:58am, Nov 28)
Last: zack

NewsblogOTP Politics November 2014: Mets Deny Bias in Ticket Official’s Firing
(5090 - 11:51am, Nov 28)
Last: Rickey! trades in sheep and threats

NewsblogBaseball's most underrated Hall of Fame candidates. | SportsonEarth.com : Anthony Castrovince Article
(5 - 11:45am, Nov 28)
Last: Blastin

NewsblogOT - November 2014 College Football thread
(600 - 11:35am, Nov 28)
Last: AuntBea

Newsblog[Cricketer NOT baseball player] Phil Hughes dies after “pitch” to the head
(16 - 11:32am, Nov 28)
Last: frannyzoo

Hall of Merit2015 Hall of Merit Ballot Discussion
(58 - 11:30am, Nov 28)
Last: Chris Cobb

NewsblogOT: NFL/NHL thread
(8731 - 11:24am, Nov 28)
Last: PASTE Thinks This Trout Kid Might Be OK (Zeth)

NewsblogOT: Monthly NBA Thread - November 2014
(1141 - 11:02am, Nov 28)
Last: Moses Taylor, Moses Taylor

NewsblogOT: NBC.news: Valve isn’t making one gaming console, but multiple ‘Steam machines’
(1186 - 11:00am, Nov 28)
Last: Random Transaction Generator

NewsblogSandy Alderson says Mets can move quickly if a shortstop becomes available - NY Daily News
(38 - 10:54am, Nov 28)
Last: formerly dp

NewsblogBoston Red Sox prove (once again) that competitive balance in baseball will never exist | cleveland.com
(44 - 10:38am, Nov 28)
Last: McCoy

NewsblogDave Cameron: A proposed three-way swap for Red Sox, Mariners, Nationals
(54 - 10:15am, Nov 28)
Last: Darren

NewsblogSource: Tomas agrees to six-year deal with D-backs | MLB.com
(25 - 10:02am, Nov 28)
Last: PreservedFish

NewsblogYankees won't get into bidding war for Chase Headley: source - NY Daily News
(14 - 9:02am, Nov 28)
Last: The Yankee Clapper

NewsblogOT:  Soccer (the Round, True Football), November 2014
(550 - 8:50am, Nov 28)
Last: Shooty Is Disappointed With His Midstream Urine

Page rendered in 0.7828 seconds
54 querie(s) executed