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 24 of 114 pages ‹ First  < 22 23 24 25 26 >  Last ›
   2301. McCoy Posted: November 05, 2012 at 06:28 PM (#4293781)
Kind Vidor
   2302. Ray (RDP) Posted: November 05, 2012 at 06:29 PM (#4293782)
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.


I meant false precision.
   2303. tshipman Posted: November 05, 2012 at 06:29 PM (#4293783)
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?


Again, the problem is the focus on "inside" information rather than accurate information.

The debate coverage is a great example. The first topic of conversation is which side's people are out there spinning more happily, not a serious discussion about points of fact or policy.

The focus is on scoops uber alles, rather than on what is actually happening--especially on the campaign trail. How many times are we told that so and so "won the week"? One of the things that Nate's analysis made very clear was that winning the week was meaningless. Pundits want to portray the race as constantly shifting and fluid. In actuality, the race was quite stable. There were two shifts: Obama's Convention and the 47% comments, and the first debate. That was it. Everything else was just ####### gravity.
   2304. Joe Kehoskie Posted: November 05, 2012 at 06:30 PM (#4293784)
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.

I believe that's Nate's sole goal, too. I just don't believe he can accomplish it — i.e., I don't believe he, or anyone, can craft a model that's 100 percent bias-free. Nate's model isn't purely objective; he has to make all sorts of subjective assumptions on which the numbers depend, from which pollsters to use to the weighting of economic factors and all sorts of other things. And no matter how hard one tries, I doubt anyone has the ability to strip away 100 percent of their biases.
   2305. DA Baracus Posted: November 05, 2012 at 06:31 PM (#4293785)
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.


Which is why it's absurd to me that Nate Silver is getting criticized for giving a likelihood but RCP isn't for not.
   2306. Joe Kehoskie Posted: November 05, 2012 at 06:37 PM (#4293787)
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.

So we need huge sample sizes in baseball, but in politics, a single election plus some post-election analysis is dispositive?

Nate's model doesn't account for October surprises, or an Election Day blizzard in Iowa, or the possibility of thousands of voters in Florida not properly punching or otherwise filling out their ballots, or a hundred other things that can affect election results. He's claiming four-decimal-point precision in an area in which it's impossible to be precise.
   2307. OCF Posted: November 05, 2012 at 06:37 PM (#4293788)
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?


One logical corollary of this position: legalize importation, subject to a tariff, and subject to Department of Agriculture inspections for purity, pests, and so on - the usual things they'd look for with any agricultural commodity. The thing I don't see clearly there is what the effect would be on Mexico. Probably in the short run the Mexican criminal organizations would continue to dominate the import/export business. But would that hold together in the long run? Anything the U.S. could do to weaken those criminal organizations would be to the benefit of Mexico.
   2308. Lassus Posted: November 05, 2012 at 06:39 PM (#4293789)
And Joe is right: According to people here, if Nate wins, he's a genius, and if he's wrong, it was faulty polls.

I'm reasonably sure absolutely no one here thinks Nate is a genius. I think they think he is very, very good at his current job. A state that is not particularly unusual, and not a genius-level qualification.
   2309. Joe Kehoskie Posted: November 05, 2012 at 06:40 PM (#4293790)
The focus is on scoops uber alles, rather than on what is actually happening--especially on the campaign trail. How many times are we told that so and so "won the week"? One of the things that Nate's analysis made very clear was that winning the week was meaningless. Pundits want to portray the race as constantly shifting and fluid. In actuality, the race was quite stable. There were two shifts: Obama's Convention and the 47% comments, and the first debate. That was it. Everything else was just ####### gravity.

We didn't need Nate's analysis for that; RCP and Pollster.com told us the same.
   2310. McCoy Posted: November 05, 2012 at 06:44 PM (#4293791)
Legalizing marijuana isn't really going to put a dent in the Mexican cartels' coffers. Legalizing importation of it wouldn't change a thing in Mexico except make the government more corrupt.
   2311. Kiko Sakata Posted: November 05, 2012 at 06:44 PM (#4293792)
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?


In theory, it should, insofar as his model is built from historical precedents of past elections. If I remember correctly, Bush under-performed his last-week polls by about 3 points; that should be a data point in Nate's model when it calculates what the probability is that Obama blows his current 2-point lead (or whatever Nate measures it as). Now, does it SUFFICIENTLY or ACCURATELY account for such possibilities? I don't know, but certainly it's the sort of thing that ought to distinguish between a poll aggregator like RCP and an election modeler like Nate.
   2312. Joe Kehoskie Posted: November 05, 2012 at 06:52 PM (#4293796)
In theory, it should, insofar as his model is built from historical precedents of past elections. If I remember correctly, Bush under-performed his last-week polls by about 3 points; that should be a data point in Nate's model when it calculates what the probability is that Obama blows his current 2-point lead (or whatever Nate measures it as).

Perhaps, but as you said, there's simply no way for Nate to accurately estimate the odds of a story like that popping up.

What were the odds of Mitt Romney's full tax returns and financial statements being leaked? What were the odds of an old Barry Soetoro arrest report popping up? What were the odds of the bin Laden death photos being leaked? We don't know, and Nate can't know.
   2313. Danny Posted: November 05, 2012 at 06:54 PM (#4293797)
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.

Silver gives probabilities and projected popular votes for all 50 states and 33 senate races. There's plenty to judge his model on; it's certainly not just a binary Obama/Romney evaluation. If he nails the electoral vote while missing 10 states, that's a bad performance.
   2314. formerly dp Posted: November 05, 2012 at 06:55 PM (#4293798)
Bunyon:

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

You vastly over-credit the American teenage boy.


Comedy like that does not deserve to get lost in the flip.
   2315. Mefisto Posted: November 05, 2012 at 06:55 PM (#4293799)
Markos Moulitsas prediction here.
   2316. Ray (RDP) Posted: November 05, 2012 at 07:02 PM (#4293800)
Silver gives probabilities and projected popular votes for all 50 states and 33 senate races. There's plenty to judge his model on; it's certainly not just a binary Obama/Romney evaluation. If he nails the electoral vote while missing 10 states, that's a bad performance.


But really the presidential election is virtually the whole ballgame.
   2317. Jolly Old St. Nick Is A Jolly Old St. Crip Posted: November 05, 2012 at 07:04 PM (#4293801)
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 ######## 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 ######## mountain.


I think we may be just talking about a different set of pundits. I'm certainly not talking about the Sunday morning bloviators, if that's what you're thinking.
   2318. Jolly Old St. Nick Is A Jolly Old St. Crip Posted: November 05, 2012 at 07:05 PM (#4293802)
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.


Steve, you can explain this elementary point a hundred more times, but they're never going to get it.
   2319. David Nieporent (now, with children) Posted: November 05, 2012 at 07:06 PM (#4293804)
The focus is on scoops uber alles, rather than on what is actually happening--especially on the campaign trail. How many times are we told that so and so "won the week"? One of the things that Nate's analysis made very clear was that winning the week was meaningless. Pundits want to portray the race as constantly shifting and fluid. In actuality, the race was quite stable. There were two shifts: Obama's Convention and the 47% comments, and the first debate. That was it. Everything else was just ####### gravity.
What? No. I firmly believe - and have said it many times, to derision from Andy - that campaigns don't matter, that these day-to-day scores/gaffes that the media focus on have little if anything to do with the outcome. But "Nate's analysis" didn't show that; Nate's analysis showed that these things don't affect his model.

And on that note, in terms of validating his model: if Nate can correctly call all the states and the overall total, that tells us (to an extent) that his equations can translate the recent polls into results. But that does not validate his claim from July that Schlabotnik had a 71.4% chance of winning. (As a separate issue, that claim doesn't really make any conceptual sense. Bayes is rolling over in his grave. But that's a different issue.)
   2320. Jolly Old St. Nick Is A Jolly Old St. Crip Posted: November 05, 2012 at 07:07 PM (#4293805)
Again, the problem is the focus on "inside" information rather than accurate information.

The debate coverage is a great example. The first topic of conversation is which side's people are out there spinning more happily, not a serious discussion about points of fact or policy.

The focus is on scoops uber alles, rather than on what is actually happening--especially on the campaign trail. How many times are we told that so and so "won the week"? One of the things that Nate's analysis made very clear was that winning the week was meaningless. Pundits want to portray the race as constantly shifting and fluid. In actuality, the race was quite stable. There were two shifts: Obama's Convention and the 47% comments, and the first debate. That was it. Everything else was just ####### gravity.


Okay, I just saw this. And if this is the sort of punditry you're talking about, I completely agree with your point.
   2321. Joe Kehoskie Posted: November 05, 2012 at 07:10 PM (#4293806)
Steve, you can explain this elementary point a hundred more times, but they're never going to get it.

No, we get it. We just know that it's impossible for Nate to know if the assumptions in his model are 100 percent accurate. It's simply not possible to prove or disprove that Mitt Romney had an 18.2 percent chance or a 26.5 percent chance or a 55.8 percent chance of winning the election as of May 12, 2012 (or September 2, 2012, or November 4, 2012).
   2322. Danny Posted: November 05, 2012 at 07:17 PM (#4293809)
Silver gives probabilities and projected popular votes for all 50 states and 33 senate races. There's plenty to judge his model on; it's certainly not just a binary Obama/Romney evaluation. If he nails the electoral vote while missing 10 states, that's a bad performance.


But really the presidential election is virtually the whole ballgame.

That's silly. Everyone and their mother predicted Obama would win on election day in 2008. The reason Silver stood out is that he nailed 49 of 50 states (plus DC) and all of the senate races, with very small average errors for each state (smaller than Sam Wang's, for example).
   2323. Best Regards, President of Comfort, Esq. Posted: November 05, 2012 at 07:20 PM (#4293812)
Legalizing marijuana isn't really going to put a dent in the Mexican cartels' coffers. Legalizing importation of it wouldn't change a thing in Mexico except make the government more corrupt.
Okay, but if you make it legal to grow and sell in the United States, but illegal to import it into the United States, that *will* hurt the cartels.
   2324. Jolly Old St. Nick Is A Jolly Old St. Crip Posted: November 05, 2012 at 07:20 PM (#4293813)
There were two shifts: Obama's Convention and the 47% comments, and the first debate. That was it. Everything else was just ####### gravity.


What? No. I firmly believe - and have said it many times, to derision from Andy - that campaigns don't matter, that these day-to-day scores/gaffes that the media focus on have little if anything to do with the outcome. But "Nate's analysis" didn't show that; Nate's analysis showed that these things don't affect his model.

David, there's a difference between saying that "campaigns matter" and saying that every little stupid "event" or "gaffe" is of some monumental consequence. Some people try to use some sort of rigid economic formula (DJA; the unemployment rate; etc.) to predict the outcome, and Silver uses several economic indices in his model, but his "genius" (respectful eyeroll quotes) is that he's able to incorporate many factors from national polls to state polls to polling bias to economic events to a handful of other "events" such as the first debate, Sandy, etc., that might actually have affected the race. It's always going to be a work in progress, but so far he's certainly been holding his own.
   2325. Johnny Sycophant-Laden Fora Posted: November 05, 2012 at 07:22 PM (#4293814)
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.

Steve, you can explain this elementary point a hundred more times, but they're never going to get it.


No they don't get it, as JoeK so unwittingly shows in 2321, or they don't want to get it, doesn't really matter.
There are intelligent criticisms to be made of Silver's model, but that's not what we're getting, we're getting the same type of crap that creationists spew out when they think they are disproving evolution
   2326. McCoy Posted: November 05, 2012 at 07:22 PM (#4293815)
Okay, but if you make it legal to grow and sell in the United States, but illegal to import it into the United States, that *will* hurt the cartels.

Which means they get more into meth, cocaine, and heroin.
   2327. Harveys Wallbangers Posted: November 05, 2012 at 07:24 PM (#4293816)
the governor's performance in the first debate inspired the faithful

period. end of story

if he had tanked a good many gop members would have just walked away

i know this firsthand because i had to listen to it nonstop up to the debate

the governor ran a solid campaign.
   2328. spike Posted: November 05, 2012 at 07:24 PM (#4293817)
Well, I can guarantee you that if Nate's wrong there will be a much more transparent public discussion on his blog about it than you'll ever get out of George Will or Dick Morris come Wednesday.
   2329. spike Posted: November 05, 2012 at 07:26 PM (#4293818)
But that does not validate his claim from July that Schlabotnik had a 71.4% chance of winning. \

That's fair.
   2330. Tripon Posted: November 05, 2012 at 07:27 PM (#4293819)
Okay, but if you make it legal to grow and sell in the United States, but illegal to import it into the United States, that *will* hurt the cartels.


But they will just move on to something else. The Cartels aren't in the business of selling MJ. They're in the business of selling contraband, and will just sell the next illegal/prohibited item that will net them revenue. We're already see this with them moving into the meth trade. And the human smuggling trade. I very much doubt that any sane country will legalize meth.
   2331. Harveys Wallbangers Posted: November 05, 2012 at 07:29 PM (#4293820)
nate silver has elevated the discussion. there is value in that alone.
   2332. formerly dp Posted: November 05, 2012 at 07:29 PM (#4293821)
Okay, but if you make it legal to grow and sell in the United States, but illegal to import it into the United States, that *will* hurt the cartels.


Why would you need to make it illegal to import? I don't understand McCoy's logic here at all-- maybe it's self-evident, but I'm missing it.

Just refreshed and saw this:
Which means they get more into meth, cocaine, and heroin.


I am missing some of this conversation-- if there's legalization across the board, it seems like you're addressing all of these in one bundle. If we're only legalizing marijuana, we're gambling that cheap, legal and safe pot will be more appealing to people than expensive, illegal and potentially dangerous harder drugs. That's not an unsafe gamble, and certainly much better than the trainwreck that is the status quo.
   2333. Steve Treder Posted: November 05, 2012 at 07:30 PM (#4293822)
nate silver has elevated the discussion. there is value in that alone.

Big time.
   2334. Joe Kehoskie Posted: November 05, 2012 at 07:37 PM (#4293823)
Okay, but if you make it legal to grow and sell in the United States, but illegal to import it into the United States, that *will* hurt the cartels.

Perhaps, but the extent is unknown. If legal marijuana is heavily taxed, that will leave the cartels with pricing opportunities in their existing distribution channels. Take away the Hollywood and Wall Street types and the rich kids who will happily pay a higher price if it means no risk of legal problems and/or bad press, and I'm guessing most remaining drug users will probably be more concerned with getting the most marijuana for their money rather than complying with Uncle Sam's tax laws.

That aside, there's a huge network of people making big money from illegal drug sales. Those people aren't going to enroll in job training or go work at the mall. They'll have every incentive to undercut the legal market.
   2335. greenback calls it soccer Posted: November 05, 2012 at 07:37 PM (#4293824)
Steve, you can explain this elementary point a hundred more times, but they're never going to get it.

That's not entirely fair. Nate Silver is a celebrity in large part precisely because people are misunderstanding the nature of the model. And while he's done an admirable job of explaining the whole modeling limitation thing in his columns, the first thing people check on his blog are the three sets of numbers (odds, E[EV], E[PV]) above the graphs. He didn't have to set up his site that way, but he understands what sells blue jeans here.

Well, I can guarantee you that if Nate's wrong there will be a much more transparent public discussion on his blog about it than you'll ever get out of George Will or Dick Morris come Wednesday.

Because of stuff like that, Nate Silver will come out of this smelling like a rose one way or the other. He's a smart, creative guy with good analytical skills. If Kevin ####### Hassett can be an economic adviser to the Romney campaign, then I'm sure there's good work waiting for Silver somewhere.
   2336. Harveys Wallbangers Posted: November 05, 2012 at 07:38 PM (#4293825)
for some reason at election time i remember my middle son and his first election because he voted with us after turning 18 just before the 1984 election. and we both voted for reagan

that was tremendous
   2337. Joe Kehoskie Posted: November 05, 2012 at 07:41 PM (#4293826)
No they don't get it, as JoeK so unwittingly shows in 2321, or they don't want to get it, doesn't really matter.
There are intelligent criticisms to be made of Silver's model, but that's not what we're getting, we're getting the same type of crap that creationists spew out when they think they are disproving evolution

Nate's model is unfalsifiable. People's religious-level belief in Nate's model is what's deserving of the "creationist" label.
   2338. Harveys Wallbangers Posted: November 05, 2012 at 07:41 PM (#4293827)
for the record i wish my party would stop nominating poor senate candidates for my state. senator johnson is a d7mb8ss and governor thompson is past his prime.

why do you do this to me?????
   2339. Howie Menckel Posted: November 05, 2012 at 07:46 PM (#4293828)

"Barone the opinionator is a complete blowhard, but Barone the walking encyclopedia of congressional districts is a national resource. Those are not contradictory statements."

bravo.

I would be stunned if anyone who has actually seen Barone's discussions of Congressional history would be dismissive of it. But he is a very partisan blowhard in terms of predictions, as you note. And in 2012, that means he is to be pilloried for every fiber of his being.

I like to note that people are complicated - but not in today's political climate, and too often not here on BBTF. It's pretty sad.

   2340. Danny Posted: November 05, 2012 at 07:53 PM (#4293830)
Final swing state polls from PPP and Rasmussen:

State           PPP  RAS
Colorado        O
+6  R+3
Florida         O
+1  R+2
Iowa            O
+2  R+1
Nevada          O
+4  O+2
New Hampshire   O+2  O+2
North Carolina  tie  R
+6
Ohio            O
+5  tie
Virginia        O
+4  R+2
Wisconsin       O
+3  tie 
   2341. Kirby Kyle Posted: November 05, 2012 at 07:57 PM (#4293833)
I just wanted to drop in and thank zonk for his election night guide in post 2057. I've printed it out for easy following.
   2342. Best Regards, President of Comfort, Esq. Posted: November 05, 2012 at 07:57 PM (#4293834)
Rasmussen has Obama +2 in New Hampshire.
   2343. Danny Posted: November 05, 2012 at 08:02 PM (#4293836)
Rasmussen has Obama +2 in New Hampshire.

Thanks, I guess they haven't updated their map yet.
   2344. Monty Posted: November 05, 2012 at 08:04 PM (#4293838)
I just got diagnosed with a kidney stone. I say this not to complain, but as a warning: tomorrow, I'm staying home and watching all the election coverage.

And I'm taking a lot of Vicodin. So I may not be coherent. Enjoy!
   2345. DA Baracus Posted: November 05, 2012 at 08:05 PM (#4293839)
Even if Romney wins every single swing state by 5+ points, as long as he hits a high percentage of Congressional races Silver at the very least can get a job with the DNC. Adding in even more data points from internal polling can only help him improve an already good model.
   2346. Random Transaction Generator Posted: November 05, 2012 at 08:05 PM (#4293840)
I now have 41 different online predictions (websites and pundits) in the Google spreadsheet, found here.

If anyone finds any others, let me know.
   2347. Random Transaction Generator Posted: November 05, 2012 at 08:07 PM (#4293841)
I just got diagnosed with a kidney stone. I say this not to complain, but as a warning: tomorrow, I'm staying home and watching all the election coverage.


I wouldn't wish that on anyone.

The kidney stone OR the all-day election coverage.
   2348. Adam M Posted: November 05, 2012 at 08:07 PM (#4293842)
I'm staying home and watching all the election coverage.

And I'm taking a lot of Vicodin.


That sounds like a good combination. I will be checking this thread frequently.
   2349. Harveys Wallbangers Posted: November 05, 2012 at 08:10 PM (#4293843)
monty

sorry there big fella
   2350. Joe Kehoskie Posted: November 05, 2012 at 08:11 PM (#4293844)
... Silver at the very least can get a job with the DNC. Adding in even more data points from internal polling can only help him improve an already good model.

This (perhaps unintentionally) hints at two of the biggest questions about Nate's model in 2012:

1. Does Nate have access to Obama's internal polling, as he did in 2008?
2. If not, how much will that lack of access potentially hurt his projections?

Nate became a political star in 2008 based in large part on the notion that he was simply analyzing the same publicly available info. to which everyone else had access. That narrative ended up being false, as we now know Nate had an NDA with the Obama campaign that gave him access to high-quality polling info. to which no other analysts apparently had access.
   2351. Random Transaction Generator Posted: November 05, 2012 at 08:11 PM (#4293846)
Nate's model is unfalsifiable. People's religious-level belief in Nate's model is what's deserving of the "creationist" label.


Why is is unfalsifiable?

If he's wrong about a lot of the swing states, that would seem to suggest it has some "false" parts to it.
If the straight poll average does better than his formula, then that would suggest he's "false".
If he's off on a large portion of the states, but in the exact same direction/amount (say, O -0.5%), then he'll have to go through his numbers and find out why.

Unfalsifiable would be Dick Morris' prediction (using his "secret polls"). If he's right, fine. If he's wrong, then his "secret polls" were off, but we'll never see his numbers.
   2352. Ray (RDP) Posted: November 05, 2012 at 08:12 PM (#4293847)
I now have 41 different online predictions (websites and pundits) in the Google spreadsheet, found here.


I couldn't recognize all the names, but in general it seems that the rightwingers are predicting a clear Romney win, and the socialists are predicting a clear Obama win.

Just like here on BTF.

   2353. DA Baracus Posted: November 05, 2012 at 08:18 PM (#4293849)
This (perhaps unintentionally) hints at two of the biggest questions about Nate's model in 2012:

1. Does Nate have access to Obama's internal polling, as he did in 2008?
2. If not, how much will that lack of access potentially hurt his projections?


Yeah I thought about that right after I hit submit. From what I understand he only got access to them at one point in time, but he did get quite a lot of them. And he doesn't have them now.
   2354. Lassus Posted: November 05, 2012 at 08:20 PM (#4293850)
and the socialists are predicting a clear Obama win. Just like here on BTF.

Not all the lefties here are predicting a clear Obama win, you doofus. And you'd have to ask Harvey specifically, but I don't think he's really touting a clear Romney win, either.


2. If not, how much will that lack of access potentially hurt his projections?

HTF could it hurt his projections? Seems like it would only help.
   2355. DKDC Posted: November 05, 2012 at 08:20 PM (#4293852)
Is anyone else not voting?

I'm pretty sure I never re-registered when I moved a few years back and even if I did it just doesn't seem worth the effort in a non-competitive state to go out of my way to try to vote.
   2356. Random Transaction Generator Posted: November 05, 2012 at 08:22 PM (#4293854)
I couldn't recognize all the names, but in general it seems that the rightwingers are predicting a clear Romney win, and the socialists are predicting a clear Obama win.


Matt Dowd (a former Republican strategist) picked Obama 303, Romney 235.
He's not on speaking terms with Rove/Bush, but I don't think he's a Democrat at this point.

   2357. Joe Kehoskie Posted: November 05, 2012 at 08:29 PM (#4293858)
Why is is unfalsifiable?

It's simply not possible to prove that Romney or Obama had X percent chance of winning as of May 12, 2012, or September 10, 2012, or even November 4, 2012. You can trust that the model was right, but you can't prove it.

***
HTF could it hurt his projections? Seems like it would only help.

If Nate doesn't have access to the highest-quality polling, as he did in 2008, then he's flying a little more blind compared to 2008.
   2358. Lassus Posted: November 05, 2012 at 08:36 PM (#4293861)
It's simply not possible to prove that Romney or Obama had X percent chance of winning as of May 12, 2012, or September 10, 2012, or even November 4, 2012. You can trust that the model was right, but you can't prove it.

Congratulations, you've debunked seeing into the future. As soon as Nate tries doing this, you're all set.
   2359. DJS and the Infinite Sadness Posted: November 05, 2012 at 08:36 PM (#4293862)
RIP Elliot Carter :-(
   2360. spike Posted: November 05, 2012 at 08:42 PM (#4293865)
He had a good run.
   2361. Joe Kehoskie Posted: November 05, 2012 at 08:43 PM (#4293866)
Congratulations, you've debunked seeing into the future. As soon as Nate tries doing this, you're all set.

???

Nate doesn't make forecasts?
   2362. Gold Star - just Gold Star Posted: November 05, 2012 at 08:43 PM (#4293867)
Do Romney's people really think it's a good idea to trash Christie because the latter didn't want to leave his hurricane-ravaged state to accompany Romney at a PA rally?
   2363. McCoy Posted: November 05, 2012 at 08:45 PM (#4293868)
I almost wish Romney was winning and Nate was predicting Romney would win. It would almost be worth it just to hear all these people who are so anti-Nate now to jump through hoops to praise him.
   2364. Morty Causa Posted: November 05, 2012 at 08:47 PM (#4293869)
and the socialists are predicting a clear Obama win.


You say "socialists" like it's a bad thing?
   2365. Kiko Sakata Posted: November 05, 2012 at 08:48 PM (#4293870)
It's simply not possible to prove that Romney or Obama had X percent chance of winning as of May 12, 2012, or September 10, 2012, or even November 4, 2012. You can trust that the model was right, but you can't prove it.


In terms of bottom line and in terms of May 12th or September 10th, you're right, but Nate makes specific projections for every state (+DC) as well as for every Senate race. In 2010, I think he had specific predictions for governors' races, and, of course, in 2008, he had the same number of projections as he has here in 2012. Add those all up and that's something like 200 or so separate predictions. From that, you can start to get general ideas: of the candidates who Nate gave 70 - 80% odds of winning, how many actually won? What was his average error? I understand that in the grand scheme of things, nobody cares if Nate picks Romney to win Alabama by 26.4 points (that's what it says as I type this) and Romney really wins the state by 40 points or 10 points, but that's exactly the sort of thing one ought to look at in evaluating how good a job Nate Silver has done in his career (and in some ways, a state like Alabama is the true test for a lot of what Nate's doing: he shows exactly two polls there, both from the same firm, one in June and one in August; if Nate can get within a couple of points in a state like that on data that thin, that's impressive).
   2366. Famous Original Joe C Posted: November 05, 2012 at 08:53 PM (#4293872)
and the socialists are predicting a clear Obama win.


Pretty sure actual socialists generally dislike Obama too.
   2367. Random Transaction Generator Posted: November 05, 2012 at 08:54 PM (#4293873)
Pretty sure actual socialists generally dislike Obama too.


Not necessarily. Up here in Canuckistan, we're big fans of Obama.
:)
   2368. The Yankee Clapper Posted: November 05, 2012 at 08:57 PM (#4293874)
Is anyone else not voting?

I've met a few people with strongly partisan views who actually don't bother to vote because their individual vote is so unlikely to affect the outcome. Free riders, you might say. Most non-voters are more politically disengaged, at least according to conventional wisdom. You can make a case that it's actually a good thing that a fair number of people don't think their lives will be affected much if the "wrong" party takes power.
   2369. Random Transaction Generator Posted: November 05, 2012 at 09:04 PM (#4293877)
Is anyone else not voting?


At my work place, there was a guy who said he didn't vote in the last Canadian federal election. He made a big point about how it would be a waste of time (since the Conservatives were going to win anyways), and we were all suckers for lining up for 10 minutes to vote after work.

A few weeks later (after the Conservatives won), he wandered by a discussion a few of us were having about the federal government.
He started to rant about how "stupid" the Conservatives were, and one of my co-workers (who voted either Liberal or NDP or Green, I'm not positive) interrupted and said "You didn't vote, so why the #### should I care what you think about it?"

Icy silence.

Non-voter walked away.

I'm not as militant, but I have shut down someone who wanted to bad-mouth the Blue Jays, but claimed he didn't have a favourite team because he wasn't a fan.
   2370. spycake Posted: November 05, 2012 at 09:05 PM (#4293878)
It's simply not possible to prove that Romney or Obama had X percent chance of winning as of May 12, 2012, or September 10, 2012, or even November 4, 2012. You can trust that the model was right, but you can't prove it.

So a professional weather forecaster who uses current conditions and past trends to figure out weather probabilities is just wasting everyone's time? Just look at the temperature today, and look at the temperature west of you, right?

WTF is your argument here? That Nate uses one number, instead of just letting people eyeball all the polls for themselves and make up their own number? It's actually not that easy to "eyeball" the polls and pick up on everything -- numerous polls, covering different times, over at least a dozen key states, with different margins for error, different state/pollster histories, etc. And how much those polls are likely to change in the time remaining between now and the election (and how likely one state's change may affect other states, or may be affected by national changes).
   2371. The Ghost of Sox Fans Past Posted: November 05, 2012 at 09:06 PM (#4293879)
Do Romney's people really think it's a good idea to trash Christie because the latter didn't want to leave his hurricane-ravaged state to accompany Romney at a PA rally?

That too? I know they were mad that he praised Obama.
   2372. spycake Posted: November 05, 2012 at 09:08 PM (#4293880)
I couldn't recognize all the names, but in general it seems that the rightwingers are predicting a clear Romney win, and the socialists are predicting a clear Obama win.

Just like here on BTF.

And since you're above it all, let me guess: you're abstaining from predictions? (Pardon me if I've missed it, this thread has been multiplying like gremlins lately)
   2373. Shredder Posted: November 05, 2012 at 09:09 PM (#4293881)
Is anyone else not voting?
I'm not, but that's because I made an unplanned trip to California to visit my parents, since my dad's been in the hospital for a few weeks. I was supposed to fly back to Chicago on Saturday, but he almost died during surgery on Friday night, and is still in ICU with a breathing tube, so I figured it was a good idea to extend my trip. Obama is going to have to find a way to carry Illinois without me. Though in all seriousness, if I lived in the Joe Walsh/Tammy Duckworth district, I'd be a bit more disappointed. First presidential election since 1992 (my first time eligible) that I've missed.
   2374. Lassus Posted: November 05, 2012 at 09:09 PM (#4293882)
RIP Elliot Carter :-(

Oh no
   2375. Rickey! trades in sheep and threats Posted: November 05, 2012 at 09:10 PM (#4293883)
Is anyone else not voting?


Well, I'm randomly in New Jersey again this week, and my flight home tomorrow lands about 6:00ish. I never got over to early voting, so there's a reasonable chance I don't get to the polls before 7:00 tomorrow night. So GA may have one less voter despite the fact that I'm obviously engaged politically.

(The only things I will really miss if I don't get there is; continuing my string of not missing elections, and voting on a charter school initiative on the GA ballot.)
   2376. Jolly Old St. Nick Is A Jolly Old St. Crip Posted: November 05, 2012 at 09:14 PM (#4293887)
"Barone the opinionator is a complete blowhard, but Barone the walking encyclopedia of congressional districts is a national resource. Those are not contradictory statements."

bravo.

I would be stunned if anyone who has actually seen Barone's discussions of Congressional history would be dismissive of it. But he is a very partisan blowhard in terms of predictions, as you note. And in 2012, that means he is to be pilloried for every fiber of his being.


I had three separate book shops in the DC area between 1984 and 2006, and Barone used to come in a lot to the one I had in Georgetown, mostly to fill out his collection of 1930's state guides. I don't remember much of anything about his political opinions then, but in terms of political knowledge it was like talking baseball with a combination of Bill James and Fred Lieb. But two other writers, William Safire and Kevin Phillips, could have given him a pretty decent run for his money in terms of their knowledge of state and local politics almost down to the precinct level. People like that are always worth listening to no matter what you might think of their political opinions, which in Phillips' case changed a lot over the years.

I like to note that people are complicated - but not in today's political climate, and too often not here on BBTF. It's pretty sad.

Sometimes you just have to listen long enough for people to surprise you. I've been surprised here more than a few times.
   2377. The Ghost of Sox Fans Past Posted: November 05, 2012 at 09:18 PM (#4293888)
Shredder, so sorry about you dad's illness. Hope he is doing better.
   2378. Best Regards, President of Comfort, Esq. Posted: November 05, 2012 at 09:19 PM (#4293889)
Nate Silver update:

91.4% Obama win. 314.4 EV's, 50.9-48.2 popular vote. If he wasn't all-in before, he is now.
   2379. Best Regards, President of Comfort, Esq. Posted: November 05, 2012 at 09:20 PM (#4293890)
Florida 52.5% for Obama. Virginia 80.3%. Ohio 91.2%
   2380. Jolly Old St. Nick Is A Jolly Old St. Crip Posted: November 05, 2012 at 09:22 PM (#4293891)
At my work place, there was a guy who said he didn't vote in the last Canadian federal election. He made a big point about how it would be a waste of time (since the Conservatives were going to win anyways), and we were all suckers for lining up for 10 minutes to vote after work.

A few weeks later (after the Conservatives won), he wandered by a discussion a few of us were having about the federal government.
He started to rant about how "stupid" the Conservatives were, and one of my co-workers (who voted either Liberal or NDP or Green, I'm not positive) interrupted and said "You didn't vote, so why the #### should I care what you think about it?"


My sentiments exactly, though it's multiplied a thousandfold when it's a Floridian who voted for Nader in 2000.
   2381. bunyon Posted: November 05, 2012 at 09:23 PM (#4293892)
Shredder, sorry to hear it and I wish your dad a speedy and full recovery. I've been through it with my dad and it's not a fun journey but much better if you're there.



Just had an online chat with a former student who was a big Obama supporter (did a bunch of stuff to GOTV on campus) who is, in Florida, voting for Johnson, saying he can't stand Obama now. He spent some time in DC after graduation...no idea what got into him. Must make time for longer conversation.
   2382. Shooty Is Disappointed With His Midstream Urine Posted: November 05, 2012 at 09:24 PM (#4293893)
91.4% Obama win. 314.4 EV's, 50.9-48.2 popular vote. If he wasn't all-in before, he is now.

Man, I was 99% confident of an Obama win in 2008 and I'm nowhere near 91% confident this time around. Looking forward to this being over. I've been invited to an election party tomorrow but I don't think I could enjoy it as my nerves will be on edge.
   2383. GregD Posted: November 05, 2012 at 09:25 PM (#4293894)
Shredder, that sucks hard. I am so sorry.
   2384. Danny Posted: November 05, 2012 at 09:26 PM (#4293895)
The National Review's poll-guy answer to Nate Silver: Anyone can be Nate Silver. Just take polls and tweak them a bit as you see fit. We're all Nate Silver-we just dont get NYT gigs.

   2385. bunyon Posted: November 05, 2012 at 09:26 PM (#4293896)
My sentiments exactly, though it's multiplied a thousandfold when it's a Floridian who voted for Nader in 2000.

Right. Because voting for the guy that best represents your views is exactly like not voting at all.

If Al wanted those votes, he should have convinced the voters he was better than Nader. Shouldn't have been hard to do.
   2386. Kiko Sakata Posted: November 05, 2012 at 09:27 PM (#4293897)
91.4% Obama win. 314.4 EV's, 50.9-48.2 popular vote. If he wasn't all-in before, he is now.


Nate is giving Romney a 13.5% chance of winning the popular vote, and a 5.4% chance of winning the popular vote but losing the election. So, he has Romney's odds of winning the election, if he wins the popular vote, at only 60%. I kind of get where that's coming from (it's the "national polls and state polls are both right and the national toss-up goes to Romney" scenario), but that's a pretty striking number.
   2387. Joe Kehoskie Posted: November 05, 2012 at 09:29 PM (#4293899)
So a professional weather forecaster who uses current conditions and past trends to figure out weather probabilities is just wasting everyone's time? Just look at the temperature today, and look at the temperature west of you, right?

WTF is your argument here?

A weather forecast that predicts what will happen tomorrow has actual real-world value. A weather forecast on February 5 that predicts there's a 16.2 percent chance of rain on November 6 doesn't have much value (although, of course, the latter forecast at least has a hundred years of history on which to base its odds, rather than ~10 unique elections).

It's actually not that easy to "eyeball" the polls and pick up on everything -- numerous polls, covering different times, over at least a dozen key states, with different margins for error, different state/pollster histories, etc.

Actually, it is, at least according to Nate. The central premise of Nate's anti-pundit stance is that elections are highly stable.
   2388. David Nieporent (now, with children) Posted: November 05, 2012 at 09:30 PM (#4293900)
So a professional weather forecaster who uses current conditions and past trends to figure out weather probabilities is just wasting everyone's time? Just look at the temperature today, and look at the temperature west of you, right?
No.
WTF is your argument here?
I'm going to go out on a limb and guess that his argument is that "It's simply not possible to prove that Romney or Obama had X percent chance of winning as of May 12, 2012, or September 10, 2012, or even November 4, 2012. You can trust that the model was right, but you can't prove it."

As people are noting, Nate is making two different predictions:

(1) that Obama will get (using Larry's updated numbers from 2378) 50.9% PV and 314.4 EV
(2) that Obama has a 91.4% chance of winning

#1 can be validated. #2 cannot be.

#2 cannot be validated now, and it certainly can't be validated for a random day during the campaign.
   2389. Shredder Posted: November 05, 2012 at 09:31 PM (#4293901)
Thanks, guys. It was actually hip replacement, but there was very little bone left to attach the prosthesis. He already has a number of other ailments. He has Multiple Myeloma, which is in remission, and he's on dialysis three days per week. It was a rough surgery, and his blood pressure crashed a couple times. He'd been in the hospital for a couple weeks getting over a blood infection before the surgery. He's doing a lot better than the doctors probably expected a few days ago, but we're still looking forward to getting that breathing tube removed. You could tell the surgeon was pretty scared when he talked to my mother and me after the procedure. Orthopedic surgeons probably don't have to deliver news like that a lot. Hip replacements aren't usually life and death.
   2390. GregD Posted: November 05, 2012 at 09:32 PM (#4293902)
Nate is giving Romney a 13.5% chance of winning the popular vote, and a 5.4% chance of winning the popular vote but losing the election. So, he has Romney's odds of winning the election, if he wins the popular vote, at only 60%. I kind of get where that's coming from (it's the "national polls and state polls are both right and the national toss-up goes to Romney" scenario), but that's a pretty striking number.
Luckily the Republicans still have their unused talking points memos from 2000 (when they thought Gore would win the EC but lose the popular and so prepared to make the claim he didn't deserve to be president) all ready to go.
   2391. Shredder Posted: November 05, 2012 at 09:33 PM (#4293903)
Just had an online chat with a former student who was a big Obama supporter (did a bunch of stuff to GOTV on campus) who is, in Florida, voting for Johnson, saying he can't stand Obama now.
He must think the economy is too strong and wants to install someone who is REALLY intent on trashing it.
   2392. zonk Posted: November 05, 2012 at 09:34 PM (#4293908)
Do Romney's people really think it's a good idea to trash Christie because the latter didn't want to leave his hurricane-ravaged state to accompany Romney at a PA rally?



That too? I know they were mad that he praised Obama.


I have my suspicions about there being much of anything to this --

The supposed quote comes from an unnamed Romney staffer to HuffPost...



The Romney rally was held at a farm in Morrisville, Pa., not more than 20 minutes from Trenton, the New Jersey capital. The physical proximity of the event to New Jersey only added to questions in the Romney campaign about why Christie chose not to come.

"You can't tell me he couldn't have gone over there for a night rally," a Romney campaign source told HuffPost.


Now, Christie has a statement out saying that's a bunch of hogwash.

My guess is that HuffPo's "source" is Some Dude who volunteers for the campaign and just venting.
   2393. Morty Causa Posted: November 05, 2012 at 09:39 PM (#4293909)
Right. Because voting for the guy that best represents your views is exactly like not voting at all.

If Al wanted those votes, he should have convinced the voters he was better than Nader. Shouldn't have been hard to do.


Although sympathetic to Jolly Old's umbrage, I agree with this substantially. Vote your mind and heart. And both the candidate who coveted the vote and the voter who didn't get his first or second choice should take the consequent regret as a learning experience. I voted for Nader, but Gore had no chance in my state, and if he had, I would have definitely voted for him.
   2394. Best Regards, President of Comfort, Esq. Posted: November 05, 2012 at 09:40 PM (#4293910)
At least it's easy to grade Nate now. If Obama loses, he was wrong. The end.
   2395. Best Regards, President of Comfort, Esq. Posted: November 05, 2012 at 09:41 PM (#4293911)
Obama and Romney interviewing with Chris Berman at halftime of the football game.

There's your game-changer.
   2396. spike Posted: November 05, 2012 at 09:41 PM (#4293913)
Oh man, regardless of electoral prescience Nate is going to go down as one of the all time great trolls in history. The crying from the pundictocracy is reaching Sandy-esque levels.
   2397. Lassus Posted: November 05, 2012 at 09:43 PM (#4293914)
I like to note that people are complicated - but not in today's political climate, and too often not here on BBTF. It's pretty sad.

:-\

Yes, if only we had your depth. Sorry.
   2398. Danny Posted: November 05, 2012 at 09:45 PM (#4293915)
Right. Because voting for the guy that best represents your views is exactly like not voting at all.

Right, you're abstaining from helping decide who will be president.
   2399. GregD Posted: November 05, 2012 at 09:46 PM (#4293918)
My guess is that HuffPo's "source" is Some Dude who volunteers for the campaign and just venting.
This has to be true. I would not at all be surprised to see Romney's staff dirty up Christie tomorrow night once the polls close but to do so today would be amateur hour since it makes Romney look like a loser. The other possibility is that it is really a Ryan staffer; those guys definitely have to be looking around now for some elbows to throw to keep their guy viable.

Ex-Veeps always get nominations (except poor Dan Quayle!) But failed VP candidates, not so good. Dole is the last one, and he wasn't really running as a former VP nominee by that point 20 years later. Mondale sort of, but he was both an ex-Veep and a failed VP nominee. Kennedy was a defeated VP nominee as he lost to Kefauver at the open convention but that just proves the point as supposedly Joe Sr wanted JFK to get enough votes to increase his profile but to lose so he wouldn't be saddled with a failed Stevenson campaign.

The last true failed VP candidate to become a presidential nominee, unless I'm mistaken, is FDR.
   2400. Kiko Sakata Posted: November 05, 2012 at 09:47 PM (#4293919)
Oh man, regardless of electoral prescience Nate is going to go down as one of the all time great trolls in history. The crying from the pundictocracy is reaching Sandy-esque levels.


I haven't read a lot of these Silver-bashing columns, and now I'm glad of it. Damn, that reads like it could have been written by Murray Chass:

An election is not a mathematical equation; it is a nation making a decision. People are weighing the priorities of their society and the quality of their leaders. Those views, at any given moment, can be roughly measured. But spreadsheets don’t add up to a political community. In a democracy, the convictions of the public ultimately depend on persuasion, which resists quantification.


Nate Silver clearly needs to get his nose out of his spreadsheet and watch a campaign.
Page 24 of 114 pages ‹ First  < 22 23 24 25 26 >  Last ›

You must be Registered and Logged In to post comments.

 

 

<< Back to main

BBTF Partner

Support BBTF

donate

Thanks to
A triple short of the cycle
for his generous support.

Bookmarks

You must be logged in to view your Bookmarks.

Hot Topics

NewsblogJerry Crasnick on Twitter: "Jake Peavy has agreed on 2 yr deal with
(1 - 6:54am, Dec 19)
Last: Harveys Wallbangers

NewsblogThe 2015 HOF Ballot Collecting Gizmo!
(62 - 5:55am, Dec 19)
Last: Gonfalon Bubble

NewsblogPadres Acquire Derek Norris – MLB Trade Rumors
(17 - 4:58am, Dec 19)
Last: Harveys Wallbangers

NewsblogOT: Monthly NBA Thread - December 2014
(708 - 4:33am, Dec 19)
Last: It's a shame about Athletic Supporter

NewsblogOT: Politics - December 2014: Baseball & Politics Collide in New Thriller
(4881 - 4:33am, Dec 19)
Last: Los Angeles El Hombre de Anaheim

NewsblogThe 4 surprisingly quiet teams of the MLB offseason
(9 - 4:21am, Dec 19)
Last: baerga1

NewsblogRoyals sign Kris Medlen to two-year deal - MLB Daily Dish
(26 - 3:58am, Dec 19)
Last: Harveys Wallbangers

NewsblogThe Dan Shaughnessy Hall Of Fame Ballot
(59 - 2:45am, Dec 19)
Last: Gonfalon Bubble

NewsblogOT: NBC.news: Valve isn’t making one gaming console, but multiple ‘Steam machines’
(1352 - 2:09am, Dec 19)
Last: Maxwn

NewsblogSaint Pete City Council Tells Rays NYET!
(3 - 1:57am, Dec 19)
Last: Dale Sams

NewsblogOT: NFL/NHL thread
(9159 - 11:35pm, Dec 18)
Last: Jolly Old St. Nick Is A Jolly Old St. Crip

NewsblogPrimer Dugout (and link of the day) 12-18-2014
(42 - 9:34pm, Dec 18)
Last: Pops Freshenmeyer

NewsblogMatt Kemp's arthritic hips hold up deal with Padres
(41 - 9:05pm, Dec 18)
Last: Jeff Frances the Mute

NewsblogAre Wil Myers' flaws fixable? | FOX Sports
(103 - 8:09pm, Dec 18)
Last: ReggieThomasLives

NewsblogHow Will MLB Handle Big Changes With Cuba? - BaseballAmerica.com
(2 - 6:13pm, Dec 18)
Last: TDF, situational idiot

Page rendered in 0.8786 seconds
48 querie(s) executed