As a fun exercise, let's see who the next Nate Silver is (he predicted 49 out of 50 states and all of the Senate elections in 2008; 2010, not so much).
Please list the states and total electoral votes for each major political party candidate in this format:
Obama - CA, CN, DE, IL, etc. - 269 electoral votes
Romney - AL, AR, AK, etc. - 269 electoral votes
If you want, you can just list the swing states because most states are already decided.
House - Rep. 242; Dem. 193
Senate - Rep. 47; Dem. 53 (if you want, you can list the states most likely to switch for bonus points).
Next Wednesday, we can see who was closest and the MTG election champion.
I will be posting my predictions over the next day or so.
My predictions:
(battleground states only)
Romney - Fl, NC, VA, PA, NH, IA, CO, MN, MI, ME2 - Total 324 Electoral votes (more likely 304)
Obama - OH - Total 208 Electoral votes
Popular vote - Romney 52%, Obama 48% (likely 53.5/46.5)
House - 243 Rep.; 192 Dem. (more likely 241 Rep.; 194 Dem.)
Senate - 49 Dem.; 51 Rep. (more likely 50/50)
I think it is a little bold, but hey that's why we do these things.
I'll jsut default to intrade.com, as thats where I'd get my predicitons form anyway. I think a lot of the swing states will be closer than their odds rpedict, but I think they are generally correct -- the only potential disagreement is Ohio.
Ohio will go for Obama. It has polled pretty consistently for Obama, Romney's targeted Ohio because of the idea a Republican presidential candidate cannot win the presidency without Ohio. The latest attack ad from Romney will due him in I think. GM and Chrysler, huge producers of jobs around Ohio and bailed out by Obama, have stepped in to defend themselves against Romney's ad that claims they are shipping jobs to China. They are actually adding jobs in Ohio, they plan on selling more cars to China, and so many local papers have stepped up to fact check it it cements the rhetoric that Romney is a liar.
Florida will go for Romney. They have a strong GOP base down there, and in a toss up they lean Republican.
Colorado will go for Romney. I am not as optimistic as Nate Silver on Colorado's electoral votes. Colorado was where it all changed for Obama in the debates. There is a strong hold by the NRA and gun rights groups in Colorado since the Columbine massacre and the recent theatre shootings. Obama's comments on assault weapons will be enough to put them over the edge for Romney.
Virginia is the most interesting to me. I think it will go Romney. Silver thinks it will go Obama. Virginia voted Republican before Obama, and if they are not convinced by Obama's record then they will vote for Romney as the safe pick. But is Romney the safe pick for them? How do they feel about Republican Governor "Ultrasound" McDonnell? He ran on jobs jobs jobs and suddenly focused on abortion in the governorship. Also at issue is how much they feel about the storm recovery effort. West Virginia got snow on top of the high winds. If and when Virginia leans towards Democrats it will be part of a march South for Democrats. North Carolina could be next. It is on the fence for Virginia. It's possible they will vote Obama, it might even be likely they will vote Obama depending on how sentiment shapes up, but I saw them voting Romney just a couple days ago without a huge backlash against Governor McDonnell.
Wisconsin is similar, but they have been traditionally Democrat with a newly installed Republican Governor Scott Walker. Scott Walker has gone through a very very controversial fight over collective bargaining rights. The Obama campaign has been mobilized in Wisconsin for a couple years over that issue, and Wisconsin has benefited from the auto bailout. So I say Wisconsin for Obama, but a lot of Republican optimists see a Wisconsin Romney win somewhere in a national march against collective bargaining rights. But when they're in a toss up, the Democrat is their safe bet.
I see New Hampshire going Obama in a similar way I see Virginia going Romney. The surrounding states go blue, and unless New Hampshire goes in their natural contrarian direction, I think they will agree with the rest of us in the NE for Obama.
Iowa is polling for Obama, but I marked them for Romney. The Des Moines Register has come out for Romney, which I would imagine in most places would make more people vote Romney, but Iowans still seem to be polling for Obama. Decide for yourself.
Nevada: Obama.
So just a few days ago I was giving Obama the win with 275 electoral votes. I see the trend going further in Obama's favor.
Obama 290, Romney 248. Romney wins Florida, North Carolina and Virginia. Obama wins Ohio, Iowa, Colorado, Nevada and New Hampshire. Everything else goes the way you'd expect.
I will note that the state-level polls seem to be trending heavily against Romney, so I don't think it's unreasonable that he'd lose more ground in the next few days.
I haven't kept up on house races, really.
For senate, I'll go 53-47, assuming you count Maine and Vermont as Democrats.
Intrade odds don't reflect the gaps in polling - just odds that they'll win. If there's a persistent gap of 3 or 4% in the polling, the odds that the state will go that way might be more like 60 or 65%.
I'll just stick with Nate - his numbers and methodology are believable to me, I'm sure he'll have some blips that are off but we'll see.
(And do note, Presidential election year prediction is drastically different from non - the Senate and House candidates get swept up with the Presidential numbers which makes it a TON easier to predict, I've seen candidates try two consecutive cycles for Senate one with a Presidential election with a strong Presidential candidate and the next being an off year - and with the same message and everything and no serious gamechanging issues they'll poll 10-15% better when there's a strong Presidential candidate in their party [or the reverse against them]
Nate's model has historical accuracy for all PRESIDENTIAL year elections, off years he admits the model is quite flawed.)
The thing with Nate is he's a salesman. No one really knows exactly how his model works. He keeps it secret because people pay him to run it. He is very accurate, and he does stay non-partisan in his analysis. The one state he missed in '08 he thought would go to McCain and it went to Obama. He got the vast majority of the House, Senate, and Governorships right in the 2010 midterms, and did predict the Republican wave in the House legislature.
The data that is used in Silver's models and the academic models are viewable by the public: Public polling, campaign contributions, political advertising, state voting history, state unemployment, state industries, state officials currently holding office and public job approval of their elected officials. It's a great big jigsaw puzzle of trying to predict what happens next.
Nate's model has historical accuracy for all PRESIDENTIAL year elections, off years he admits the model is quite flawed.)
Yes, but that means his sample size is still one. It is not about whether or not you can create a model that can retroactively describe the elections, because then you end up with a model and prediction like the CU Boulder study that said Romney by 330+ electoral votes. Rather it is about success as a predictive model and Silver has only been successful once so far. He's never been unsuccessful, but still.
Private Mod Note
():
Rollback Post to RevisionRollBack
Asking people to remove quotes in their signatures is tyranny! If I can't say something just because someone's feelings are hurt then no one would ever be able to say anything! Political correctness is stupid.
Right, but as I said, he made sure before he started publicizing it that the formula made sure that all existing polled presidential elections fit - I believe (although I've barely started thumbing through his book, so talking about an interview) he quoted that he had the algorithm running privately during 2004 and it worked as well though.
But yea, for absolute accuracy wise he's 1:1 or 2:2 at best - but his book gives some phenomenal insight into how statistics and stuff works which gives me quite a bit of confidence, especially being someone that previous to his book found math to be simple besides statistics. (Although we'll see how much of a grasp I have after I finish it)
No, I'm pretty sure his sample size is very large, you're referring to his results or successful predictions, not his sample size. A sample size of one would be asking a single random voter who they are going to vote for, and then declaring that candidate the likely victor.
Rather it is about success as a predictive model and Silver has only been successful once so far. He's never been unsuccessful, but still.
Other models don't even have that kind of success. The true test will be a second election, but it's not a point against his model that it was successful the only time it was tested.
Nate's model has historical accuracy for all PRESIDENTIAL year elections, off years he admits the model is quite flawed.)
Yes, but that means his sample size is still one. It is not about whether or not you can create a model that can retroactively describe the elections, because then you end up with a model and prediction like the CU Boulder study that said Romney by 330+ electoral votes. Rather it is about success as a predictive model and Silver has only been successful once so far. He's never been unsuccessful, but still.
The number of surveys or samples (that is, elections) is one; the sample size, n, however, is not one.
I think the Democrats/Obama will have a narrow-margin win over the Republicans/Romney. I don't really know anything more than that and what I read from the news but I don't really care that much (really, what difference does it make? :frown:).
The sample size in determining the accuracy of Nate Silver's methodology as a predictive model is one. How hard is this to understand? I'm not talking about the methodology only having a sample size of one in determining the elections, but rather the accuracy having only one sample size. He's only predicted one presidential election. This isn't complicated people.
Private Mod Note
():
Rollback Post to RevisionRollBack
Asking people to remove quotes in their signatures is tyranny! If I can't say something just because someone's feelings are hurt then no one would ever be able to say anything! Political correctness is stupid.
The sample size in determining the accuracy of Nate Silver's methodology as a predictive model is one. How hard is this to understand? I'm not talking about the methodology only having a sample size of one in determining the elections, but rather the accuracy having only one sample size. He's only predicted one presidential election. This isn't complicated people.
A fair point, though the sample size would be more accurately represented as 50, as he didn't just call the election but rather called state by state numbers and was close to dead on in every single one in 2000. I think even Silver would tell you to have a healthy skepticism until samples get much larger, particularly in pointing to the relative paucity of presidential election polling until VERY recently.
As long as the point is "Silver's only done it once, therefore I will not rely too much on his method until it's worked a few more times", it's a fair point. Too often I've seen "Silver's a hack" as the argument, or "Silver's only done it once, therefore he's totally wrong", which basically translate to "I have partisan blinders on."
Given that I am a HUGE Nate Silver fan I don't think it's fair to read my response as "therefore he's wrong" but rather "therefore he's not God's gift to Presidential Election predictions."
As far as the 2004 argument goes, I had not heard that. That may lend more credence. The 2010 elections definitely do not because he even admitted his methodology was not fit for Congress elections and it was pretty much a crapshoot on his part.
Private Mod Note
():
Rollback Post to RevisionRollBack
Asking people to remove quotes in their signatures is tyranny! If I can't say something just because someone's feelings are hurt then no one would ever be able to say anything! Political correctness is stupid.
Believe me, I'm not reading your response that way. As I said, it's totally fair to point out that we need to see this work quite a few more times before putting anything like absolute faith in it.
If you look at an aggregate of swing state polls like Real Clear Politics and how much you trust the individual polls they average, Obama wins the election.
The sample size in determining the accuracy of Nate Silver's methodology as a predictive model is one. How hard is this to understand? I'm not talking about the methodology only having a sample size of one in determining the elections, but rather the accuracy having only one sample size. He's only predicted one presidential election. This isn't complicated people.
Why isn't it 50? He predicted outcomes for 50 states. He got 49 right. (50 for 51 if you count DC)
what many people don't realize is that BEFORE you factor in the swing states, obama has a significant lead over romney.
romney has to win most of the swing state electoral votes, whereas even if obama got a little less than half of them, he'd still win.
obama is doing pretty well for a president who is running on a recession. every single incumbent who ran on a recession lost. the exception is franklin roosevelt, but the economy did improve significantly under his first term.
what many people don't realize is that BEFORE you factor in the swing states, obama has a significant lead over romney.
romney has to win most of the swing state electoral votes, whereas even if obama got a little less than half of them, he'd still win.
obama is doing pretty well for a president who is running on a recession. every single incumbent who ran on a recession lost. the exception is franklin roosevelt, but the economy did improve significantly under his first term.
Ronald Reagan not only ran for a second term with unemployment in the mid 7% -- he won nearly all the states. There are a lot of similarities between Reagan and Obama, maybe even a landslide re-election.
If each swing state is a coin flip, this would still give Obama a significantly larger change to win. But then we also have this (along with non-predictive market polls which say essentially the same thing, but with smaller margins): http://www.intrade.com/v4/misc/key-races/states/
Even granting Romney the unlikely result of winning both Florida AND Ohio, Romney STILL only has roughly the same number of ways to win as Obama. I wouldn't be very confident in Romney when granting him TWO highly contested swing states only just puts him on equal footing...
Edit - Oh yeah, I should probably include an actual prediction. I think Romney will win Florida and NC; while Obama will win Ohio, Iowa, and Wisconson, granting him the victory. The other non-affiliated states aren't very important to my prediction.
Ohio and Wisconsin are trending for Obama in a big way. Certain polls have updated in the RCP aggregate of polls.
Iowa has a slight nod to Obama, but the big story is Romney losing votes to probably Gary Johnson. Both candidates are losing votes in Colorado. Obama is steady in Nevada, Romney is losing votes. Obama is losing votes in Florida while Romney stays steady. And Obama is gaining in a narrow race for Virginia.
All polls are from my previous post of RCP averages.
It isn't like the polls are the be all end all of the election, Harry Reid was expected to lose Nevada by 6 points in 2010, and he was re-elected. The polls are sometimes wrong, even when political bias is absent.
Private Mod Note
():
Rollback Post to RevisionRollBack
To post a comment, please login or register a new account.
Please list the states and total electoral votes for each major political party candidate in this format:
Obama - CA, CN, DE, IL, etc. - 269 electoral votes
Romney - AL, AR, AK, etc. - 269 electoral votes
If you want, you can just list the swing states because most states are already decided.
House - Rep. 242; Dem. 193
Senate - Rep. 47; Dem. 53 (if you want, you can list the states most likely to switch for bonus points).
Next Wednesday, we can see who was closest and the MTG election champion.
I will be posting my predictions over the next day or so.
My predictions:
(battleground states only)
Romney - Fl, NC, VA, PA, NH, IA, CO, MN, MI, ME2 - Total 324 Electoral votes (more likely 304)
Obama - OH - Total 208 Electoral votes
Popular vote - Romney 52%, Obama 48% (likely 53.5/46.5)
House - 243 Rep.; 192 Dem. (more likely 241 Rep.; 194 Dem.)
Senate - 49 Dem.; 51 Rep. (more likely 50/50)
I think it is a little bold, but hey that's why we do these things.
Florida will go for Romney. They have a strong GOP base down there, and in a toss up they lean Republican.
Colorado will go for Romney. I am not as optimistic as Nate Silver on Colorado's electoral votes. Colorado was where it all changed for Obama in the debates. There is a strong hold by the NRA and gun rights groups in Colorado since the Columbine massacre and the recent theatre shootings. Obama's comments on assault weapons will be enough to put them over the edge for Romney.
Virginia is the most interesting to me. I think it will go Romney. Silver thinks it will go Obama. Virginia voted Republican before Obama, and if they are not convinced by Obama's record then they will vote for Romney as the safe pick. But is Romney the safe pick for them? How do they feel about Republican Governor "Ultrasound" McDonnell? He ran on jobs jobs jobs and suddenly focused on abortion in the governorship. Also at issue is how much they feel about the storm recovery effort. West Virginia got snow on top of the high winds. If and when Virginia leans towards Democrats it will be part of a march South for Democrats. North Carolina could be next. It is on the fence for Virginia. It's possible they will vote Obama, it might even be likely they will vote Obama depending on how sentiment shapes up, but I saw them voting Romney just a couple days ago without a huge backlash against Governor McDonnell.
Wisconsin is similar, but they have been traditionally Democrat with a newly installed Republican Governor Scott Walker. Scott Walker has gone through a very very controversial fight over collective bargaining rights. The Obama campaign has been mobilized in Wisconsin for a couple years over that issue, and Wisconsin has benefited from the auto bailout. So I say Wisconsin for Obama, but a lot of Republican optimists see a Wisconsin Romney win somewhere in a national march against collective bargaining rights. But when they're in a toss up, the Democrat is their safe bet.
I see New Hampshire going Obama in a similar way I see Virginia going Romney. The surrounding states go blue, and unless New Hampshire goes in their natural contrarian direction, I think they will agree with the rest of us in the NE for Obama.
Iowa is polling for Obama, but I marked them for Romney. The Des Moines Register has come out for Romney, which I would imagine in most places would make more people vote Romney, but Iowans still seem to be polling for Obama. Decide for yourself.
Nevada: Obama.
So just a few days ago I was giving Obama the win with 275 electoral votes. I see the trend going further in Obama's favor.
I will note that the state-level polls seem to be trending heavily against Romney, so I don't think it's unreasonable that he'd lose more ground in the next few days.
I haven't kept up on house races, really.
For senate, I'll go 53-47, assuming you count Maine and Vermont as Democrats.
(And do note, Presidential election year prediction is drastically different from non - the Senate and House candidates get swept up with the Presidential numbers which makes it a TON easier to predict, I've seen candidates try two consecutive cycles for Senate one with a Presidential election with a strong Presidential candidate and the next being an off year - and with the same message and everything and no serious gamechanging issues they'll poll 10-15% better when there's a strong Presidential candidate in their party [or the reverse against them]
Nate's model has historical accuracy for all PRESIDENTIAL year elections, off years he admits the model is quite flawed.)
Re: People misusing the term Vanilla to describe a flying, unleash (sometimes trample) critter.
There are academic models viewable by the public to try and predict the election
http://journals.cambridge.org/action/displayJournal?jid=PSC
The data that is used in Silver's models and the academic models are viewable by the public: Public polling, campaign contributions, political advertising, state voting history, state unemployment, state industries, state officials currently holding office and public job approval of their elected officials. It's a great big jigsaw puzzle of trying to predict what happens next.
Yes, but that means his sample size is still one. It is not about whether or not you can create a model that can retroactively describe the elections, because then you end up with a model and prediction like the CU Boulder study that said Romney by 330+ electoral votes. Rather it is about success as a predictive model and Silver has only been successful once so far. He's never been unsuccessful, but still.
But yea, for absolute accuracy wise he's 1:1 or 2:2 at best - but his book gives some phenomenal insight into how statistics and stuff works which gives me quite a bit of confidence, especially being someone that previous to his book found math to be simple besides statistics. (Although we'll see how much of a grasp I have after I finish it)
Re: People misusing the term Vanilla to describe a flying, unleash (sometimes trample) critter.
No, I'm pretty sure his sample size is very large, you're referring to his results or successful predictions, not his sample size. A sample size of one would be asking a single random voter who they are going to vote for, and then declaring that candidate the likely victor.
Other models don't even have that kind of success. The true test will be a second election, but it's not a point against his model that it was successful the only time it was tested.
I think the Democrats/Obama will have a narrow-margin win over the Republicans/Romney. I don't really know anything more than that and what I read from the news but I don't really care that much (really, what difference does it make? :frown:).
Re: People misusing the term Vanilla to describe a flying, unleash (sometimes trample) critter.
A fair point, though the sample size would be more accurately represented as 50, as he didn't just call the election but rather called state by state numbers and was close to dead on in every single one in 2000. I think even Silver would tell you to have a healthy skepticism until samples get much larger, particularly in pointing to the relative paucity of presidential election polling until VERY recently.
As long as the point is "Silver's only done it once, therefore I will not rely too much on his method until it's worked a few more times", it's a fair point. Too often I've seen "Silver's a hack" as the argument, or "Silver's only done it once, therefore he's totally wrong", which basically translate to "I have partisan blinders on."
As far as the 2004 argument goes, I had not heard that. That may lend more credence. The 2010 elections definitely do not because he even admitted his methodology was not fit for Congress elections and it was pretty much a crapshoot on his part.
Ohio: Obama +2.3
Wisconsin: Obama +5.0
Iowa: Obama +2.0
Colorado: Obama +0.9
Nevada: Obama +2.7
New Hampshire: Obama +2.0
Florida: Romney +1.2
Virginia: Romney +0.5
Obama wins with 290
Re: People misusing the term Vanilla to describe a flying, unleash (sometimes trample) critter.
Why isn't it 50? He predicted outcomes for 50 states. He got 49 right. (50 for 51 if you count DC)
http://www.intrade.com/v4/misc/electoral-map/
There was also a CU Boulder study mentioned that predicts Romney will win with 330 electoral votes. The study looks at state by state economic conditions.
http://www.colorado.edu/news/releases/2012/10/04/updated-election-forecasting-model-still-points-romney-win-university
1. He's basically John Kerry 2.0
2. The Tea Party's legacy is starting to loose it's branding image
The Tea Party needs to do something to start winning over major metropolitan areas if the states are won in the urban and microurban areas.
Ambition must be made to counteract ambition.
Individualities may form communities, but it is institutions alone that can create a nation.
Nothing succeeds like the appearance of success.
Here is my principle: Taxes shall be levied according to ability to pay. That is the only American principle.
romney has to win most of the swing state electoral votes, whereas even if obama got a little less than half of them, he'd still win.
obama is doing pretty well for a president who is running on a recession. every single incumbent who ran on a recession lost. the exception is franklin roosevelt, but the economy did improve significantly under his first term.
Ronald Reagan not only ran for a second term with unemployment in the mid 7% -- he won nearly all the states. There are a lot of similarities between Reagan and Obama, maybe even a landslide re-election.
http://www.nytimes.com/interactive/2012/11/02/us/politics/paths-to-the-white-house.html
If each swing state is a coin flip, this would still give Obama a significantly larger change to win. But then we also have this (along with non-predictive market polls which say essentially the same thing, but with smaller margins):
http://www.intrade.com/v4/misc/key-races/states/
Even granting Romney the unlikely result of winning both Florida AND Ohio, Romney STILL only has roughly the same number of ways to win as Obama. I wouldn't be very confident in Romney when granting him TWO highly contested swing states only just puts him on equal footing...
Edit - Oh yeah, I should probably include an actual prediction. I think Romney will win Florida and NC; while Obama will win Ohio, Iowa, and Wisconson, granting him the victory. The other non-affiliated states aren't very important to my prediction.
Iowa has a slight nod to Obama, but the big story is Romney losing votes to probably Gary Johnson. Both candidates are losing votes in Colorado. Obama is steady in Nevada, Romney is losing votes. Obama is losing votes in Florida while Romney stays steady. And Obama is gaining in a narrow race for Virginia.
All polls are from my previous post of RCP averages.
It isn't like the polls are the be all end all of the election, Harry Reid was expected to lose Nevada by 6 points in 2010, and he was re-elected. The polls are sometimes wrong, even when political bias is absent.