3 Mathematical Models Predict the Next President

[spngr311]

What's funny is I didn't even have to read the first post to know what the results were! All I had to do was see that the OP was rn79870. He's like a democrat version of Ann Coulter.

[rn79870]

Are you disputing the findings of those three mathematical models? (and they were published before the polls started the 7+ point swing in favor of Obama)

[Beancooker]

You can not calculate opinion with mathmatics. Plain and simple.

You may get close, given the fact that you assume what people will vote for, but then it's just a hypothesis, not a calculation.

[rn79870]

Opinions are reflected in polls, not models. Mathematical models are more like Bayesian logic in that are a form of decision making and inferential statistics that deals with probability inference: using the knowledge of prior events to predict future events. That's why they have been so extreemly accurate, even absolutely accurate over the years.

[S13_love]

Opinions are reflected in polls, not models. Mathematical models are more like Bayesian logic in that are a form of decision making and inferential statistics that deals with probability inference: using the knowledge of prior events to predict future events. That's why they have been so extreemly accurate, even absolutely accurate over the years.

^ ...uh what he said

[mtcookson]

Hmm... this is interesting. d!ck Morris' Map:

http://w3.newsmax.com/a/morris...AB2-1

[rn79870]

Polls, like opinions (as you've cited here) are subjective.Models are objective. Two different animals, two different validities.

CNN puts the EV @ 233 for Obama, 189 for McCain. Gallop currently puts the popular vote at 48% for Obama, 44% for McCain. All these are subjective and not valid models. In that they are subjective in their data, you can correctly assume it will change (and change I'm sure it will) between now and November 4th.

[96Qowner]

Ah, but our esteemed American journalists almost always fail to quote Obama correctly. The phrase is "working Americans", meaning payroll taxes. Only those who earn more than $250,000 will see an increase in their income tax rates. Of course, Obama also intends to restore capital gains taxes to a 28% rate from the current 15%, so that's an 87% increase for any family who holds mutual funds, and that's about half of all US families.

Oops, I posted a falsehood. Obama only proposes to raise Capital Gains taxes to 20% and only for those with earnings over $250,000. It was good of the learned Obama supporters, who are SO very familiar with his plan, to correct me - oh ... yeah, they didn't. Whatever.

Capital Gains: Families with incomes below $250,000 will continue to pay the capital gains rates that they pay today. For those in the top two income tax brackets – likewise adjusted to affect only families over $250,000 – Obama will create a new top capital gains rate of 20 percent. Obama’s 20%rate is equal is the lowest rate that existed in the 1990s and the rate that President Bush proposed in 2001. It is almost a third lower than the rate that President Reagan signed into law in 1986

http://www.barackobama.com/pdf...L.pdf

My apologies, folks.

[spngr311]

Are you disputing the findings of those three mathematical models? (and they were published before the polls started the 7+ point swing in favor of Obama)

Not at all, just the constant posts from you that are always one sided. You always seem to find all the news stories that give positive news to Obama.

[marlin29311]

Mathematicl Models = Wrong

Real Models = Hot (i'd post pictures of hot women, but I am at work)

On a real note - they are PREDICTIVE models, not something uncommon from ones used by your local insurance company. Sure, something could model out to be the best risk or a sure thing, but then still have a plane fall on it.

The law of large numbers goverens insurance modeling - and it's a huge amount of exposures. Generally, and i mean GENERALLY, predictive modeling works. But not always. And it sucks when a your house burns down, even though it wasn't supposed to....

Haven't we also been predicted to have a huge asteroid hit the earth, have a supervolcano erupt, krackatoa (sp?) fall into the ocean, and a mountain of other things?

I wouldn't get too excited about how a model says the election is going to go. For all we know, Obama or McCain could go on TV tomorrow and say that they hate everyone who has anything but blue eyes and blond hair (gasp!)....

Until election day, I won't believe a thing....and I probably won't believe it after it too...

[HashiriyaS14]

Mathematicl Models = Wrong

Real Models = Hot (i'd post pictures of hot women, but I am at work)

On a real note - they are PREDICTIVE models, not something uncommon from ones used by your local insurance company. Sure, something could model out to be the best risk or a sure thing, but then still have a plane fall on it.

The law of large numbers goverens insurance modeling - and it's a huge amount of exposures. Generally, and i mean GENERALLY, predictive modeling works. But not always. And it sucks when a your house burns down, even though it wasn't supposed to....

Haven't we also been predicted to have a huge asteroid hit the earth, have a supervolcano erupt, krackatoa (sp?) fall into the ocean, and a mountain of other things?

I wouldn't get too excited about how a model says the election is going to go. For all we know, Obama or McCain could go on TV tomorrow and say that they hate everyone who has anything but blue eyes and blond hair (gasp!)....

Until election day, I won't believe a thing....and I probably won't believe it after it too...

This post wins.

That is all.

[rn79870]

Mathematicl Models = Wrong

Real Models = Hot (i'd post pictures of hot women, but I am at work)

On a real note - they are PREDICTIVE models, not something uncommon from ones used by your local insurance company. Sure, something could model out to be the best risk or a sure thing, but then still have a plane fall on it.

The law of large numbers governs insurance modeling - and it's a huge amount of exposures. Generally, and i mean GENERALLY, predictive modeling works. But not always. And it sucks when a your house burns down, even though it wasn't supposed to....

Haven't we also been predicted to have a huge asteroid hit the earth, have a supervolcano erupt, krackatoa (sp?) fall into the ocean, and a mountain of other things?

I wouldn't get too excited about how a model says the election is going to go. For all we know, Obama or McCain could go on TV tomorrow and say that they hate everyone who has anything but blue eyes and blond hair (gasp!)....

Until election day, I won't believe a thing....and I probably won't believe it after it too...

You're mistaking actuarial tables for models.Example.A mathematical model can provide that the sun will rise tomorrow, even the time the sun will rise. An actuarial table will provide insight on whether or not I'll be there to witness it. One predicts averages based on group study, the other outcome based on prior events. Google Bayesian logic and you'll understand how powerful math can be in this regard. Of course, the validity of any model is how well it has performed in the past. Did I mention 100% here?

This is best evidenced by the fact that all three models point to a Dem win, not to mention their outstanding (even remarkable) accuracy rate. One model, maybe not, two models, looking pretty good. Three models, it's a sure bet.

Not at all, just the constant posts from you that are always one sided. You always seem to find all the news stories that give positive news to Obama.

Anymore it's pretty hard to find anything positive about McCain. (Seriously.)

[marlin29311]

Im not talking about actuarial tables either. I'm talking about CAT management tables, or those used to determine the likelyhood of a natural disaster occuring and such. The types of tables that determine when a storm of size "x" will occur, etc.....100 year flood plains...all that type of stuff. These are based off mathematic algorithms - and they've been wrong. Katrina is a nice example of a storm that happened too soon, as well as the recent disasters in Asia as well...

[Jesda]

I can't help it if the three models predict a resounding win for Obama. I'm sure that if they showed a win for McCain someone would have posted about that.

Here's the problem with you, this thread, and this post.

You are more engaged in the "sport" of politics, beating the GOP, and beating McCain, and less interested in actual discussions of economics, science, and society. I do believe Barack Obama will win, but you should wipe the drool off your face and pretend to be a reasonable human being.

This entire thread is irrelevant to the discussions about issues that actually matter.

Actually, I'd go a step further and say that this thread is a microcosm of American politics today. We waste more time talking about lipstick and whacko preachers than debating economic policy. On a juvenile level, I suppose its more satisfying to beat the other "team" no matter what, but in the end you still lose if both teams "win" by pissing on American peace and prosperity.

I guess hacked emails bring higher cable news ratings than interviews and debates with economists, scientists, and people in business and healthcare. Scandal sells, but I expect more from a forum of people I know.

[SmithSR]

Here's the problem with you, this thread, and this post.

You are more engaged in the "sport" of politics, beating the GOP, and beating McCain, and less interested in actual discussions of economics, science, and society. I do believe Barack Obama will win, but you should wipe the drool off your face and pretend to be a reasonable human being.

This entire thread is irrelevant to the discussions about issues that actually matter.

Actually, I'd go a step further and say that this thread is a microcosm of American politics today. We waste more time talking about lipstick and whacko preachers than debating economic policy. On a juvenile level, I suppose its more satisfying to beat the other "team" no matter what, but in the end you still lose if both teams "win" by pissing on American peace and prosperity.

I guess hacked emails bring higher cable news ratings than interviews and debates with economists, scientists, and people in business and healthcare. Scandal sells, but I expect more from a forum of people I know.

Quoted for clarity of thought.

(his comment clarifies my thoughts on the subject)

[rn79870]

Here's the problem with you, this thread, and this post.

You are more engaged in the "sport" of politics, beating the GOP, and beating McCain, and less interested in actual discussions of economics, science, and society. I do believe Barack Obama will win, but you should wipe the drool off your face and pretend to be a reasonable human being.

Actually, impressing you is right about down there with....let me check....there it is, impressing you is right next to milking a water buffalo in the Ganges on my list of must do things. But seriously, I do enjoy your posts.

This entire thread is irrelevant to the discussions about issues that actually matter.

Psssttt you clicked the title to open it. You didn't have to you know. It mattered to me or I wouldn't have posted it. I'm trying to help people understand the futility of supporting McCain/Palin. (you agree with at least that part)

Actually, I'd go a step further and say that this thread is a microcosm of American politics today. We waste more time talking about lipstick and whacko preachers than debating economic policy. On a juvenile level, I suppose its more satisfying to beat the other "team" no matter what, but in the end you still lose if both teams "win" by pissing on American peace and prosperity..

Speaking or preachers, did you hear the one about Palin's preacher declaring a woman a witch. A witch who caused traffic accidents. Then he ran that poor African woman out of her home in order to protect the town from her witchcraft. Believe it or not, that's true and it's the same preacher Palin credits prayed her into the Governors mansion no less. Kind of makes Wright look sane by comparison.

I guess hacked emails bring higher cable news ratings than interviews and debates with economists, scientists, and people in business and healthcare. Scandal sells, but I expect more from a forum of people I know.

You know, all those would be interesting threads, and you'd probably start an interesting discussion if you did start one.

Quoted for clarity of thought.

(his comment clarifies my thoughts on the subject)

Quoted for clarity of thought. That's funny. I can picture you moving your mouth as you read those posts and clarify your thoughts.

[marlin29311]

Model #1 The Abramowitz modelEmory political science professor Alan Abramowitz has a mathematical election model that works. Abramowitz's system has correctly predicted the popular vote winner within two percentage points for every presidential election since 1988. This year, it's predicting an Obama win: 54.3 percent, versus McCain's 45.7 percent.

Now that i have more than a minute to look at this at work, I would love to point out a few things. I love selected data...

A) That's great that it hasn't failed since '88, but how about before that? You failed to mention that the model failed 4 out of 11 times during the 1948 - 1988 years. Its like saying the Yankees are 7-3 in their last ten, but for the season they're 65-83. Please give the whole picture when you are giving statistics.

B) So you're saying that using an investment hedging model is a great way to decide who's going to win the election? In the same economy where AIG had to be bailed out by the government, Lehman Bro's and Merril Lynch (and possibly more to come...)? Sorry, I'm not buying.

C) I'm too bored to find out all of the times it failed in the past, but I'll just say that it didn't work in 2000.

And now for the fun of it, I'm going to include a model that has McCain winning.

http://www.sciencedaily.com/re...3.htm

I stand by my original posts. Model's mean nothing when it comes to having millions of people make up their minds on a day in november. I'll wait until that day before I believe who will be the next president.

[rn79870]

And that was my claim...

Abramowitz’s forecast has correctly predicted the popular vote winner within two percentage points or less in every presidential election since 1988.

Same level of accuracy with the other two. Since we're post 1988, that is revelant.

Here is another one for you. 75 sites predicting the electoral college outcome. Guess who has a 60 to 15 advantage? That's what, 80% of the sited predicting the outcome of the EV are predicting an Obama victory?

http://3bluedudes.com/ProjectDatabase.htm

[SmithSR]

Quoted for clarity of thought. That's funny. I can picture you moving your mouth as you read those posts and clarify your thoughts.

You live for this.

[smockers83]

Polls, like opinions (as you've cited here) are subjective.Models are objective.

Wrong all over the place, as usual when it comes to these kinds of things. Polls and surveys (the real polls, not ones done by CNN, Fox, or ESPN that are unscientific) are a random sample of the population designed to represent the population of either a certain area or the country as a whole. The same questions are asked of everyone and everyone provides their own answers. That's objective. So it really depends on the poll as it could be surveying a specific population, which is why I asked before that we create a rule that when citing polls that a link is provided to the poll and its data so we can see how it was conducted and the questions asked to determine how valid the poll really is.

Models on the other hand attempt to interpret data. There can be numerous different ways to develop a model to interpret the same data set. This means we can have different models representing the same data yet the information used in each model may be different. That is subjective as the model is subject to the information plugged into it. I can try and model the unprecedented economic growth seen in Europe right after WWII, and many people have attempted to do so. This is something I studied for a good 4 months and I found only one model that really seemed to get it right without being subjective, but even then it was still subjective. The other models I researched were very subjective as they excluded many possibilities and data that should have been included or were not even thought of by the author/creator. But all of the models were pretty much correct in interpreting the growth of Europe, they just had problems interpreting the decline in growth experienced in the 70s. So in essence, they were all right for 30 some years, the one that was the least subjective was the best model as it allowed for the situations that caused the decrease in growth.

You said you put your kids through college, right? And if I remember correctly they went to very good schools. Maybe you should sit in on some classes at their schools. And also, you should apply the same rules to everyone, including yourself.

[spngr311]

I really hope these models prove to be wrong. I work as an engineer in the aerospace industry and I don't want to have to look for another job. It really sucks to hear a candidate say he wants to fund math and science education by cutting funding from an organization that provides jobs for many people who excelled and studied hard in math and science. There is a lot of irony there.

[rn79870]

You know Smocky, when you wake up Nov 5th. and you see that Obama is the president-elect, are you going to argue with the newspaper?

And spngr311 I certainly hope you don't lose your job. But the economy may play a part in that too.

[smockers83]

You know Smocky, when you wake up Nov 5th. and you see that Obama is the president-elect, are you going to argue with the newspaper?

I said I wasn't discrediting the models because I can't see the models, what goes into them, and how they work. Without that, its hard to do so. And apparently they have been accurate, so if that's the case, that's the case, big deal.

What I was discrediting is the claim that mathematical models are 100% accurate. To say that a model is 100% accurate before the real data is even in in light of what the model predicts, in this case a victory for Obama and given an Obama supporter, is just a taunt to those on the right and others supporting McCain, a trolling post if you will to stir the pot. That claim is also just flat out wrong. Am I actually scared of who's going to be the next president? No, I just wish it was over already because I'm honestly getting sick of it. Each one has their own take on the country, but I think either one, given the two choices, would do just fine. Just waiting for the debates to make my final decision.

It becomes more and more obvious that you don't read very well.

Umm, ok? That made a whole lot of sense, Bob. Nice way to interject politics into a nonpolitical issue. Sorry that you don't get it, I guess, and consequently attack me and others for it because we prove you wrong or tell you things you don't want to hear/believe? Can I just suggest something? Learn to be able to become vulnerable and deal with it appropriately and accept humility. They are great traits to have in one's personality.

Would I argue with the newspaper? No, because I don't get a newspaper for one thing, only Democrats get papers (joke, but I really don't get a paper). Would I argue against real results? No, because real results are real results. If Obama wins, Obama wins. If a third party wins, a third party wins. Big deal. If they win, obviously a majority of the people have faith in them so can that necessarily be completely bad? No.

Do I have a basis for arguing against a forecast? Every single person does. How many times do you get mad at the weatherman or say he's stupid? If the models were presented by someone else on the right side and made the same claims about models as you did, I would be making the same arguments that I am now, which quite frankly, have nothing to do with politics--this is what Jesda and myself attempted to point out to you earlier.

[smockers83]

And spngr311 I certainly hope you don't lose your job. But the economy may play a part in that too.

Not when the government directly removes the funding for the project he's on, which I'm guessing is a defense project since he mentions pulling out funding. Just a guess. But if Obama wants to help middle America, he can lower taxes but he can't raise taxes on the institutions that hire them. When you raise taxes on a producer, the first thing they do is reduce output, then they have to reduce costs and scale back their inputs to match output. That scaling back will be those in middle America. So exactly how does that help those people and how does that have to do with economy when the government directly affects the economy through taxation in this example?

[spngr311]

And spngr311 I certainly hope you don't lose your job. But the economy may play a part in that too.

I hope I don't lose my job either - but if Obama pulls funding from NASA to fund his "plan", the chances are pretty good that I (as an engineer) will lose my job as subcontractor that designs elements for NASA. It is not a direct result of the economy (NASA had a limited budget during the Clinton era as well), but a direct result of government budgets.

I feel a victim of irony if I lose my job. I'm very good at math and science, I worked hard in school, went to a good college and got a job in engineering and now I have to worry about it because of some idealist's concept of budgeting education by hurting the industry that will need it. Obama has said this in the past (though he flipped when he was in Brevard County Florida, home of KSC), so I have a legitimate concern. Talk about stressing me out - I have a lot riding on Obama not winning.

[smockers83]

I felt this was appropriate.

[smockers83]

You're mistaking actuarial tables for models.Example.A mathematical model can provide that the sun will rise tomorrow, even the time the sun will rise. An actuarial table will provide insight on whether or not I'll be there to witness it. One predicts averages based on group study, the other outcome based on prior events.

I love this statement. Actuarial tables are developed from statistical models designed to model and predict outcomes. A mortality table shows the probability of someone dying or how long a person born in a certain year is expected to live. These are all based on statistical models which are subject to the information plugged into them. Each insurance company uses their own models for the underwriting process, some are better able to predict perils better than others. Insurance statistics used to determine risk are developed from statistical models and previous data through the use of the Law of Large Numbers, as previously mentioned. In group insurance, there's a minimum number of people needed to qualify for group insurance so that the insurance company can classify the risk correctly--the more people in the group, the better numbers the insurance company can come up with for risk assessment and premium rates.

So we have a model that predicts if and when the sun will rise. We get an answer = Yes, it will rise and it will rise at 6:47am. We also have a model that will predict the chance of me getting hit by someone else in a car. We get an answer = Yes, its possible with a probability of .05 with a 95% confidence interval. If it looks and acts like a model, its probably a model.

A model that is 100% correct all the time is a mathematical law or formula, such as volume of a cube = x^3, the area of a circle = pi*r^2, or the area under a given curve is its derivative.

[AZhitman]

I heard there was a monkey in a zoo somewhere that can predict the outcome of a sports event with 99.1% accuracy. BFD.

Polls don't tell me the candidate is RIGHT.

They tell me Obama has done a much better job of obfuscation and smoke-screening.

If the MSM would actually report on his contradictions and confabulations, the numbers would swing. Alas, most don't.

[marlin29311]

I heard there was a monkey in a zoo somewhere that can predict the outcome of a sports event with 99.1% accuracy. BFD.

I think you are confusing that with the book Biff took from Marty in Back to the Future 2

And yea - I'm still not buying that "models" say that a candidate will win. And thank you smockers for talking about the insurance thing - i just didn't have the will to respond to that.

I actually created a new model that says that Ralph Nader is going to win. Unfortunatly, this is only based on the 3 people that sit around me at work while we were joking around....

Who will the next president be? WAIT until election day.

[AZhitman]

^ Solid.

And I stand corrected on "Eighties Movie Trivia", Alex.