538.com map

[tokar]

October 28th

Changes
In McCain's Favor:
North Carolina from LEAN Obama to TOSS-UP
Virginia from SAFE Obama to LIKELY Obama

In Obama's Favor:
New Hampshire from LIKELY Obama to SAFE Obama

Sen. Barack Obama/Sen. Joe Biden: 349 (-15)
Sen. John McCain/Gov. Sarah Palin: 152 (NC)
Toss-up: 37 (+15)

Snapshot

Sen. Barack Obama/Sen. Joe Biden: 375 (NC)
Sen. John McCain/Gov. Sarah Palin: 163 (NC)

Changes
In McCain's favor:
Arkansas from 6-10 to 10+
Florida from DEM3-6 to DEM0-3 (at +2.5)
Minnesota from DEM10+ to DEM6-10 (at +9.5)
Pennsylvania from DEM10+ to DEM6-10 (at +9.0)

In Obama's favor:

0-3 point lead = 30%, 3-6 point lead = 40%, 6-10 point lead = 60%, 10+ point lead = 80%

[Alcon]

WHY THE HELL DID NATE SILVER WEIGHT THE LA TIMES POLL SO DAMN HIGH?!

I'm thinking of taking the site off of my bookmarks.  It's becoming freaking aggravating.

Nate's methodology has its screwy points, one of which is probably over-emphasizing polls with limited track records.  But he can't weigh based on how much he likes a poll.  That being said, I wish he included more past record stuff from Presidential elections so as to down-weigh that sort of thing.

But he can't just go, "LA Times, kerplunk," or anything.

[Ronnie]

I just saw the new map on 538.  It seems reasonable, but if Nate Silver weighted the crappy polls properly, the race could be a lot closer.

For example, he weighted the LA Times OH poll (Obama + 9) at 0.91 and the Big Ten Indiana Poll (Obama +9.5) at 0.77.  Now, those weightings aren't that high, but they're enough to change his ratings a lot.

[Verily]

I just saw the new map on 538.  It seems reasonable, but if Nate Silver weighted the crappy polls properly, the race could be a lot closer.

For example, he weighted the LA Times OH poll (Obama + 9) at 0.91 and the Big Ten Indiana Poll (Obama +9.5) at 0.77.  Now, those weightings aren't that high, but they're enough to change his ratings a lot.

His ratings are objective, not subjective. There's not that much data for most polls, which is unfortunate, but you can't just say, "This poll looks bad," and dismiss it if you're going to do a scientific analysis. Nate Silver's formula is designed to be a mathematical modeling of the election, not a personal forecast. In a forecast, you can be as arbitrary as you want; in science, you have to be objective in weightings.

[muon2]

I have seen comments on a number of threads that suggest that though there might be some tightening of the race towards McCain, the state polls disagree and remain solid for Obama. As a counterpoint I would note the significant increase in combinations in the EC moving toward McCain on 538. That is indicative of movement detected by the 538 analysis.

It's become even more polarized on 538 after this week's polling. The combination with 375 EV now dominates and makes up almost 20% of their simulated outcomes. That is equivalent to saying that there is very little left for tossups, since they give the flexibility to create additional combinations. I suspect it is also due to the closeness of the election date, since there is less time now for states to shift.



There also seems very little more for Obama to get in their model. It drops off precipitously after 378 EV, and I see no combinations beyond 410 EV.

[Verily]

I have seen comments on a number of threads that suggest that though there might be some tightening of the race towards McCain, the state polls disagree and remain solid for Obama. As a counterpoint I would note the significant increase in combinations in the EC moving toward McCain on 538. That is indicative of movement detected by the 538 analysis.

It's become even more polarized on 538 after this week's polling. The combination with 375 EV now dominates and makes up almost 20% of their simulated outcomes. That is equivalent to saying that there is very little left for tossups, since they give the flexibility to create additional combinations. I suspect it is also due to the closeness of the election date, since there is less time now for states to shift.



There also seems very little more for Obama to get in their model. It drops off precipitously after 378 EV, and I see no combinations beyond 410 EV.

Generally speaking, this is because McCain has made gains in the states he has absolutely no business losing -- North Carolina and Indiana, and maybe Missouri (although the polling there is ambiguous). None of these states win McCain the election, they just reduce the Obama landslide. McCain hasn't made gains in any of the essential states, however.

[Ronnie]

I just saw the new map on 538.  It seems reasonable, but if Nate Silver weighted the crappy polls properly, the race could be a lot closer.

For example, he weighted the LA Times OH poll (Obama + 9) at 0.91 and the Big Ten Indiana Poll (Obama +9.5) at 0.77.  Now, those weightings aren't that high, but they're enough to change his ratings a lot.

His ratings are objective, not subjective. There's not that much data for most polls, which is unfortunate, but you can't just say, "This poll looks bad," and dismiss it if you're going to do a scientific analysis. Nate Silver's formula is designed to be a mathematical modeling of the election, not a personal forecast. In a forecast, you can be as arbitrary as you want; in science, you have to be objective in weightings.

If Nate wants to have accurate results, he needs to looks at the internals.  I mean in some SUSA polls, the internals were just plain wrong, and he weighted the polls above average.  I imagine that the internals in the Indiana Big Ten Poll, for example, was off.  

I mean, I understand Nate Silver wanting to give pollsters a chance until they are proven wrong, but that theory pretty much lacks common sense.  

[Alcon]

How would you suggest he adjusts for internals?  That's pretty damn complicated stuff -- although I agree using primary performance as the exclusive baseline is a bad idea.

[Verily]

I just saw the new map on 538.  It seems reasonable, but if Nate Silver weighted the crappy polls properly, the race could be a lot closer.

For example, he weighted the LA Times OH poll (Obama + 9) at 0.91 and the Big Ten Indiana Poll (Obama +9.5) at 0.77.  Now, those weightings aren't that high, but they're enough to change his ratings a lot.

His ratings are objective, not subjective. There's not that much data for most polls, which is unfortunate, but you can't just say, "This poll looks bad," and dismiss it if you're going to do a scientific analysis. Nate Silver's formula is designed to be a mathematical modeling of the election, not a personal forecast. In a forecast, you can be as arbitrary as you want; in science, you have to be objective in weightings.

If Nate wants to have accurate results, he needs to looks at the internals.  I mean in some SUSA polls, the internals were just plain wrong, and he weighted the polls above average.  I imagine that the internals in the Indiana Big Ten Poll, for example, was off.  

I mean, I understand Nate Silver wanting to give pollsters a chance until they are proven wrong, but that theory pretty much lacks common sense.  

Mathematics lacks common sense. You're missing the point entirely.

[Ronnie]

I just saw the new map on 538.  It seems reasonable, but if Nate Silver weighted the crappy polls properly, the race could be a lot closer.

For example, he weighted the LA Times OH poll (Obama + 9) at 0.91 and the Big Ten Indiana Poll (Obama +9.5) at 0.77.  Now, those weightings aren't that high, but they're enough to change his ratings a lot.

His ratings are objective, not subjective. There's not that much data for most polls, which is unfortunate, but you can't just say, "This poll looks bad," and dismiss it if you're going to do a scientific analysis. Nate Silver's formula is designed to be a mathematical modeling of the election, not a personal forecast. In a forecast, you can be as arbitrary as you want; in science, you have to be objective in weightings.

If Nate wants to have accurate results, he needs to looks at the internals.  I mean in some SUSA polls, the internals were just plain wrong, and he weighted the polls above average.  I imagine that the internals in the Indiana Big Ten Poll, for example, was off.  

I mean, I understand Nate Silver wanting to give pollsters a chance until they are proven wrong, but that theory pretty much lacks common sense.  

Mathematics lacks common sense. You're missing the point entirely.

Fine, if Nate Silver wants to play around with his mathematical formulas, let him go ahead.  But if he weights the polls wrong, what's the point?

[muon2]

Generally speaking, this is because McCain has made gains in the states he has absolutely no business losing -- North Carolina and Indiana, and maybe Missouri (although the polling there is ambiguous). None of these states win McCain the election, they just reduce the Obama landslide. McCain hasn't made gains in any of the essential states, however.

I agree that none of this gets McCain to 270. But it does show movement at the state level, at least based on the 538 analysis. Tightening in the heavily contested states makes sense, just as the hard Obama states may well be seeing an excess surge away from McCain due to last month resource allocation with bandwagon effect in those solid O states.

[Verily]

I just saw the new map on 538.  It seems reasonable, but if Nate Silver weighted the crappy polls properly, the race could be a lot closer.

For example, he weighted the LA Times OH poll (Obama + 9) at 0.91 and the Big Ten Indiana Poll (Obama +9.5) at 0.77.  Now, those weightings aren't that high, but they're enough to change his ratings a lot.

His ratings are objective, not subjective. There's not that much data for most polls, which is unfortunate, but you can't just say, "This poll looks bad," and dismiss it if you're going to do a scientific analysis. Nate Silver's formula is designed to be a mathematical modeling of the election, not a personal forecast. In a forecast, you can be as arbitrary as you want; in science, you have to be objective in weightings.

If Nate wants to have accurate results, he needs to looks at the internals.  I mean in some SUSA polls, the internals were just plain wrong, and he weighted the polls above average.  I imagine that the internals in the Indiana Big Ten Poll, for example, was off.  

I mean, I understand Nate Silver wanting to give pollsters a chance until they are proven wrong, but that theory pretty much lacks common sense.  

Mathematics lacks common sense. You're missing the point entirely.

Fine, if Nate Silver wants to play around with his mathematical formulas, let him go ahead.  But if he weights the polls wrong, what's the point?

Your definition of "wrong" is a non-mathematical one. To weight them any way other than 538 is weighting them would be wrong. You can't weight polls differently based on your gut. Unless there is data to support weighting them weakly (which there isn't, just a sense that they have something wrong), you can't justify weighting them weakly in a mathematical sense.

[Verily]

Generally speaking, this is because McCain has made gains in the states he has absolutely no business losing -- North Carolina and Indiana, and maybe Missouri (although the polling there is ambiguous). None of these states win McCain the election, they just reduce the Obama landslide. McCain hasn't made gains in any of the essential states, however.

I agree that none of this gets McCain to 270. But it does show movement at the state level, at least based on the 538 analysis. Tightening in the heavily contested states makes sense, just as the hard Obama states may well be seeing an excess surge away from McCain due to last month resource allocation with bandwagon effect in those solid O states.

True, although even then most of the movement towards McCain in 538 has been based on national polls (which are also factored in), not state polls. The only state polls which have shown movement to McCain of late are in North Carolina; Indiana has stayed steady in the state polling but is dragged towards McCain based on national polls.

[Ronnie]

I just saw the new map on 538.  It seems reasonable, but if Nate Silver weighted the crappy polls properly, the race could be a lot closer.

For example, he weighted the LA Times OH poll (Obama + 9) at 0.91 and the Big Ten Indiana Poll (Obama +9.5) at 0.77.  Now, those weightings aren't that high, but they're enough to change his ratings a lot.

His ratings are objective, not subjective. There's not that much data for most polls, which is unfortunate, but you can't just say, "This poll looks bad," and dismiss it if you're going to do a scientific analysis. Nate Silver's formula is designed to be a mathematical modeling of the election, not a personal forecast. In a forecast, you can be as arbitrary as you want; in science, you have to be objective in weightings.

If Nate wants to have accurate results, he needs to looks at the internals.  I mean in some SUSA polls, the internals were just plain wrong, and he weighted the polls above average.  I imagine that the internals in the Indiana Big Ten Poll, for example, was off.  

I mean, I understand Nate Silver wanting to give pollsters a chance until they are proven wrong, but that theory pretty much lacks common sense.  

Mathematics lacks common sense. You're missing the point entirely.

Fine, if Nate Silver wants to play around with his mathematical formulas, let him go ahead.  But if he weights the polls wrong, what's the point?

Your definition of "wrong" is a non-mathematical one. To weight them any way other than 538 is weighting them would be wrong. You can't weight polls differently based on your gut. Unless there is data to support weighting them weakly (which there isn't, just a sense that they have something wrong), you can't justify weighting them weakly in a mathematical sense.

I guess it's not mathematical, but, uhm, I thought Nate Silver's primary intention was accuracy?  Pollsters develop different strategies over time, which makes their accuracy vary.  I think it's inappropriate for him to base a lot of his weightings over previous elections, which include the primaries.  Plus, he seems to give very arbitrary weightings to newer pollsters.

If you're advocating that it's okay to be completely wrong based on "mathematics", I really don't understand. 

[tokar]

October 29th

Changes
In McCain's Favor:
Nebraska CD1 from LIKELY McCain to SAFE McCain

In Obama's Favor:
Nevada from LEAN Obama to LIKELY Obama

Sen. Barack Obama/Sen. Joe Biden: 349 (NC)
Sen. John McCain/Gov. Sarah Palin: 152 (NC)
Toss-up: 37 (NC)

Snapshot

Sen. Barack Obama/Sen. Joe Biden: 375 (NC)
Sen. John McCain/Gov. Sarah Palin: 163 (NC)

Changes
In McCain's favor:
Colorado from DEM6-10 to DEM3-6 (at +5.5)
Ohio from DEM3-6 to DEM0-3 (at +2.9)
Virginia from DEM6-10 to DEM3-6 (at +5.9)

In Obama's favor:
Nevada from 0-3 to 3-6 (at +3.3)

0-3 point lead = 30%, 3-6 point lead = 40%, 6-10 point lead = 60%, 10+ point lead = 80%

[Bob Dole '96]

These maps are so grossly wrong it's not even funny.  Silver is wrong.

The polls need to be weighted to realistic partisan ID (NOT REGISTRATION -- TOTALLY DIFFERENT THINGS).  Using 2006 would be a starting point.  Some of these polls are using registration and that's why you're getting these horrible Obama numbers.

For example, the other day a Survey USA poll came out that had Obama +10 in Virginia.  Guess what?  It had a D+9 advantage in the poll.  Well, in 2006, it was R+3 in party ID at the polls.  Reweight it -- and guess what -- McCain is ahead.

Also, take a look at early voting numbers.  They are not meeting Obama's expectations.

On the flip side, McCain's problem is that due to Obama's broken promises on campaign spending, he has a smaller map to work with -- basically two paths -- either winning Ohio, Florida, Pennsylvania, and Virginia -- or, Ohio, Florida, Virginia, Colorado, and Nevada.  (this is assuming he holds other things).  Iowa could replace Colorado.  This is different from 2004 when Bush had Minnesota, Wisconsin, Oregon, and Michigan in play.

[Bacon King]

Question. What exactly is the "return on investment" map?

[The Ex-Factor]

Question. What exactly is the "return on investment" map?

"The ratio of a state’s Tipping Point percentage to the number of eligible voters in each state, calibrated such that an average state has a Return on Investment Index of 1.0. This is intended to represent the marginal return from spending one additional dollar (or other type of campaign resource) in that state."

So right now Nevada's ROI is 13.8 - which means that a dollar spent there is 13.8 times more useful than a dollar spent nationwide.

[??????????]

The correlation between refusal to accept (the closest thing we have to) fact and blue avatars is astounding to the point of embarrassment.

Quit being like Supersoulty.

[Horus]

The correlation between refusal to accept (the closest thing we have to) fact and blue avatars is astounding to the point of embarrassment.

Quite being like Supersoulty.

You mean reasonable? The fact is McCain is behind, and this map makes a fairly logical projection about how the race would turn out if voting were today. Nothing more, nothing less.

[??????????]

Yes, but the election won't be held today, .538 is a partisan Dem site, et al.

[Horus]

Its predictions always seemed about 1-2 points Dem-friendly, but to predict a McCain loss is not being negative, it is being realistic. Not to say he can't win, but if I were a Republican, I'd be quite down right now.

[Person Man]

Yes, but the election won't be held today, .538 is a partisan Dem site, et al.

...and Republican sites say the same thing.

EP and CHt, for example.

[Ronnie]

Yes, but the election won't be held today, .538 is a partisan Dem site, et al.

That's going a bit too far, but Nate's weightings sometimes skew the model toward Obama.

[??????????]

It's frustrating to see people acting like children just because the guy they don't like is going to win.  Especially when I'm a part of that group and people are going to associate me with that because of others' behavior.

Yes, because supporting your candidate means you're "childish". Sorry but this whole board is turning into nothing but left wing hackery. Just because I choose not to be all glum, gloom and doom doesn't make me 'childish' as you say. It makes me a person who believes my candidate DOES have a shot. Are all the other people who volunteer at McCain HQs around the country "children who can't wake up to reality"? Please. I'll fight for my candidate to the bitter end, if that's not the course you choose to go, whatever.

And btw, their are PLENTY of reasons for any real conservative to dislike Obama. I'm very concerned about how dangerous Obama will be, mostly because he is naive and untried. Sadly this country will learn the hard way I suppose.