Presidential Prediction...Our own Crystal Ball

[ukpumacat]

The early voting numbers are up primarily because more people are being encouraged to vote earlier due to COVID. A lot of locales include mail in ballots in those early vote numbers.

Good point. I still think voter turnout will be huge. But, you are right, early voting doesn't necessarily prove that (although it would if they were down with Covid).

[ukpumacat]

You put a lot of faith in polls that use a model with an 18% lower GOP turnout than actually voted in 2016.

I thought you might enjoy this article. I found it interesting. The biggest change pollsters made after 2016 was the weight they give educated voters. They (educated) tend to be Dems and they tend to be more likely to answer polls. So now, these state polls weight more for that. Its why most of them were so much more accurate in 2018.

In short, if 70% of those they are polling are college graduates and 30% are not....they weight the answer of the 30% far more. Because only 30% of actual voters are actually college educated. Clinton won "educated voters" by huge margins but Trump turned out "non-educated voters" in record numbers. Now, they are expecting that again.

P.S. None of this is post is in any way a hidden slight of any Republican etc. I am simply using terms they use with these polls. The bottom line is that Trump wins big in rural areas where there are less college educated voters and they are less likely to answer a poll.

https://www.wpr.org/polls-missed-mar...different-2020

[KeithKSR]

I have a number of pro-union relatives in the Midwest. The vast majority will vote for Trump. That used to be a stabile Democratic demographic, and many have been registered Democrats for many decades. There is a realignment of the support of the parties, and this makes polling much harder.

Populists candidates seem to always outperform polls, and it isn’t just happening in the United States.

Demographics are also shifting. Georgia and North Carolina are becoming more purple as people flee the Northeast. The Midwest becomes more purple as the Dems move away from their middle class base. Texas is becoming more purple as people continue to flee California.

[KeithKSR]

IBD/TIPP has closed dramatically. https://nypost.com/2020/10/20/poll-t...ap-with-biden/

[ukpumacat]

Demographics are also shifting. Georgia and North Carolina are becoming more purple as people flee the Northeast. The Midwest becomes more purple as the Dems move away from their middle class base. Texas is becoming more purple as people continue to flee California.

This isn't factual. I have seen this exact narrative on Fox News countless times (I am not saying you saw it there) and its simply fodder.

There are raw numbers to check this.
We do not have 2019 numbers yet (they are usually released in November) but in 2018 86,164 Californians moved to Texas. But, 37,810 Texans moved to California.
So, a net of 48,354 for Texas.
Of course, this number includes children. And those who are adults...not all are Democrats. In fact, if they are moving to Texas many are probably not (I have only had 2 friends move to Texas from here and both are conservatives). But let's just pretend ALL 48,354 of the net gain are liberal adults.

8,969,226 people voted in Texas in 2016. It will be more in 2020. It will take about 100 years of "Californians fleeing" to make a dent.
Texas is turning purple for 2 reasons (and neither have anything to do with California):
1. Young People
2. Hispanics

Its the same reason every red state that is turning purple is turning purple. If Republicans do not do a better job at winning this demographic, they will lose more and more states in the future. Its that simple.

Btw, on a side note...California (and #2 Texas and #3 Florida) is always going to have a huge number of people moving from it. Its a HUGE state with a ton of people. It is true that in the last few years, more are moving away than are moving to.
But, that is to be expected...its expensive! But tons are still moving to. In fact, in 2018 (the last full year we have numbers for), California was #3 on states people moved TO.

[bigsky]

Gallatin County Montana is getting
Bluer as the people flee the states they’ve already ruined.

[catmanjack]

Just don’t get that, just like all the immigrants from other countries that come to America for a fresh start and bring the same things with them.

[blueboss]

I’m starting to be very worried that Biden is going to pull this off. Meaning Kamela will be the next POTUS.


Sent from my iPhone using Tapatalk

[ukpumacat]

Just don’t get that, just like all the immigrants from other countries that come to America for a fresh start and bring the same things with them.

Good. Lord.

[catmanjack]

Yes I see people like your Democratic Omar ruining America!!! So stop with the good lord crap!

[bigsky]

I’m starting to be very worried that Biden is going to pull this off. Meaning Kamela will be the next POTUS.


Sent from my iPhone using Tapatalk

You should be!

[ukpumacat]

I’ve said before on here and I’ll say it again: my only hope is that whoever wins it’s by a lot. Enough to be clear.
And that means even if that’s Trump.
America is strong but a divided election in which we wait months to find out who won or one that is decided in the Supreme Court would lead to utter and total chaos. I care about who wins for numerous reasons. But I care about that more.
Despite the “end of the world” predictions which comes every 4 years America will survive and be strong no matter who wins this race. Politics are cyclical and usually work on a pendulum. It’s one of the things that keeps her balanced.
But we are a divided nation right now. More than in decades. And a contested election on top of where we are now would simply be catastrophic.
As a Biden supporter, I’m hoping he pulls off Florida. If he wins it, not only does that mean he will won but it means he will win big and we will know that night (as Florida counts their mail in votes early). I would guess Trump pulls Florida out but my fingers are crossed on it.

If Trump wins, let him do it big. Make it clear. Just...no contested election. Just no. On either side.

P.s. I know many of us think this election will be close and that neither will win big. BUT....in 2016, Trump won big. Electorally. Not like a landslide but big enough it was clear.
And that race was decided by 3 counties in 3 states (Michigan, Wisconsin and Pennsylvania) with a total of 77,000 people in them. And he lost the popular vote.
So even a very very close election can have a decisive winner. That’s all I mean above.

[ukpumacat]

Trafalgar is interesting in the polling world. They were closer to the final tally in a few of the swing states in 2016 and the Florida governor race in 2018. But they have been pretty awful on many Other races.
They openly admit they add a “secret Trump voter” bump in their polls. They call it the Social desirability bias.
So, depending on how strong that is...will depend on how correct they call a race. They always lean Republican in their polls because of it.
But here is the big caveat: they will always be closer in races where the late undecideds break toward the Republican (because these people cannot be polled) and they will be way off when the race breaks for the Democrat late.
This is why the 538 model is so good. They do not poll. They simply rate pollsters based on past success and then average all polls (including Trafalgar).

More here on Trafalgar which makes my point I quoted. They just released a few more Trump friendly polls (as I said before, they always lean Republican so if a race goes that way they tend to be more accurate and when they go Dem they tend to be very wrong).

In the polls they released....here is the breakdown of their numbers:

Attachment 9754

If you look closely, they have Biden polling/winning 25% of Republican votes. And they have Trump polling/winning 28% of Dem voters. This is just asinine. They also have Trump winning 50% of hispanics (haha) surveyed and 30% of Black voters.

Again, this is why the quality of the pollsters over time matters. The most respected pollsters over time is consistently a handful of very good ones: Fox News, ABC, Sienna College, Marist, Selzer, Wall Street Journal, Emerson College, etc.

[bigsky]

Feels more landslidey every day. You see Trump minions dismissing Biden’s record fundraising or poll results and it sounds like whistling in the dark.

[KSRBEvans]

If you want to start playing with the electoral map, if you start with the 2016 electoral map, Trump can lose both Michigan and Pennsylvania and still win:

Capture 10-26 20.jpg

But there's no margin for error. IA, WI, Arizona, NC, even one NE congressional district gets him below 270.

[VirginiaCat]

Ok, as far as the changing state affiliations, the movement from Northeast Urban sprawl is real.

Virgina was a very red state up to the 1980's.. Wilder started that change with is W.

I would call VA a more blue than red state from a National Election perspective because 3-4 counties have populations that are about 40% of total state population. They also surround DC and are VERY BLUE. Even In state politics tinge blue now. If I did not have to live here I would leave.

NC was trending that way because the Triangle is VERY BLUE with is liberal universities. GA is all about Atlanta population.

We are becomming a nation of city states. THe urban centers population override the full state demographic choices..... It is a HUGE problem because there is not an electoral college to proctect at the state level.

[ukpumacat]

If you want to start playing with the electoral map, if you start with the 2016 electoral map, Trump can lose both Michigan and Pennsylvania and still win:

Capture 10-26 20.jpg

But there's no margin for error. IA, WI, Arizona, NC, even one NE congressional district gets him below 270.

Yes, of course. There are actual ways to get there. Just, unrealistic. Trump has basically conceded Wisconsin as he has pulled almost all ads out of there.
His smallish path is to hold on to the other 2016 states and win Michigan or Pa. PA is likely his best shot which is why I think its the state most likely to decide the election.

[ukpumacat]

Feels more landslidey every day. You see Trump minions dismissing Biden’s record fundraising or poll results and it sounds like whistling in the dark.

That seems to be the best bet....odds wise. Trump had a few very good (and reliable) polls out of Texas this morning. I never thought Biden could win there but he had some hope late last week as a few reliable polls showed him tied there.

[ukpumacat]

That seems to be the best bet....odds wise. Trump had a few very good (and reliable) polls out of Texas this morning. I never thought Biden could win there but he had some hope late last week as a few reliable polls showed him tied there.

And just like that Biden just did another big TV buy in Texas. And now Kamala is going there.

This is sort of where we are in this whole thing. An update:

For Trump to win, he has to defy the odds and polls (again). Its worse this year though.

At this time last election (and this doesn't get talked about enough), Hillary's numbers were collapsing.
Nationally, she had gone from up +9 to +3 in the last week. And the final LA Times/USC poll had Trump +3. This NEVER gets talked about. But the National polls were not wrong...they were very accurate. Her numbers were flailing. The Comet letter killed her and late deciders went to the "outsider and unknown" Trump.

So, why didn't more pundits have Trump winning then? The polls were not wrong, the analysis was. The issue with the polls were with the state polls. This mostly happened because many of the most reliable pollsters just simply did not do state polls. So, the ones we had were not as accurate AND they simply had a wrong formula (which I wrote about above and these were changed after 2016). Since, the most reliable pollsters have been very accurate (and killed it in 2018).

Back to this year: Biden's numbers are stable. He's not flailing like Clinton did. You have the outlier Rasmussen national poll from yesterday (which today is already back to Biden +2). I could write plenty more about Rasmussen but won't for now. Just to say he was very off in 2018 and he doesn't release his methodology which is always strange. But all polls have margins of errors. So the important thing is to take an average. As of today, Biden is at +9.1. There are a lot of +6 and +7's in there so it seems with including the margin of error, that he is anywhere between +5 and +11 nationally.

Of course, we are not a true democracy and we elect based on the electoral college (although common sentiment is that its impossible to lose the electoral college if you win the popular vote by 5 or more). So, state polls matter.

In 2016, Clinton's numbers were falling in Pennsylvania and Michigan (and she lost). Wisconsin was the state that state polls missed the most. Pennsylvania and Michigan had the race within the margin of error so those were not as big of shocks.

Today, Biden has pretty solid state poll leads in each of these. The race has tightened some in Pennsylvania. Biden is +5 there now. That seems to be where it is likely to finish as the final polls are mostly in and there will be a few more tomorrow. The debate seems to have done nothing as the latest polls (post debate) there are mostly in the +6 or +7 range.
In Michigan and Wisconsin, Biden is +8.

The truest toss up states are Georgia, Florida, Iowa and North Carolina.
The last 5 polls in Georgia all have the race tied or Biden leading.
North Carolina has had Biden +2 for a few weeks now. The only two polls that have had Trump with any lead there are Trafalgar and Rasmussen.
Iowa is Biden +1. The last few best polls have Biden winning there but its very close (all within the margin of error).
Florida is anyone's guess. Right now, polls have Biden +2. But, there has been a back and forth in Florida for some time.

Other states:
Arizona is within the margin of error. Biden is +3 there. However, I would say the most reliable polls there have been pretty consistent for Biden.

Texas is Trump +1. Sienna College/NY Times is one of the best pollsters out there and it has Trump +4. If not for that poll, Biden would actually have a slight lead there. The last several polls have had him tied or winning. I suspect Trump wins there still.

Nevada is Biden +7.

New Hampshire is Biden +11

Nebraska 2 (not the state but the smaller District) is Biden +7. This matters. For Trump to win, he has to hold Florida, Georgia, Iowa and North Carolina. And he has to win Pennsylvania. But, if Biden wins Nebraska 2 (1 electoral vote) that would give Biden the Presidency. So Trump would have to win Pennsylvania AND Arizona (or Michigan).

In summary:

I am as fascinated by the polls as anytime in history. If they are wrong AGAIN, some of these companies will go out of business. A lot of these forecasts (and Vegas oddsmakers) are starting to use the early voting numbers. They are historic.

An almost final thought: there is a part of me that wonders if Trump's whole "you can't trust mail in ballots" has backfired. I know some of the early voting numbers are about Covid. But I don't think that's the whole story. I think A LOT of it is people on both sides of the aisle that are fired up about the election. And I also think so many people do not trust the mail right now and want to make sure they go and vote early. Dems especially are super suspicious that Trump will do something slimy. I hear it a lot. So they are coming out in record numbers to get their vote in now. I genuinely think all of his talk on it has backfired and he has driven people out early to get there vote in.

And a final thought: Most states do not start counting early votes until election day. So, the earliest numbers that come in will be Election Day, in person numbers only. As the night goes on, the early voting numbers will start to come in (absentee and in person both).
I do not envy news desks that have to report on election results. Its going to be very difficult this year because of Covid and we will very possibly not know who won on election night. If Biden wins Florida, its over and we will know that night (even if they haven't called the race).

[bigsky]

All day long the absentee votes can be counted (in some states). So the first reports will be mail in. In Montana. Can’t speak for other states

[ukpumacat]

All day long the absentee votes can be counted (in some states). So the first reports will be mail in. In Montana. Can’t speak for other states

Ya, every state is different. Here is a great article listing how many of the top "toss up" states do theirs:

https://www.washingtonpost.com/polit...ction-results/

Bottom line: We will likely know Florida, Minnesota and Arizona on election night.

Michigan is sort of a hybrid...larger counties can start counting early ballots before Election Day. Smaller counties cannot.

Ohio, Wisconsin, North Carolina and Pennsylvania will take days. They do not start counting until Election Day.

Some states begin counting before Election Day, some on Election Day, and some cannot start until all polling places are closed.

So, if Biden wins Florida, he is going to be the President and we will have a great idea of that on election night. If Trump wins Florida, we just don't know anything. If Trump wins Florida and Arizona, it looks better for him but still don't know. If he wins Minnesota also on election night, he is winning.

For the other states, we will likely just not know for days.

[ukpumacat]

These early voting (early voting = mail in/absentee and in person) numbers are simply staggering:

This is the current early vote turnout as a percentage of the total 2016 turnout.

Texas: 91.2%
Montana: 80.7%
Georgia: 76.8%
North Carolina: 76.1%
Arizona: 74.4%
Colorado: 72.8%
Florida: 72.2%

Two thoughts:

1. We are not only going to have a record turnout, but its going to be by a lot (I think this benefits Biden, but we will see).
2. Again, a reminder that in many states, these votes will not be tallied/counted until Election Day. In some, they do not start counting them until after polls close. And in Alaska, they do not start until a week after Election Day.
The point: We will very likely not know the winner until days later (this is also why Trump continues to say the votes should not count after Election Day. There is a huge Dem advantage in how many are voting early). We will need to be very patient. And early numbers will likely be very skewed.

But just wow how many people have already voted.

[ukpumacat]

I will say one good thing is that if we can all ignore the noise on election night (tweets, pundits etc), the election desks at the major networks are very very good at what they do (Fox, CNN, NBC, CBS). All of them have a slew of experts that are the best in the business at numbers and analysis.
All of them will be reminding us exactly which states do what, etc. And when the network calls a state for a candidate, they are putting their entire reputation on the line. I have no doubt all of us are skeptical of the bias different networks have for different candidates...but those decision desks DO care about getting it right first and foremost.

Here is a quick walk back in history in 2012 when Karl Rove called out the Fox News Decision Desk for naming Obama the winner of Ohio. The Fox decision desk stuck to their guns and they were right of course (he ended up winning by 3 points). https://www.theatlantic.com/politics...n-desk/321670/

[VirginiaCat]

Ok, 5 days. My prediction.

Trump Loses AZ, MN, and MI

Trump Wins WI, PA, OH, FL, GA, NC

[ukpumacat]

Ok, 5 days. My prediction.

Trump Loses AZ, MN, and MI

Trump Wins WI, PA, OH, FL, GA, NC

Fair enough. Just an fyi...Wisconsin is an extreme long shot (which Trump knows because he pulled out much of his ad buys there this week). The last several polls have Biden up pretty big there. Even Trafalgar has Trump behind.

Let's just say, hypothetically, that what you say above happens...Trump would win. If all of that happens and Biden wins Wisconsin (which I think he does), Biden would win. Meaning, Trump can win PA, but if he loses Arizona and NE2, he loses.

[TRUCKERCATFAN]

Trump can lose MI and WI. As long as he holds PA and the other generally red toss up states, he will win. I really hope so. I despise ever Democrat politician in our government.

[ukpumacat]

Trump can lose MI and WI. As long as he holds PA and the other generally red toss up states, he will win. I really hope so. I despise ever Democrat politician in our government.

I am not sure if you are considering Arizona a "red toss up state" or not. But, Trump can win Pennsylvania and win the "red states" and still lose. If he loses AZ and NE2 (as I said above) he will lose 270-268.
Fwiw, Arizona was one state where the state polls were right on in 2016.

[ukpumacat]

I came across this from 4 years ago. As I've said, some of the state polls in 2016 were wrong but the National polls were very good. Check out this ABC News/Washington Post Poll (one of the best) tracker.
As you can see, Clinton's lead flat sunk in the last 10 days (Comey letter). They ended up exactly right, Clinton by 2.

Attachment 9761

P.S. They have focused their polling more on state polls this year. Their last National poll was on Oct 10th where they had Biden +11.

[KeithKSR]

Yes, of course. There are actual ways to get there. Just, unrealistic. Trump has basically conceded Wisconsin as he has pulled almost all ads out of there.
His smallish path is to hold on to the other 2016 states and win Michigan or Pa. PA is likely his best shot which is why I think its the state most likely to decide the election.

Ads weren’t pulled in Wisconsin. The RNC is running the ads there as part of the combined ad strategy.

[bigsky]

Florida decides the election early if it goes blue. Georgia, North Carolina and Texas l think the same. If those hold for Trump then we riot until Pennsylvania come in