The End is Nigh: US Presidential Election Thread Part XVI

[Oregoropa]

that's better than the toss-off column of spirit cooking

If I'm wrong maybe I should whip up some spirit cooking and drink it. Plenty of available deer blood in my back yard.

Oregoropa thinks stats are for chumps -- can't replace intangibles!

RCP has moved Virginia, Maine, Michigan, New Mexico into the toss-up column today. Not as far fetched as you think. Momentum is on Trump's side

 

[bobsaget77]

Here is my map. Not as bullish as what I posted last week, but I still think we are looking at a 1980 style win for Trump.



Here are states where I think the results will be within 2.5%



Wisconsin (razors edge victory to Trump within 0.2%. Closest of the night)

New Mexico (within 1.5% to Hillary)

Minnesota

Maine

Colorado

Nevada



2.5 - 5%



Michigan

Pennsylvania

New Jersey

Connecticut

Virginia

Delaware

Oregon

Rhode Island

New Hampshire

Florida

That would be a great victory.
I've just been praying for a victory. Don't care if it's 538-0 or 274-264, just win baby. Take down these Soros owned, Devil worshipping, spirit cooking, luciferian lunatics.
#DRAIN THE SWAMP
I'm going out on a limb and say that he loses Florida or NC (not both), and wins Penn, Michigan, New Mexico, Nevada, Colorado.

that's better than the toss-off column of spirit cooking

 

[Mrs. Garrison]

did someone say whole hog?

 

[LemonMelon]

"Miss Piggy"

 

[Axver]

Just stop.

 

[anitram]

Just stop.

It astounds me how blind so many men are to just how many women there are out there of Hillary's generation or close thereto who are going to go to the polls and stick it to every Trump they've ever come across in their lives. And sadly how many more of us there are of the younger generation who have had to deal with much of the same.

Brace yourselves fellas.

 

[Mrs. Garrison]

If I'm wrong maybe I should whip up some spirit cooking and drink it. Plenty of available deer blood in my back yard.

Save it for when the apocalypse hits, which is shortly after that no-fly zone is enforced in Syria, when all the hell breaks loose.

 

[LuckyNumber7]

If I'm wrong maybe I should whip up some spirit cooking and drink it. Plenty of available deer blood in my back yard.







RCP has moved Virginia, Maine, Michigan, New Mexico into the toss-up column today. Not as far fetched as you think. Momentum is on Trump's side

For the ten-bazillionth time, I don't know how many more times I have to say this... RCP POLL AVERAGES HAVE NO STATISTICAL MEANING. You can't average a set of distributions that end up bimodal.

Oh, and do your research. "Trafalgar Group," ie propaganda poll, is the one with Trump up everywhere by massive margins. They cited the Cuban vote as to why Trump has a "solid lock" in Florida. Good one. Not to mention their entire report is done in EXCEL.

Please stop citing RCP. It's not an accurate predictor.

 

[LemonMelon]

Can we talk about how Obummer commanded all the illegals to vote?! The librulz must have seen Oregoropa's map.

 

[Irvine511]

I hope the Trump men in here remember to smile tomorrow night, win or lose.

They're so much prettier when they smile.

 

[LuckyNumber7]

I hope the Trump men in here remember to smile tomorrow night, win or lose.

They're so much prettier when they smile.

I made this for anyone who cites RCP averages as to why their candidate is going to win... a big frowny average of polls.

My Paint skills are on point.

 

[Oregoropa]

Live by the polls, die by the polls. Easy to cite when they are in your favor. I'm looking at the overall trends. Past performance by both candidates versus poll projections. Also much can be said as to where the campaigns are spending their final days (based upon their internal polling)

 

[anitram]

Live by the polls, die by the polls. Easy to cite when they are in your favor. I'm looking at the overall trends. Past performance by both candidates versus poll projections. Also much can be said as to where the campaigns are spending their final days (based upon their internal polling)

If Hillary Clinton wins close to/around 350 EVs will you declare her victory a landslide tomorrow?

 

[Headache in a Suitcase]

Here is my map. Not as bullish as what I posted last week, but I still think we are looking at a 1980 style win for Trump.

Here are states where I think the results will be within 2.5%

Wisconsin (razors edge victory to Trump within 0.2%. Closest of the night)
New Mexico (within 1.5% to Hillary)
Minnesota
Maine
Colorado
Nevada

2.5 - 5%

Michigan
Pennsylvania
New Jersey
Connecticut
Virginia
Delaware
Oregon
Rhode Island
New Hampshire
Florida

Lol

 

[LuckyNumber7]

Live by the polls, die by the polls. Easy to cite when they are in your favor. I'm looking at the overall trends. Past performance by both candidates versus poll projections. Also much can be said as to where the campaigns are spending their final days (based upon their internal polling)

No. NO NO NO. A thousand times NO. This is not how you do statistics. I'm not doing ANYTHING in my favor. You're just citing an improper statistic. This is a fundamental flaw with RealClearPolitics. For the bazillionth time and one, if you average a multivariable (i.e. multiple candidates, in this case) sample with other multivariable samples, what you create is a statistically meaningless number with how you're trying to use it. If five sample says Trump is at 40 and the next five says Trump is at 60, Trump is NOT at 50. No sample has suggested that Trump is at 50. Trump is most likely at 40 or at 60 -- since his results are bimodal, with a gap in between, you can't just average these numbers. At least, not without understanding what the result generates.

As you know, RCP creates a Poll Average vs. Time chart. This DOES adequately describe how the candidates have been performing in polls, over time. i.e. YES, recent polls on average are higher for Trump. That ABSOLUTELY IS NOT THE SAME THING as suggesting Trump is LEADING in the polls. He led in A SINGLE POLL. A single poll is a single indicator.

The only way you can ever blindly average these numbers like RCP does with 100% guarantee is if the poll was a simple "are you voting for Donald Trump? Y/N." Then yes, you could average multiple samples without consequence.

Jesus. You must be the guy who makes EXCEL produce 8th degree polynomial fits just because it makes your r-squared value perfect.

 

[Irvine511]

So Trump is in MN telling people how horrible the Somali immigrants are and what a threat they are.

What a horrible, racist person.

 

[PhilsFan]

Here is my map. Not as bullish as what I posted last week, but I still think we are looking at a 1980 style win for Trump.

Here are states where I think the results will be within 2.5%

Wisconsin (razors edge victory to Trump within 0.2%. Closest of the night)
New Mexico (within 1.5% to Hillary)
Minnesota
Maine
Colorado
Nevada

2.5 - 5%

Michigan
Pennsylvania
New Jersey
Connecticut
Virginia
Delaware
Oregon
Rhode Island
New Hampshire
Florida

The commonwealth is not going for Trump, that I can assure you.

 

[Oregoropa]

The commonwealth is not going for Trump, that I can assure you.

My part of the commonwealth is going for Trump by 20 points.

Everybody that wants to harp on my map, put up your own. The big horserace is tomorrow. Let's put our cards on the table.

 

[BVS]

Here is my map. Not as bullish as what I posted last week, but I still think we are looking at a 1980 style win for Trump.



Here are states where I think the results will be within 2.5%



Wisconsin (razors edge victory to Trump within 0.2%. Closest of the night)

New Mexico (within 1.5% to Hillary)

Minnesota

Maine

Colorado

Nevada



2.5 - 5%



Michigan

Pennsylvania

New Jersey

Connecticut

Virginia

Delaware

Oregon

Rhode Island

New Hampshire

Florida

I'm quoting this just in case you're here on Wednesday, I'd like to ask what you were thinking.


Sent from my iPhone using U2 Interference

 

[Axver]

LN7, I take it you're enjoying the spat between Nate Silver and that guy at HuffPo?

 

[LuckyNumber7]

LN7, I take it you're enjoying the spat between Nate Silver and that guy at HuffPo?

I tend not to follow. But I did just do a brief read up since you mentioned it.

I do not pretend to be a pollster, nor do I pretend to have the full experience of someone who runs prediction models. But I do have a firm grasp of fundamental statistics. So, based upon what little I read about what you're talking about... yeah, whoever thinks data exists out there that has Clinton winning at 98% chance is biased and has toyed with statistics. He's 100% right about how you use numbers... you can't just produce something and assume that because it mathematically works out, it physically has a meaning.

Case in point... I used to work in a microgravity physics laboratory. We would take high speed videos of particles colliding in near zero-G environments, and have a computer code track these particles positions per camera frame. The result is a series of x-y points. In order to determine how fast that particle might be moving, you can stick a regression line on it and statistically, with some certainty, decipher the velocity. So what kind of regression line do you use? Linear or quadratic? Well, it's in a zero-G environment. So, no gravity. Simple physics says that particles moving without a force on them should be... linear! If the quadratic regression line statistically fits better, that doesn't mean it's correct. Chances are, it was poorly tracked by the computer, or perhaps the particle bounced somewhere. You need to go look at the video. You can't happily just stick a quadratic fit on something that is physically not quadratic. You bastardize the purpose of statistics when you do that.

The same can be said for anyone lazily reporting biased numbers. I have minimal opinion of Nate Silver. I legitimately know very little about him. But judging his rant on Twitter, I know where he's coming from.

 

[Oregoropa]

I'm quoting this just in case you're here on Wednesday, I'd like to ask what you were thinking.


Sent from my iPhone using U2 Interference

He's peaking at the moment of decision

 

[LuckyNumber7]

He's peaking at the moment of decision

Literally NOTHING suggests this at all.

 

[kiwilad]

LuckyNumber7, in your opinion, which appears a hell of a lot better educated than mine, which of the polls should I/we be looking at with the most faith?

 

[bono_212]

Our local elections are also tomorrow, and I'll be voting for the very first time. Wish I was living in the U.S. so I could cast my vote for Hillary.

I thought Puerto Rico votes for the president?

Lol

I can't even.

 

[trojanchick99]

ANN Coulter remains a pathetic human being who is also fucking dumb.


[TWEET]795833821501460480[/TWEET]

 

[bono_212]

 

[DaveC]

No. NO NO NO. A thousand times NO. This is not how you do statistics. I'm not doing ANYTHING in my favor. You're just citing an improper statistic. This is a fundamental flaw with RealClearPolitics. For the bazillionth time and one, if you average a multivariable (i.e. multiple candidates, in this case) sample with other multivariable samples, what you create is a statistically meaningless number with how you're trying to use it. If five sample says Trump is at 40 and the next five says Trump is at 60, Trump is NOT at 50. No sample has suggested that Trump is at 50. Trump is most likely at 40 or at 60 -- since his results are bimodal, with a gap in between, you can't just average these numbers. At least, not without understanding what the result generates.

As you know, RCP creates a Poll Average vs. Time chart. This DOES adequately describe how the candidates have been performing in polls, over time. i.e. YES, recent polls on average are higher for Trump. That ABSOLUTELY IS NOT THE SAME THING as suggesting Trump is LEADING in the polls. He led in A SINGLE POLL. A single poll is a single indicator.

The only way you can ever blindly average these numbers like RCP does with 100% guarantee is if the poll was a simple "are you voting for Donald Trump? Y/N." Then yes, you could average multiple samples without consequence.

Jesus. You must be the guy who makes EXCEL produce 8th degree polynomial fits just because it makes your r-squared value perfect.

 

[prbiker15]

I thought Puerto Rico votes for the president?

Nope, we can only participate in the primaries.

[TWEET]795833821501460480[/TWEET]

How stupid of her.

 

[DaveC]

here's my map from a few days ago: http://www.u2interference.com/forum...ion-thread-part-xv-220364-45.html#post8115302

i'm still pretty happy with it, although i wish i had flipped ohio and florida and made it 322-216.