Go Back  PPRuNe Forums > PPRuNe Social > Jet Blast
Reload this Page >

US Politics Hamsterwheel V3.0

Jet Blast Topics that don't fit the other forums. Rules of Engagement apply.

US Politics Hamsterwheel V3.0

Old 24th Aug 2020, 20:35
  #2381 (permalink)  
 
Join Date: Jan 2008
Location: Denver
Posts: 1,102
Originally Posted by Uncle Fred View Post
Could someone tell me when I should really start paying attention to the polls?

I actually mean this seriously. The election falls on 3 November (IIAC) and we are still in August. Just over two months hence.....
For anyone who has an appreciation for how stastical models and predictions work, we see that the polls in 2016 were not wrong, per se, but rather erroneously interpreted by simpky glomming on to a 80% chance for HRC meant certainty...which is exactly not what it was doing. Something about missing that day in maths when the words possible, probable, and certain were discussed.

I know many here appreciate how to work with statistics and will watch the polls and reports with a measured eye. This then goes back to my first sentence. At what point did they in 2016 start to show what might (and eventually did) happen?

Lots of muck to fly now, but at what point does it narrow sharply and take shape. Heck, Biden could be, at the day, 70 points ahead in every state and I wouldn't bat an eye. 9 to ten week is a long road. Lots of changes ahead.
Actually, I've been mulling making a brief (hah!) Observers' Guide to American Presidential Elections post. You give me a good excuse to do it now. I'll try to make this neutral - but I'm sure our friends on the other side will correct me if I slip up.

Labor Day (first weekend in Sept.) is traditionally when the real head-to-head campaign begins. The conventions are over, the candidates are officially nominated, and have received whatever "bump" those will provide, and voters are back from summer vacations (or getting the crops in) and starting to focus. This year, of course, is vastly abnormal. So that may not be as big a gear-shift as usual. I think the general trends are pretty well set, but there is always the chance of black-swan events, right up until Election Day (although less so these days, with significant "early voting" allowed - some as early as next month). In other words, it will be a count-down, and it becomes a judgement call whether October's polls will be somehow more accurate than August's.

Ignore the national polls. We really run 50 separate State elections for President, each (except for two, Nebraska and Maine**) being winner-take-all. Clinton and Gore won the National vote - to no effect. So look at the State polls, and then add up the total Electoral votes each candidate is likely to acquire.

Statistically, most polls (based on the sample size, usually 600-1200) will have a margin of error of 3-4%. Takes a poll with 2500 respondents to squeeze that down to 2%. If one wants to be cautious (I do) anything less than a 5% lead is an effective tie - from now until Election Day. As our friends in the UK well know, 52-48 polls can easily result in a 48-52 final vote. Of note: "undecideds" or "others" voters do not always split proportionally, so one has to be extra-careful where those are 5% or more. They may or may not decline as we get closer to Election Day.

Even the State polls are just raw data. It is useful to get some analysis from people who don't poll themselves, but apply local and historical knowledge to each State's polls, such as: accuracy of past polls in that State, the mean a State may revert to, how "similar" States are polling, past voter turnout, demographics, and whether a particular poll is "all adults" or "registered voters" or "likely voters." Or even ignore polls and go strictly with sociological data about the voter pool: race, education, age, lifestyle (urban, suburban, rural).

However, as you mention, I tend to ignore the "80% probability" predictions. Those are based on running several thousand "simulated" elections, and for me that is just too much of a derivative value, that loses touch with the humans making their decisions. Probabilities may be necessary evils in quantum mechanics and for bookies - but for me, I want to decide if Schrödinger's Cat (the election) will be alive or dead, not what probabilistic state it occupies before then. BTW I'm still sticking with 50/50 - too close at this point and too many variables.

Turnout is key. Clinton failed to get out or energize her voters in the three States of the "Blue Wall" (Michigan, Pennsylvania, Wisconsin). They mostly did not switch to Trump - they just failed to vote at all. So if you want to play "pundit" and dig deeper into how solid a State's polls are, you need to look at that as well as the raw numbers. Personally I favor polls of "likely voters," although even that is still subject to error.

Pollsters do try to learn from their mistakes. In 2016, as they themselves will tell you, they missed the impending low Democratic turnout in the Blue Wall, and they failed to correctly estimate and include rural and non-college voters. They have tried to re-balance their polling models - whether they succeeded we won't know until after the election.

I prefer to get my data from 270towin.com. They are pretty neutral in gathering and averaging not just polls, but the estimates of the "pundits" or election-scientists who dig deeper and apply additional local knowledge. With links so that anyone can make their own evaluation of the reasoning used. The pundits themselves are on hiatus until after Labor Day, waiting for things to firm up a bit.

First I look at the "concensus electoral map": https://www.270towin.com/maps/consen...l-map-forecast

Then I look at the "pundits' maps" - as well as going on to their own web pages. it is of note that at this point, the conservatives' bête-noire, CNN, is the most cautious or conservative - they aren't assigning Biden a majority yet:
https://www.270towin.com/2020-electi...t-predictions/

Then I look at the State-by-State polling - and apply my own local knowledge of things such as ballot referendums or other races (for the Senate, for example) that may affect or reinforce turnout or produce "coattail effects," to massage whether, say, a 2-point lead for Biden somewhere is "real." For example, I just don't buy that Georgia is ripe to go Democratic this time, or even really on a knife-edge. I'm slightly more confident of Arizona, simply because both Biden and the Democratic Senate nominee are showing leads:

https://www.270towin.com/2020-polls-biden-trump/
_____________________
** personally, I rather like the idea of semi-proportional distribution of electoral votes within a State. (Two votes for the whole state for the state-wide winner, plus individual electoral votes from each Congressional District)

I wouldn't mind seeing red patches in California or Washington or Colorado, giving West Coast or obgraham (for example) more valuable votes, or blue patches in Texas or Georgia or Kansas. But it won't happen, because this is up to each State, and having a juicy, solid block of Electoral votes going entirely to one person with just 50%+1 of the popular vote tends to enhance their power. And, of course, it might make it much, much harder to "call" the Election on the night. But we shouldn't run our elections for the convenience of the pundits.

My salute to both Maine and Nebraska for bucking the majority.

Last edited by pattern_is_full; 24th Aug 2020 at 21:09.
pattern_is_full is online now  
Old 24th Aug 2020, 20:58
  #2382 (permalink)  
 
Join Date: Jan 2008
Location: Denver
Posts: 1,102
Originally Posted by West Coast View Post
Looking for some context, I went into the archives and reviewed US politics 2.0 thread, the months leading up to the 2016 election. The amount of smugness (including you) about an impending Clinton victory was thick. You should go take a look.
Yep - live and learn.
pattern_is_full is online now  
Old 24th Aug 2020, 20:59
  #2383 (permalink)  
 
Join Date: Aug 2005
Location: E.Wash State
Posts: 0
I've no real dispute with your post, Pattern.

Although i don't put that much effort into deciphering what others say, or the polls in general. All I can control is one vote. The rest, including what I write here, is just bloviating.
obgraham is offline  
Old 24th Aug 2020, 21:04
  #2384 (permalink)  
 
Join Date: Apr 2001
Location: surfing, watching for sharks
Posts: 3,839
Originally Posted by pattern_is_full View Post
Yep - live and learn.

It was interesting to see what held the attention of our group 4 years ago, and to see who is no longer with us.

RIP Conpilot
West Coast is offline  
Old 25th Aug 2020, 00:29
  #2385 (permalink)  
 
Join Date: Nov 2013
Location: apogee
Age: 67
Posts: 69
I have a hard time figuring out the current crop of conservative Bible thumpers who support Trump.
It's easy. They are, if nothing else, predictable.

'Don't criticize what I do, do as I say. We are the right right.'
meadowrun is offline  
Old 25th Aug 2020, 00:32
  #2386 (permalink)  
 
Join Date: Jul 2013
Location: Everett, WA
Age: 66
Posts: 3,410
When I was in school nearly 50 years ago, my Social Studies teach gave an excellent explanation of why the polls are often wrong (or at least misleading) - and that polls usually underestimate the Republican candidate. I've seen little over the years to put his explanation into question.
This election (and to an extent the 2016 election) have exaggerated that for one simple reason: The hatred of Trump by many on the left is so strong - and their reactions so virulent - that many people are literally afraid to admit they plan to vote for Trump. Heck, wearing a "MAGA" hat is enough to get you assaulted around here.

As a result I'm taking the poll results with a grain of salt this year. With so much mail-in voting this year, we may not know the result of the only poll that really matters for weeks after the election. We've had near 100% mail-in voting in Washington state for years - close races routinely take a week or more after election day for the results to be known with any certainty.
tdracer is offline  
Old 25th Aug 2020, 01:16
  #2387 (permalink)  
 
Join Date: Aug 2000
Location: Los Angeles, CA USA
Posts: 27
Originally Posted by tdracer View Post
The hatred of Trump by many on the left is so strong - and their reactions so virulent - that many people are literally afraid to admit they plan to vote for Trump. Heck, wearing a "MAGA" hat is enough to get you assaulted around here
Why do you suppose that is so? I would really like to know.

Last edited by Senior Pilot; 25th Aug 2020 at 06:23. Reason: Fix quote insert
Dea Certe is offline  
Old 25th Aug 2020, 01:21
  #2388 (permalink)  
 
Join Date: Jun 2011
Location: The World
Posts: 1,499
Originally Posted by obgraham View Post
As is so often the case, Dre, your explanation of the 2016 polls is at variance with what nearly everyone else understood.
I posted a link from 538, a well known election site. The quote you posted from Pew Research is exactly what I was saying when I stated “(the pollsters) underestimated the percentage of white non college educated voters who turned out, a majority of whom favoured Trump”

And 538.org, allied with the New York Times (another great Republican outlet!) on the morning of the election itself, had Clinton 71% likely to win and stated that other polls were running 85-99% likely Clinton. Their assessment of the Electoral College was 305 to 235 or so for Clinton.
Uh huh. As I said the poll average didn’t take into account the late shift to Trump after the Comey letter. They were still within the margin of error. And Trump snuck over the line by a nose (0.2 in Michigan, 0.7 in WI/PA). 70-30 isn’t that much of a sure thing.

It's a little self serving to come along now, 4 years later and try to tell us that most everyone knew Trump had a good likelihood of winning. It just was not true.
Again you have misconstrued my post. What I said was that it was tighter than most media were reporting, not that pollsters said Trump had a good likelihood of winning.
dr dre is offline  
Old 25th Aug 2020, 02:40
  #2389 (permalink)  
 
Join Date: Aug 2008
Location: Vendee
Posts: 141
Originally Posted by pattern_is_full View Post
Actually, I've been mulling making a brief (hah!) Observers' Guide to American Presidential Elections post. You give me a good excuse to do it now. I'll try to make this neutral - but I'm sure our friends on the other side will correct me if I slip up.

Labor Day (first weekend in Sept.) is traditionally when the real head-to-head campaign begins. The conventions are over, the candidates are officially nominated, and have received whatever "bump" those will provide, and voters are back from summer vacations (or getting the crops in) and starting to focus. This year, of course, is vastly abnormal. So that may not be as big a gear-shift as usual. I think the general trends are pretty well set, but there is always the chance of black-swan events, right up until Election Day (although less so these days, with significant "early voting" allowed - some as early as next month). In other words, it will be a count-down, and it becomes a judgement call whether October's polls will be somehow more accurate than August's.

Ignore the national polls. We really run 50 separate State elections for President, each (except for two, Nebraska and Maine**) being winner-take-all. Clinton and Gore won the National vote - to no effect. So look at the State polls, and then add up the total Electoral votes each candidate is likely to acquire.

Statistically, most polls (based on the sample size, usually 600-1200) will have a margin of error of 3-4%. Takes a poll with 2500 respondents to squeeze that down to 2%. If one wants to be cautious (I do) anything less than a 5% lead is an effective tie - from now until Election Day. As our friends in the UK well know, 52-48 polls can easily result in a 48-52 final vote. Of note: "undecideds" or "others" voters do not always split proportionally, so one has to be extra-careful where those are 5% or more. They may or may not decline as we get closer to Election Day.

Even the State polls are just raw data. It is useful to get some analysis from people who don't poll themselves, but apply local and historical knowledge to each State's polls, such as: accuracy of past polls in that State, the mean a State may revert to, how "similar" States are polling, past voter turnout, demographics, and whether a particular poll is "all adults" or "registered voters" or "likely voters." Or even ignore polls and go strictly with sociological data about the voter pool: race, education, age, lifestyle (urban, suburban, rural).

However, as you mention, I tend to ignore the "80% probability" predictions. Those are based on running several thousand "simulated" elections, and for me that is just too much of a derivative value, that loses touch with the humans making their decisions. Probabilities may be necessary evils in quantum mechanics and for bookies - but for me, I want to decide if Schrödinger's Cat (the election) will be alive or dead, not what probabilistic state it occupies before then. BTW I'm still sticking with 50/50 - too close at this point and too many variables.

Turnout is key. Clinton failed to get out or energize her voters in the three States of the "Blue Wall" (Michigan, Pennsylvania, Wisconsin). They mostly did not switch to Trump - they just failed to vote at all. So if you want to play "pundit" and dig deeper into how solid a State's polls are, you need to look at that as well as the raw numbers. Personally I favor polls of "likely voters," although even that is still subject to error.

Pollsters do try to learn from their mistakes. In 2016, as they themselves will tell you, they missed the impending low Democratic turnout in the Blue Wall, and they failed to correctly estimate and include rural and non-college voters. They have tried to re-balance their polling models - whether they succeeded we won't know until after the election.

I prefer to get my data from 270towin.com. They are pretty neutral in gathering and averaging not just polls, but the estimates of the "pundits" or election-scientists who dig deeper and apply additional local knowledge. With links so that anyone can make their own evaluation of the reasoning used. The pundits themselves are on hiatus until after Labor Day, waiting for things to firm up a bit.

First I look at the "concensus electoral map": https://www.270towin.com/maps/consen...l-map-forecast

Then I look at the "pundits' maps" - as well as going on to their own web pages. it is of note that at this point, the conservatives' bête-noire, CNN, is the most cautious or conservative - they aren't assigning Biden a majority yet:
https://www.270towin.com/2020-electi...t-predictions/

Then I look at the State-by-State polling - and apply my own local knowledge of things such as ballot referendums or other races (for the Senate, for example) that may affect or reinforce turnout or produce "coattail effects," to massage whether, say, a 2-point lead for Biden somewhere is "real." For example, I just don't buy that Georgia is ripe to go Democratic this time, or even really on a knife-edge. I'm slightly more confident of Arizona, simply because both Biden and the Democratic Senate nominee are showing leads:

https://www.270towin.com/2020-polls-biden-trump/
_____________________
** personally, I rather like the idea of semi-proportional distribution of electoral votes within a State. (Two votes for the whole state for the state-wide winner, plus individual electoral votes from each Congressional District)

I wouldn't mind seeing red patches in California or Washington or Colorado, giving West Coast or obgraham (for example) more valuable votes, or blue patches in Texas or Georgia or Kansas. But it won't happen, because this is up to each State, and having a juicy, solid block of Electoral votes going entirely to one person with just 50%+1 of the popular vote tends to enhance their power. And, of course, it might make it much, much harder to "call" the Election on the night. But we shouldn't run our elections for the convenience of the pundits.

My salute to both Maine and Nebraska for bucking the majority.
Thank you Pattern. I really appreciate the thought and effort that went into your reply and it well answers my question. You are indeed spot on about how we in the UK were reminded in the Brexit vote that the obverse of a 52-48 poll lies not far from 48-52.

Again, many thanks!
Uncle Fred is offline  
Old 25th Aug 2020, 02:58
  #2390 (permalink)  
 
Join Date: Aug 2005
Location: E.Wash State
Posts: 0
Here are some numbers about rejected mail-in ballots, just for interest. My state votes completely by mail, or you drop it in a county collection box. Has to be in and postmarked by the vote deadline. That's the law.

The 2 counties mailed out 160,000 ballots for this month's primary election.
49% were returned, a record.
2100 of those were rejected. 1220 for being late, about 800 for not being signed, or signature not varifiable, even after a phone call from the election board.

So looks like roughly 2.3% rejection rate, in a state with 10 years of voting by mail. Make of that what you will.
obgraham is offline  
Old 25th Aug 2020, 05:11
  #2391 (permalink)  
 
Join Date: Jul 2013
Location: Everett, WA
Age: 66
Posts: 3,410
Originally Posted by Dea Certe View Post
Why do you suppose that is so? I would really like to know.
Why is the far left so violent? I've long wondered that myself, but the nature of the "mostly peaceful" protests this summer has shown that there is plenty of violence on the far left.
tdracer is offline  
Old 25th Aug 2020, 11:14
  #2392 (permalink)  
 
Join Date: Apr 2005
Location: RH11 UK
Posts: 106
Originally Posted by obgraham View Post
Here are some numbers about rejected mail-in ballots, just for interest. My state votes completely by mail, or you drop it in a county collection box. Has to be in and postmarked by the vote deadline. That's the law.

The 2 counties mailed out 160,000 ballots for this month's primary election.
49% were returned, a record.
2100 of those were rejected. 1220 for being late, about 800 for not being signed, or signature not varifiable, even after a phone call from the election board.

So looks like roughly 2.3% rejection rate, in a state with 10 years of voting by mail. Make of that what you will.
Would be great to compare this 2.3% with the rejection rate on in-person ballots. Is this better or worse?
Revnetwork is offline  
Old 25th Aug 2020, 14:50
  #2393 (permalink)  
 
Join Date: Mar 2002
Location: Florida
Posts: 0
Originally Posted by tdracer View Post
Why is the far left so violent? I've long wondered that myself, but the nature of the "mostly peaceful" protests this summer has shown that there is plenty of violence on the far left.
It may not be as inclusive as the far left, but only a moment of opportunity for some
lomapaseo is offline  
Old 26th Aug 2020, 09:39
  #2394 (permalink)  
 
Join Date: Jun 2007
Location: Beyond the Blue Horizon
Age: 61
Posts: 896
Lompaso
Who was the right wing nutcase who drove his car into a very peaceful protest, in I think Charlotte the other year. Somewhat right of centre I would suggest. But to be honest you appear to have so many "incidents" college shootings, nightclubs etc over there it is difficult to keep up. Basically if you allow people to purchase semi automatic assault rifles with little checks, and even less with hand guns and sporting rifles what do you expect. You have idiots on both sides who in your country can obtain said weapons with comparative ease, and can then carry out their own warped agenders ad nauseam and nobody over their seems to have the guts to say enough is enough and clamp down. We had our own idiots who carried out a massacre at Hungerford which resulted in the banning of assault weapons here, and then another at Dunblane did for handguns. I speak as someone who shoots (Deer / Game birds) but your gun laws are frankly idiotic and help these extreme causes.

Kind regards
Mr Mac
Mr Mac is offline  
Old 26th Aug 2020, 14:00
  #2395 (permalink)  
 
Join Date: Mar 2002
Location: Florida
Posts: 0
How did we get into gun laws? I thought we were talking about sticks, hammer, rocks and shopping bags.
lomapaseo is offline  
Old 26th Aug 2020, 17:22
  #2396 (permalink)  
 
Join Date: Jun 2007
Location: Beyond the Blue Horizon
Age: 61
Posts: 896
lomapaseo
I thought the two people waving an M16, and what looked like an automatic chromed up pistol were Republicans in St Louis the other week, because did not Donald invite them to speak at his convention ? That was how it is reported over here In that incident, and we are told nobody threw a rock or anything. As for the car hitting and killing the demonstrator in Charlotte, was that not also a peaceful protest. ? The point is there are as many mad right wingers in your country as those on the left, and US left is pretty middle of the road in the rest of the world.
Kind regards
Mr Mac.
Mr Mac is offline  
Old 26th Aug 2020, 18:07
  #2397 (permalink)  
 
Join Date: Oct 2002
Location: West Wiltshire, UK
Age: 69
Posts: 412
Originally Posted by Mr Mac View Post
lomapaseo
I thought the two people waving an M16, and what looked like an automatic chromed up pistol were Republicans in St Louis the other week, because did not Donald invite them to speak at his convention ? That was how it is reported over here In that incident, and we are told nobody threw a rock or anything. As for the car hitting and killing the demonstrator in Charlotte, was that not also a peaceful protest. ? The point is there are as many mad right wingers in your country as those on the left, and US left is pretty middle of the road in the rest of the world.
Kind regards
Mr Mac.

The couple of lawyers who were filmed, standing holding an AR15 and a chrome finish handgun in St Luis, said that they believed they were under threat, because the protesters opened the gates to their private estate and walked down the private access road, taking a shortcut to get to the mayor's residence. The woman showed a scary lack of understanding of gun safety, as every photo I've seen shows her finger on the trigger, rather than outside the trigger guard, whilst the gun was pointed directly at the protesters. I was taught to never point a loaded gun at someone unless you intended to kill them. She could very easily have accidentally fired that gun, and although the situation may have seemed a bit threatening, it seemed clear from the voices that can be heard on some of the video recordings that the protesters weren't threatening them, or even trespassing on their property, they stayed on the road and made it clear they were peaceful.

That couple were, I think lucky. Imagine what might have happened had one of those guns been fired, even accidentally, given the tensions that existed at that time.
VP959 is offline  
Old 26th Aug 2020, 21:15
  #2398 (permalink)  
 
Join Date: Dec 2000
Location: Arlington, Tx. US
Posts: 638
Big surprise! The murderer in Wisconsin is an Alt-Right white teenager stoking violence to further Trump’s agenda.
The Sultan is offline  
Old 26th Aug 2020, 21:36
  #2399 (permalink)  
 
Join Date: Mar 2002
Location: Florida
Posts: 0
Originally Posted by The Sultan View Post
Big surprise! The murderer in Wisconsin is an Alt-Right white teenager stoking violence to further Trump’s agenda.

How in hell do you point at an agenda from Trump as a cause. Alt-rights have their own agenda and take a counter view to BLM. We have laws for this kind of behavior that are supposed to work on both sides. It sounds to me that the law is being applied impartially in this case. However if we allow this to be inflamed beyond the charges than you can expect Trump to step in.to enforce laws. and that does seem to be his agenda
lomapaseo is offline  
Old 26th Aug 2020, 21:38
  #2400 (permalink)  
 
Join Date: Jan 2006
Location: EU
Posts: 487
Originally Posted by The Sultan View Post
Big surprise! The murderer in Wisconsin is an Alt-Right white teenager stoking violence to further Trump’s agenda.
No way! Shurely some fake news right there. Couldn’t possibly have been a sad, dangerous, armed nutter; incited by a sad, dangerous, narcissistic nutter.

Next you‘ll be telling us he was a member of Proud Boys or Boogaloo Bois (but missed the police and killed some real people - in Bois speak).

Nobody doing more healing than The Frump: “4 more years. 4 more years“
Torquetalk is offline  

Thread Tools
Search this Thread

Contact Us - Archive - Advertising - Cookie Policy - Privacy Statement - Terms of Service - Do Not Sell My Personal Information -

Copyright © 2021 MH Sub I, LLC dba Internet Brands. All rights reserved. Use of this site indicates your consent to the Terms of Use.