What's new
Fantasy Football - Footballguys Forums

This is a sample guest message. Register a free account today to become a member! Once signed in, you'll be able to participate on this site by adding your own topics and posts, as well as connect with other members through your own private inbox!

POLLS/538/Politico (1 Viewer)

SHIZNITTTT

SHIZNITTTT

Footballguy
:lol:   How could so many polls be so wrong?  Hillary was going to win bigly, and in the end "they" all had it wrong by a LARGE margin.  If anything maybe the way polling is done will change.  It looks as if the polling was very lazy or simply just bad.  I didn't see a single poll that had Trump winning, but I saw several polls that showed Hillary winning over 300 electoral votes.   Hint. I guess the days of calling someone on a land line is not real polling.  I can hear the pollsters now, but, but, but, it was the FBI at the end that changed people's votes, or people were to embarrassed to say they were voting for Trump.  Meh, polls are dead no one is going to trust them now. 

 
SHIZNITTTT said:
:lol:   How could so many polls be so wrong?  Hillary was going to win bigly, and in the end "they" all had it wrong by a LARGE margin.  If anything maybe the way polling is done will change.  It looks as if the polling was very lazy or simply just bad.  I didn't see a single poll that had Trump winning, but I saw several polls that showed Hillary winning over 300 electoral votes.   Hint. I guess the days of calling someone on a land line is not real polling.  I can hear the pollsters now, but, but, but, it was the FBI at the end that changed people's votes, or people were to embarrassed to say they were voting for Trump.  Meh, polls are dead no one is going to trust them now. 
Click to expand...
I might be wrong, but I'm pretty sure the single poll that had Brexit correct also was one of a couple that had Trump heading into yesterday. I'd like to see their model. 

 
There's a guy who posted in the Hillary thread that had a lot of money on Hillary winning at a certain odds. He then put what sounded like everything else he had on a separate odds of her winning. I feel bad for that guy and I'm sure he's not too happy about the polls leading up to this election. 

 
How?  Didn't you read the WikiLeaks? Oversampling & media collusion.  they didn't trick all of us.

 
Last edited by a moderator:
  • Despite what the pollsters tell you, it is not science.  Polling is an art filled with assumptions, poor methodology, and biases which you can throw the math out the window.  The margin of error calculation used by pollsters assume you have a known population and that you are able to obtain a perfectly random sample of the entire population.  This assumption is not even remotely close to reality.  The pollsters don't know who is going to vote and who is not going to vote, they make educated guesses.  There is a lot of uncertainty that a person who says they are going to vote for X might actually end up voting for Y.  And there is certainly no way the pollsters are really able to obtain anything close to a random sample.  So when a pollster tells you their poll has a margin of error of +/-3% you can laugh in their face because the truth is, they have no idea what their true margin of error is.    
 
Quez said:
How?  Didn't you read the WikiLeaks? Oversampling & media collusion.  they didn't trick all of us.
Click to expand...


Exactly. Stupid is as stupid does. It's hard to hear anything when you're too busy running your own big mouth. Lots of big mouths around here and in the national punditry.

 
bigmarc27 said:
I might be wrong, but I'm pretty sure the single poll that had Brexit correct also was one of a couple that had Trump heading into yesterday. I'd like to see their model. 
Click to expand...
Their model is only as accurate as their ability to predict who is motivated to vote.  Right now they have their finger on the population better than other pollsters.  I can guarantee you, their model would have completely fallen apart in the Obama 2008 election as the dynamic of who was motivated was completely different.

 
The polls are skewed. Most polls had even R's and D's responding, but they skewed it to a D+8 like Obama had in 2008. They can do whatever they want with the raw data.

 
Turns out the only Poll we should have been following was the USC/LA Times

As of Tuesday morning, the poll's final forecast for the election showed Trump leading by a little over 3 points, 46.8% to 43.6%.

The poll’s findings caused dismay — even outrage — among some readers, especially Democrats, who have denounced it and often criticized The Times for running it.

But just as four years ago the poll was one of the few that did not underestimate President Obama’s support, it seems as though it may have been on its own in not underestimating Trump’s.
Click to expand...
:mellow:

 
bigmarc27 said:
I might be wrong, but I'm pretty sure the single poll that had Brexit correct also was one of a couple that had Trump heading into yesterday. I'd like to see their model. 
Click to expand...
You are wrong. Its a common held belief - but its not true.

@ForecasterEnten - Why do folks continue to argue that most polls in Brexit showed remain winning? During June, Leave led in 17 polls, Remain in 14, 3 ties.

 
whoknew said:
You are wrong. Its a common held belief - but its not true.

@ForecasterEnten - Why do folks continue to argue that most polls in Brexit showed remain winning? During June, Leave led in 17 polls, Remain in 14, 3 ties.
Click to expand...
I just saw it thrown out there a bunch. Wasn't positive, that's why I said as much. 

 
Just thankful the hurricane forecasters did a much better job than the political forecasters this year.

 
Trump ran the Jesse Ventura campaign on a national level.  From the start, it was predicated on driving a different demographic to the polls that are typically unlikely voters.  All the polls utterly missed Ventura's victory as govrenor of Minnesota and they missed Trump winning the White House.

 
Riversco said:
Trump ran the Jesse Ventura campaign on a national level.  From the start, it was predicated on driving a different demographic to the polls that are typically unlikely voters.  All the polls utterly missed Ventura's victory as govrenor of Minnesota and they missed Trump winning the White House.
Click to expand...
VERY :goodposting:  

The only hope is he doesn't drive the USA over the cliff like Ventura did to Minnesota :oldunsure:

 
I'm not sure they were that wrong.  In the last week, the polls really tightened.  I think it was just assumed that there was no way Trump could possibly run the table in a bunch of basic coin flip state elections.

I still can't believe he won all of them.  Just impossible...

 
I know I started to see polling as being more accurate than it turned out to be. Hopefully won't lose sight of that in the future.

 
One of the CNN panelists last night said that the USC/LA Times poll was consistently correct, and the major procedural difference between it and other polls was that they used the same voter pool every time. No trying to pull a different 1,000 likely voters every week.

 
538's model had Trump with a 28.6% chance of winning. That is not an insignificant number. I don't turn off the TV when a guy with a .286 on-base percentage comes to the plate, or walk away from the poker table when I've got 13 outs on the river.

It's human nature to say that 538 "got it wrong" because the candidate with the higher percentage didn't win, but really the model always acknowledged the uncertainty of relying on poll numbers. If their predicted favorite had a 71.4% chance of winning every election, they'd have to be "wrong" 2 out of every 7 times to truly be right.

What's kinda funny is that HuffPo ran this piece 2 days ago blasting 538's methodology because it didn't have Clinton as a 98 or 99 percent favorite like their own numbers did.

 
Yeah, not sure why Silver's the one getting hate in here. He was one of the few poll nerds who insisted on Trump having a real shot. Lots of other places had Hillary at a 98-99% chance.

 
Brony said:
:lmao:

Electoral college idea is so stupid.  Not a democracy if a vote in OH, FL is more important than a vote in TX or CA.
Click to expand...
Thanks to our wonderful educational system you don't even know what type of government we have.  We have a democratic republic, not a straight democracy.

 
Spartans Rule said:
538's model had Trump with a 28.6% chance of winning. That is not an insignificant number. I don't turn off the TV when a guy with a .286 on-base percentage comes to the plate, or walk away from the poker table when I've got 13 outs on the river.

It's human nature to say that 538 "got it wrong" because the candidate with the higher percentage didn't win, but really the model always acknowledged the uncertainty of relying on poll numbers. If their predicted favorite had a 71.4% chance of winning every election, they'd have to be "wrong" 2 out of every 7 times to truly be right.

What's kinda funny is that HuffPo ran this piece 2 days ago blasting 538's methodology because it didn't have Clinton as a 98 or 99 percent favorite like their own numbers did.
Click to expand...
Yeah I'm not sure why people are so quick to #### on Nate Silver all the time.  His numbers are purely poll-based, and his methodology was one of the only ones giving Trump a fighting chance.  If anything we should realize that his model was, yet again, far more accurate than your average pollster.

538 had the Bills as a 31% chance of beating the Seahawks on MNF this week, and the Bills had 4 shots to score a game-winning TD (of course, they couldn't do so).  It's not unreasonable for something with 25-35% odds to hit.

 
538 takes the most heat because they are the most visible pollster. And they were wrong, just not as wrong as the other pollsters.

And, quite frankly, when people try to defend them on the grounds that they weren't as wrong, the optics are pretty bad, especially the day after being proved wrong definitively.

 
bud29 said:
Yeah, not sure why Silver's the one getting hate in here. He was one of the few poll nerds who insisted on Trump having a real shot. Lots of other places had Hillary at a 98-99% chance.
Click to expand...
Agreed, he was actually the closest on the mark and said basically the results were within polling error and you had three possible results (narrow, Trump win, narrow Clinton win or big Clinton win).  Huffpost with its 98% Clinton win and statements that no way she losses was totally off base. 

 
Black Box said:
538 takes the most heat because they are the most visible pollster. And they were wrong, just not as wrong as the other pollsters.

And, quite frankly, when people try to defend them on the grounds that they weren't as wrong, the optics are pretty bad, especially the day after being proved wrong definitively.
Click to expand...
Are you talking about me?  I didn't vote for Clinton.

538's model had Trump showing a 28.6% chance to win going into yesterday.  An underdog but certainly well within the realm of possibility to win.  Silver has said repeatedly that one of the failings of his model is that it views states individually and doesn't correlate across state lines (ie if Trump wins PA, he's probably going to win OH and perhaps MI).  That appears to have been the case for sure last night.

I suppose where you and I differ is that I view 538's prognostications as simply a predictive model, absent of political bias.  Silver himself may have biases, and probably does.  But his model is just that, a predictive model.  It's not perfect and I'm sure he'll tweak it, because that's how predictive models work.  No matter whether Clinton won or Trump won, it doesn't necessarily prove or disprove Silver's analysis, though I would say it's clear that his model was not as accurate in this election as it was in the past.  Models should always be adjusted and tweaked.  You're a pretty logical guy, not sure where the disconnect lies here.

 
Steve Tasker said:
Are you talking about me?  I didn't vote for Clinton.

538's model had Trump showing a 28.6% chance to win going into yesterday.  An underdog but certainly well within the realm of possibility to win.  Silver has said repeatedly that one of the failings of his model is that it views states individually and doesn't correlate across state lines (ie if Trump wins PA, he's probably going to win OH and perhaps MI).  That appears to have been the case for sure last night.

I suppose where you and I differ is that I view 538's prognostications as simply a predictive model, absent of political bias.  Silver himself may have biases, and probably does.  But his model is just that, a predictive model.  It's not perfect and I'm sure he'll tweak it, because that's how predictive models work.  No matter whether Clinton won or Trump won, it doesn't necessarily prove or disprove Silver's analysis, though I would say it's clear that his model was not as accurate in this election as it was in the past.  Models should always be adjusted and tweaked.  You're a pretty logical guy, not sure where the disconnect lies here.
Click to expand...
No, I get that polls will tweak their methodology moving forward, I was just attempting to answer the question why 538 takes so much heat.

 
Riversco said:
Trump ran the Jesse Ventura campaign on a national level.  From the start, it was predicated on driving a different demographic to the polls that are typically unlikely voters.  All the polls utterly missed Ventura's victory as govrenor of Minnesota and they missed Trump winning the White House.
Click to expand...
Only low energy people were polled. I know of no one who was. Low energy follows Hillary around like a shadow.

 
Steve Tasker said:
Are you talking about me?  I didn't vote for Clinton.

538's model had Trump showing a 28.6% chance to win going into yesterday.  An underdog but certainly well within the realm of possibility to win.  Silver has said repeatedly that one of the failings of his model is that it views states individually and doesn't correlate across state lines (ie if Trump wins PA, he's probably going to win OH and perhaps MI).  That appears to have been the case for sure last night.

I suppose where you and I differ is that I view 538's prognostications as simply a predictive model, absent of political bias.  Silver himself may have biases, and probably does.  But his model is just that, a predictive model.  It's not perfect and I'm sure he'll tweak it, because that's how predictive models work.  No matter whether Clinton won or Trump won, it doesn't necessarily prove or disprove Silver's analysis, though I would say it's clear that his model was not as accurate in this election as it was in the past.  Models should always be adjusted and tweaked.  You're a pretty logical guy, not sure where the disconnect lies here.
Click to expand...
i find it strange that he hasn't corrected for that, since he has recognized it. it should be well within his analytic capabilities.

btw, this was the same error that modelers made in assessing the risk of mortgage securities - assuming that the risk of default for a given house was independent of the risk for another house, when in fact, default risk was highly correlated.

 
Brony said:
:lmao:

Electoral college idea is so stupid.  Not a democracy if a vote in OH, FL is more important than a vote in TX or CA.
Click to expand...
Electoral college is way more exciting than a popular vote.  Take last night.  We spent like an hour in Fl.  Then once that become clear, we moved to VA, NC and OH.  Then we moved to WI and MI.  The time zones really help to stagger the excitement as well.  Electoral college and time zones is the winning formula. It would be ideal if a a west coast state was in play but then you're getting into the really late hours and I'm normally in bed by 11 ET.

 
NutterButter said:
Electoral college is way more exciting than a popular vote.  Take last night.  We spent like an hour in Fl.  Then once that become clear, we moved to VA, NC and OH.  Then we moved to WI and MI.  The time zones really help to stagger the excitement as well.  Electoral college and time zones is the winning formula. It would be ideal if a a west coast state was in play but then you're getting into the really late hours and I'm normally in bed by 11 ET.
Click to expand...
Almost like it was designed for TV Drama!

 
You must log in or register to reply here.

Similar threads

Jayrod
The need to be "right" is the real sickness here
2
Replies
87
Views
3K
Nathan R. Jessep
Nathan R. Jessep
GordonGekko
Do You Believe That Effective Modern "Work / Life Balance" Is Even Possible? (19/30)
2
Replies
54
Views
3K
Joe Bryant
Joe Bryant
G
Beatles album review....Part 1 Revolver
Replies
23
Views
950
Guido Merkins
G

Users who are viewing this thread

Total: 1 (members: 0, guests: 1)
Back
Top