- My Forums
- Tiger Rant
- LSU Recruiting
- SEC Rant
- Saints Talk
- Pelicans Talk
- More Sports Board
- Fantasy Sports
- Golf Board
- Soccer Board
- O-T Lounge
- Tech Board
- Home/Garden Board
- Outdoor Board
- Health/Fitness Board
- Movie/TV Board
- Book Board
- Music Board
- Political Talk
- Money Talk
- Fark Board
- Gaming Board
- Travel Board
- Food/Drink Board
- Ticket Exchange
- TD Help Board
Customize My Forums- View All Forums
- Show Left Links
- Topic Sort Options
- Trending Topics
- Recent Topics
- Active Topics
Started By
Message
Posted on 10/27/20 at 11:21 am to nc_tiger
quote:Some of these polls don’t understand how reality works.
this place doesn't understand how the field of statistics works.
And if Nate can’t correct for that, I don’t think he understands how statistics work.
Posted on 10/27/20 at 11:21 am to Scruffy
quote:
The guy is such a chode.
Man, I need to start using that insult again. It’s such a good one
Posted on 10/27/20 at 11:21 am to tigerskin
Baked into that "12% chance for Trump" is the fact that Nate Bronze has Biden with a 33% chance to win Texas, and a 67% chance to win Florida.
Biden will not come remotely close to winning Texas, and he won't win Florida either.
Once you turn Texas and Florida red on Nate Bronze's map, I'd imagine Trump's chances of winning jump from about 12% to >40%.
And that's before even mentioning states like North Carolina, where Nate Bronze gives Biden a 65% chance of winning.
Biden will not come remotely close to winning Texas, and he won't win Florida either.
Once you turn Texas and Florida red on Nate Bronze's map, I'd imagine Trump's chances of winning jump from about 12% to >40%.
And that's before even mentioning states like North Carolina, where Nate Bronze gives Biden a 65% chance of winning.
Posted on 10/27/20 at 11:22 am to upgrayedd
It just sounds insulting.
Chode
Old school
Chode
Old school
Posted on 10/27/20 at 11:23 am to LNCHBOX
quote:
Underlined the key part. Now go back to the rest of his models for that election.
on Oct 27 2016, he had Trump with a 17.8% chance. Oct 19 was 12.6%
he adjusted late in the race, just like he is about to do this cycle
Posted on 10/27/20 at 11:23 am to Oates Mustache
to build a statistical model that's based on a set of priors and available polls to predict the probability of election outcomes, not the outcome it self.
Posted on 10/27/20 at 11:24 am to Oates Mustache
quote:
So then why does he exist?
The purpose of the vast majority of pollsters is NOT to gauge public opinion but rather to GUIDE it.
The polls are done to affect a specific outcome:
Currently the polling numbers serve several purposes:
1. demoralize the right to get them to give up and not vote next Tuesday
2. Keep hope up for the left to come and vote next week to make sure they maintain the House
3. IF the POTUS wins a close election, justify the investigation that he must have cheated because the polls had him getting killed
4. If they successfully cheat and win, justify the victory by pointing to the polls showing they were ahead the entire time
Posted on 10/27/20 at 11:25 am to skullhawk
I think part of the issue is he is heavily reliant on the information provided by pollsters. Unfortunately the population of the US has an all time low level of trust in the media and that viewpoint is then manifested in what people tell the media driven polls.
If you don’t trust the media and therefore their pollsters, you disregard or discount their data collection. This isn’t a big deal if the level of distrust is balanced across the population as a whole. Unfortunately it is lopsided and so there is a a major loss of real information tied to the republicans.
There is no such thing as a “shy Trump supporter” they struggle to measure, but instead it’s Republicans who don’t trust the media and therefor either don’t play the game or purposely give false information due to their lack of trust. Silver is stuck because garbage in garbage out modeling-wise. If he uses the data provided, based on his model it will be skewed heavily to Biden because of bad data. If he tries to build in the distrust to the media, there is no real justification how to tweak the values and so he is just pumping in more error of a different type.
Damned if you do, damned if you don’t.
If you don’t trust the media and therefore their pollsters, you disregard or discount their data collection. This isn’t a big deal if the level of distrust is balanced across the population as a whole. Unfortunately it is lopsided and so there is a a major loss of real information tied to the republicans.
There is no such thing as a “shy Trump supporter” they struggle to measure, but instead it’s Republicans who don’t trust the media and therefor either don’t play the game or purposely give false information due to their lack of trust. Silver is stuck because garbage in garbage out modeling-wise. If he uses the data provided, based on his model it will be skewed heavily to Biden because of bad data. If he tries to build in the distrust to the media, there is no real justification how to tweak the values and so he is just pumping in more error of a different type.
Damned if you do, damned if you don’t.
Posted on 10/27/20 at 11:26 am to the808bass
quote:
I’m sure he will reward you for your service.
You’re a smart guy and a good poster, but your anger and sarcasm are a real blind spot
Posted on 10/27/20 at 11:27 am to nc_tiger
quote:
to build a statistical model that's based on a set of priors and available polls to predict the probability of election outcomes, not the outcome it self.
And there's no way to know if his statistical model is worth a crap or not, which is why I don't understand all the fawning.
Posted on 10/27/20 at 11:28 am to laxtonto
Plenty of Trump voters are flat-out lying to pollsters. Especially in a time when cities are being burned by those who oppose him.
Posted on 10/27/20 at 11:30 am to Scruffy
quote:
Some of these polls don’t understand how reality works.
And if Nate can’t correct for that, I don’t think he understands how statistics work.
as stated yesterday, Nate gives yougov a B grade in his algorithm, even though its an internet only, you have to be registered, and they offer free gifts after you participate so many times.
people dont think that criteria is going to slant LARGELY liberal and young?
bad data in, bad data out
Posted on 10/27/20 at 11:32 am to VoxDawg
Fat Thom Yorke with Jason Alexander's 1990 hairline.
ESPN and the MSM give this guy gobs of money for percentages and odds..that are a joke.
Billy Walters, Billy Baxter, Spiro, all the Wiseguys & Sharps, etc..would love it if Nate made the lines instead of Kenny White, Jay Kornegay & Co, CRIS, Pinnacle, Grande, 5Dimes, etc..
Posted on 10/27/20 at 11:35 am to Flats
quote:
And there's no way to know if his statistical model is worth a crap or not, which is why I don't understand all the fawning.
the model literally aced the 2012 election. then in 2016 gave trump a higher chance of winning that anyone else out there (71 hillary to 29 trump). so right now history says it's worth a crap, but we live in different times. I think if Trump wins this year polling in general, and then downstream models like 538, will require a harder self analysis than they claimed they did post-2016
Posted on 10/27/20 at 11:35 am to nc_tiger
quote:
to build a statistical model that's based on a set of priors and available polls to predict the probability of election outcomes, not the outcome it self.
You understand who pays him right? He fits his data into their narratives, 538 are good at cool animation and graphics that make Dems feel good about themselves
But any measure of common sense would tell you the polls they use in the models are flawed or outright corrupted. Over sampling of never trumpers in the “Republicans” Measurables is a hint.
So their probabilities are corrupted, but the CLIENT is happy
Then Nate can say “well the polls were wrong” and walk away
Posted on 10/27/20 at 11:37 am to LuckyTiger
quote:
Probably.
I think he gave Trump an 8% chance against Hillary.
28.6 against Hillary
Posted on 10/27/20 at 11:40 am to bayoubengals88
quote:On Election Day.
28.6 against Hillary
He had Trump down in the teens, if not lower, before Election Day.
Posted on 10/27/20 at 11:42 am to Scruffy
quote:
He had Trump down in the teens, if not lower, before Election Day.
and at another point before Election Day he had the odds near 50/50
Posted on 10/27/20 at 11:43 am to nc_tiger
quote:
the model literally aced the 2012 election. then in 2016 gave trump a higher chance of winning that anyone else out there (71 hillary to 29 trump). so right now history says it's worth a crap,
How do you "ace" an election if all you give are percentages? Since you claim to know how statistics work, you know that 2 iterations of something this complex tell us nothing.
Popular
Back to top
Follow TigerDroppings for LSU Football News