Register a SA Forums Account here!
JOINING THE SA FORUMS WILL REMOVE THIS BIG AD, THE ANNOYING UNDERLINED ADS, AND STUPID INTERSTITIAL ADS!!!

You can: log in, read the tech support FAQ, or request your lost password. This dumb message (and those ads) will appear on every screen until you register! Get rid of this crap by registering your own SA Forums Account and joining roughly 150,000 Goons, for the one-time price of $9.95! We charge money because it costs us money per month for bills, and since we don't believe in showing ads to our users, we try to make the money back through forum registrations.
 
  • Locked thread
The Protagonist
Jun 29, 2009

The average is 5.5? I thought it was 4. This is very unsettling.

Xandu posted:

He's still using that garbage polls-plus model on the site though.

Wasn't the problem he ignored his own model?

Adbot
ADBOT LOVES YOU

Vox Nihili
May 28, 2008

Xandu posted:

He's still using that garbage polls-plus model on the site though.

Hey, it's only done slightly worse than the polls-only model!

Typo
Aug 19, 2009

Chernigov Military Aviation Lyceum
The Fighting Slowpokes
haha nate silver's self-criticism struggle session

Xandu
Feb 19, 2006


It's hard to be humble when you're as great as I am.

Vox Nihili posted:

Hey, it's only done slightly worse than the polls-only model!

This is my problem with it. Yeah it's close-ish, but it's also objectively worse than just using the polls and that's been clear for a while, so why not get rid of it or make the model beter?

Both his models also wildly overestimate the chance of the winner winning, which he sort of acknowledges but has done nothing to correct.

Edit:

quote:

Sanders won in Indiana when our “polls-only” forecast gave him just a 15 percent chance and our “polls-plus” forecast gave him only a 10 percent chance. More impressively, he won in Michigan, where both models gave him under a 1 percent chance.

This isn't impressive, it just means your model is lovely. Or more accurately, your inputs are.

stephenfry
Nov 3, 2009

I AM AN IDIOT.
I AM AN IDIOT.
I AM AN IDIOT.
I AM AN IDIOT.
I AM AN IDIOT.
I AM AN IDIOT.
I AM AN IDIOT.
I AM AN IDIOT.

Tayter Swift posted:

Realtalk I've been working in forecasting myself for about six years and the primary thing it's taught me is that Nobody Likes Forecasters

I mean when was the last time you thought 'drat thanks weatherperson' when they nailed the rain forecast to the hour

Don't do forecasting kids

the industry comes with some upsides https://weather.com/news/news/meteorologists-least-creepy-profession

Vox Nihili
May 28, 2008

Xandu posted:

This is my problem with it. Yeah it's close-ish, but it's also objectively worse than just using the polls and that's been clear for a while, so why not get rid of it or make the model beter?

Both his models also wildly overestimate the chance of the winner winning, which he sort of acknowledges but has done nothing to correct.

Edit:


This isn't impressive, it just means your model is lovely. Or more accurately, your inputs are.

Yeah, my post was a joke, much like his plus model.

Mirthless
Mar 27, 2011

by the sex ghost

METEOROLOGISTS LEAST LIKELY PEOPLE TO MAKE YOU INTO A SKIN SUIT - Seriously, just get in the van, you're 100% safe

oystertoadfish
Jun 17, 2003

more like beta danalyst

etalian
Mar 20, 2006

Nate Silver will knock off Trump at the convention so his past prediction will be accurate

Mayor Dave
Feb 20, 2009

Bernie the Snow Clown
What I don't get is why he uses excel at all. He's already said that he uses Stata as part of his analytical toolset, and I'm confused as to what benefit some massive excel macro gives him over any of the excellent statistical software packages available.

etalian
Mar 20, 2006

Mayor Dave posted:

What I don't get is why he uses excel at all. He's already said that he uses Stata as part of his analytical toolset, and I'm confused as to what benefit some massive excel macro gives him over any of the excellent statistical software packages available.

It's because Nate Silver is a waste, Nate Silver is a big fat mess

Ork of Fiction
Jul 22, 2013
Once Trump is ushered into power and National Bullying Corps is implemented, Nate Silver will be getting so many noogies.

etalian
Mar 20, 2006

I hope Predictit has a market guessing how long until Nate Silver goes bald.

R. Guyovich
Dec 25, 1991


this guy is a college friend of mine and i have made so many jokes about the politics side of the site to him while i've been in town

he's cool about it, all things considered

The Brown Menace
Dec 24, 2010

Now comes in all colors.


Concerned Citizen posted:

is this thread now transforming into one of those hate threads in which the target receives irrationally huge amounts of hate and then inevitably ends up actually succeeding?

that's actually every thread about trump other than the official trump thread

ate shit on live tv
Feb 15, 2004

by Azathoth

lol

He's really pathetically self-loathing.

quote:

it’s time to ask whether user scores are truly as objective as they purport to be.
:allears:

ate shit on live tv has issued a correction as of 07:00 on May 19, 2016

Yudo
May 15, 2003

Mayor Dave posted:

What I don't get is why he uses excel at all. He's already said that he uses Stata as part of his analytical toolset, and I'm confused as to what benefit some massive excel macro gives him over any of the excellent statistical software packages available.

Excel is legitimately confusing. Why can't I just type out what I want? SAS, Stata, SPSS, R et al. are not exactly lisp level hard and even a clunky script can be used over and over. Most importantly, clicking on poo poo is for suckers.

The Brown Menace
Dec 24, 2010

Now comes in all colors.


Yudo posted:

Excel is legitimately confusing. Why can't I just type out what I want? SAS, Stata, SPSS, R et al. are not exactly lisp level hard and even a clunky script can be used over and over. Most importantly, clicking on poo poo is for suckers.

*nate silver furiously typing "donald trump not becoming president" into his by this point sentient excel sheet*

jojoinnit
Dec 13, 2010

Strength and speed, that's why you're a special agent.

The Brown Menace posted:

*nate silver furiously typing "donald trump not becoming president" into his by this point sentient excel sheet*

He's just a man getting through life the best he can.

TheGreyGhost
Feb 14, 2012

“Go win the Heimlich Trophy!”

Mayor Dave posted:

What I don't get is why he uses excel at all. He's already said that he uses Stata as part of his analytical toolset, and I'm confused as to what benefit some massive excel macro gives him over any of the excellent statistical software packages available.

Excel is great if you have to tweak your data set before actually running calculations/initially uploading the data to the application because doing that stuff in Stata's data editor is stupid and takes more time. I've worked under a few people that write at the website now, and we will always make sure the initial .csv of data is cleaned up in excel first just to avoid variables importing in the wrong type or other stupid poo poo that's easy to miss. That's literally the only point to using excel in conjunction with any platform like that though. Every quant-focused polisci major gets trained on something like SPSS or Stata these days because it's the best suited for the job, and occasionally you'll get some bastard that teaches R because they sperg about how the graphics look too much. I think of Stata as the best middle ground since it's a little more user friendly than something like R, but way more robust than SPSS.


Yudo posted:

Excel is legitimately confusing. Why can't I just type out what I want? SAS, Stata, SPSS, R et al. are not exactly lisp level hard and even a clunky script can be used over and over. Most importantly, clicking on poo poo is for suckers.

If you use macros/Visual basic for excel, you basically can just like a Stata do file would. It's an astonishingly stupid and brittle program/method, and you have to learn a language/system of references that's dependent on the physical placement of numbers in the sheet, but you can do it.

Now, why you would ever want to use accounting software that's closer to Quickbooks than a stats package is beyond me, but you could do it.

Mayor Dave
Feb 20, 2009

Bernie the Snow Clown

TheGreyGhost posted:

Excel is great if you have to tweak your data set before actually running calculations/initially uploading the data to the application because doing that stuff in Stata's data editor is stupid and takes more time. I've worked under a few people that write at the website now, and we will always make sure the initial .csv of data is cleaned up in excel first just to avoid variables importing in the wrong type or other stupid poo poo that's easy to miss. That's literally the only point to using excel in conjunction with any platform like that though. Every quant-focused polisci major gets trained on something like SPSS or Stata these days because it's the best suited for the job, and occasionally you'll get some bastard that teaches R because they sperg about how the graphics look too much. I think of Stata as the best middle ground since it's a little more user friendly than something like R, but way more robust than SPSS.


If you use macros/Visual basic for excel, you basically can just like a Stata do file would. It's an astonishingly stupid and brittle program/method, and you have to learn a language/system of references that's dependent on the physical placement of numbers in the sheet, but you can do it.

Now, why you would ever want to use accounting software that's closer to Quickbooks than a stats package is beyond me, but you could do it.

Don't get me wrong, I clean my data in excel just like the next guy, but I wouldn't even do summary statistics in excel much less try to make a monster model to invent actual psychohistory. That's why I thought it was weird; when I was learning graduate statistics it only took a semester or two to turn even the least savvy users into people who can do most anything in Stata. The fact that Nate Silver prefers Excel is just weird to me.

Combed Thunderclap
Jan 4, 2011



Nate started using Excel way way back when he was just a baby sabermetrics nerd and he doesn't seem interested in ever fully switching (there's a rant about it somewhere back in this thread already, actually).

My hatred has subsumed significantly since reading his mea culpa article, which is actually really good.

Then I try to listen to the 538 podcast argggggh

Tayter Swift
Nov 18, 2002

Pillbug
Excel owns pivot tables own don't do gigantic forecasts in excel tho namaste

Laphroaig
Feb 6, 2004

Drinking Smoke
Dinosaur Gum

hillary is going on Ellen w/ the cast of the ghostbusters reboot

it is now a relevant political issue to the 2016 presidential election.

nothing matters.

Mia Wasikowska
Oct 7, 2006


what an rear end in a top hat

The Brown Menace
Dec 24, 2010

Now comes in all colors.


lol just lol at nate's hairline beating a tactical retreat before he did

Helical Nightmares
Apr 30, 2009
*waves spreadsheet* Why don't my graphs affect reality? Whyyyyyyy?

The Brown Menace posted:

lol just lol at nate's hairline beating a tactical retreat before he did

:sbahj:

All hair must kneel before the Trump.

Harold Stassen
Jan 24, 2016
Can someone repost the tweet or whatever with the pic of Nate Silver with cuck hair

citybeatnik
Mar 1, 2013

You Are All
WEIRDOS




Combed Thunderclap posted:

Nate started using Excel way way back when he was just a baby sabermetrics nerd and he doesn't seem interested in ever fully switching (there's a rant about it somewhere back in this thread already, actually).

My hatred has subsumed significantly since reading his mea culpa article, which is actually really good.

Then I try to listen to the 538 podcast argggggh

He's been salty on the podcasts for a while now - are you talking about the bit where he just railed against the NYT for a while during it?

point of return
Aug 13, 2011

by exmarx
the parts where delusional Mitt Romney guys got mad at him were funny though

The Brown Menace
Dec 24, 2010

Now comes in all colors.


point of return posted:

the parts where delusional Mitt Romney guys got mad at him were funny though

and now he's the delusional mitt romney guy getting mad

Sancho
Jul 18, 2003

Harold Stassen posted:

Can someone repost the tweet or whatever with the pic of Nate Silver with cuck hair

Just go to Nate Silver's twitter page it's usually in the top 5 replies for like any tweet he does.

etalian
Mar 20, 2006

Trump first executive order will be to give Nate Silver a swirly

Mia Wasikowska
Oct 7, 2006


this article is so bad it ought to negate any goodwill from his psuedo mea culpa

PERPETUAL IDIOT
Sep 12, 2003
Isn't he a climate change denier/"skeptic" as well?

Xandu
Feb 19, 2006


It's hard to be humble when you're as great as I am.

PERPETUAL IDIOT posted:

Isn't he a climate change denier/"skeptic" as well?

Not exactly, he's just out of his element and trying to be clever. His basic argument in his book is that climate models haven't been very accurate and they've predicted the earth would be warmer than it actually is. But he doesn't really bother to account for governments reducing emissions.

583 did publish a pretty widely-panned article on climate change, but they ended up retracting it.

Pomplamoose
Jun 28, 2008

Donald Trump is withholding that sweet, sweet data. http://fivethirtyeight.com/features/trumps-scorning-of-data-may-not-hurt-him-but-itll-hurt-the-gop/

Trumps Baby Hands
Mar 27, 2016

Silent white light filled the world. And the righteous and unrighteous alike were consumed in that holy fire.

the article posted:

Data doesn’t win elections; candidates do.

Voters don't decide primaries; the Parties do.

Mia Wasikowska
Oct 7, 2006

this is a few months old but my god
http://fivethirtyeight.com/features/why-a-former-twitch-employee-has-one-of-the-most-reproduced-faces-ever/?ex_cid=538twitter
:kappa: :frankerz:

Adbot
ADBOT LOVES YOU

Pomplamoose
Jun 28, 2008

  • Locked thread