home studios equipment staff/friends booking/rates for sale forum contact

Dude: Nate Silver

Vote and debate.

Moderators: kerble, Electrical-Staff

Dude: Nate Silver

Person who renders all pundits, talking heads, and self-proclaimed experts useless
32
71%
NEEEEEEEEEEERDD!!!!
13
29%
 
Total votes : 45

Re: Dude: Nate Silver

Postby Wood Goblin on Tue Oct 30, 2012 1:12 pm

P.J. Craven wrote:And if it doesn't matter to you that Silver was dramatically off the mark in the UK in '10 when the more "ordinary" pundits/bone rollers had it correct, then I don't know what to tell ya.


It doesn't matter to me all. Bill James might be brilliant, but nobody expects him to crack the sport of cricket.
User avatar
Wood Goblin
Heaven-Sent Hero
Heaven-Sent Hero
 
Posts: 6065
Joined: Tue Jun 06, 2006 12:43 pm
Location: South Loop, Chicago

Re: Dude: Nate Silver

Postby P.J. Craven on Tue Oct 30, 2012 1:25 pm

I just think, if this subject interests you, that it would be wise to look at a lot more information than the information neatly packaged by Nate Silver. His is a well-researched perspective, but another perspective nonetheless.
Last edited by P.J. Craven on Tue Oct 30, 2012 1:26 pm, edited 1 time in total.
ty-lot wrote:Sick bro
P.J. Craven
Humankind's Greatest Musical Genius
Humankind's Greatest Musical Genius
 
Posts: 4049
Joined: Tue Sep 14, 2010 10:55 am
Location: sedentary af

Re: Dude: Nate Silver

Postby whoisalhedges on Tue Oct 30, 2012 1:26 pm

P.J. Craven wrote:He's got Romney winning FL by about 65%-35% from a probabilistic perspective. I will be interested to see if that prediction proves correct. Also interesting is that early voting counts in North Carolina currently have Obama up by 281,000 votes or so. And Silver has Romney as an 85% (approximately) favorite there. Not saying he's going to be wrong, I am just interested.

This means that Romney wins FL in 65% of simulations given current data. That means Obama wins 35% of the time. If Obama wins Florida, that does not make the model wrong - it means the model was wrong (in this case) 65% of the time, and right 35% of the time.
"Country music must go on!" -- Austin J. Hall, September 11, 2001
User avatar
whoisalhedges
Eternal Bosom of Hot Love
Eternal Bosom of Hot Love
 
Posts: 3387
Joined: Mon Sep 26, 2011 2:27 pm
Location: Where they put chili on spaghetti, BBQ sauce on kielbasa, and men on the moon.

Re: Dude: Nate Silver

Postby P.J. Craven on Tue Oct 30, 2012 1:27 pm

whoisalhedges wrote:
P.J. Craven wrote:He's got Romney winning FL by about 65%-35% from a probabilistic perspective. I will be interested to see if that prediction proves correct. Also interesting is that early voting counts in North Carolina currently have Obama up by 281,000 votes or so. And Silver has Romney as an 85% (approximately) favorite there. Not saying he's going to be wrong, I am just interested.

This means that Romney wins FL in 65% of simulations given current data. That means Obama wins 35% of the time. If Obama wins Florida, that does not make the model wrong - it means the model was wrong (in this case) 65% of the time, and right 35% of the time.


God, I'd never thought of it that way before.
ty-lot wrote:Sick bro
P.J. Craven
Humankind's Greatest Musical Genius
Humankind's Greatest Musical Genius
 
Posts: 4049
Joined: Tue Sep 14, 2010 10:55 am
Location: sedentary af

Re: Dude: Nate Silver

Postby MWilke on Tue Oct 30, 2012 1:34 pm

P.J. Craven wrote:One guy's methodology is somehow more "rigorous" than the other. Does it really matter when they've both had it very wrong in equal measures?


Yes it does. You're looking at this all wrong. I'm not looking for a prophet. I'm looking at a forecast. You have to understand the difference. Everything you've posted indicates you don't see the distinction. You're predicting uncertain outcomes dependent on several factors of aggregate human psychology. That effectively means no one knows for sure. But if one has a more rigorous thought process over another guy, I'm still going to pay attention to the harder worker in the end.


P.J. Craven wrote:He's got Romney winning FL by about 65%-35% from a probabilistic perspective. I will be interested to see if that prediction proves correct.


Do you play poker?

Think of it this way:

Romney and Obama are playing poker in Florida. Romney flopped top 2 pair. Obama has a flush draw.

EV.com's analysis: Romney's winning this hand.

Silver: Romney's got a 65% win probability. We'll see what happens in the events to come, but unless something really steers Obama's way, Romney has good odds here.


That's the difference. I prefer the more thorough analysis.
User avatar
MWilke
Lode Star of the Twenty-First Century
Lode Star of the Twenty-First Century
 
Posts: 2269
Joined: Mon May 01, 2006 2:48 pm
Location: West Burbs

Re: Dude: Nate Silver

Postby cmay on Tue Oct 30, 2012 1:45 pm

P.J. Craven wrote:When he gets something dramatically wrong, it doesn't matter to you that Nate Silver was wrong?


Pretty sure what the others are pointing to is language like above. "Wrong"/"right" is not the correct way to interpret his predictions. He gives probabilities of outcomes. The only way that he could possibly be "wrong" is if he were to say that he's 100% certain that an outcome will occur. Which, of course, he's never done. Just saying.

It's like saying the weatherman was wrong when he forecasted a 60% chance of rain and it didn't rain that day. The weatherman isn't wrong, he just reported the probability of a measurable amount of rain falling in his broadcast area, which includes a probability that it won't rain at all.
User avatar
cmay
lt. kara 'starbuck' trace
lt. kara 'starbuck' trace
 
Posts: 696
Joined: Mon Jan 24, 2011 12:31 pm
Location: austin, tx

Re: Dude: Nate Silver

Postby P.J. Craven on Tue Oct 30, 2012 1:58 pm

MWilke wrote:
P.J. Craven wrote:One guy's methodology is somehow more "rigorous" than the other. Does it really matter when they've both had it very wrong in equal measures?


Yes it does. You're looking at this all wrong. I'm not looking for a prophet.


You framed it in terms of "he got Indiana and North Carolina right!" Which he didn't by that metric, he had Indiana wrong. But suddenly, now it's a forecast.

So when he's "right", he's not really right, and when he's "wrong", he's not really wrong. Unless he calls North Dakota 100% for Romney and it goes to Obama, the man's method is light years beyond all others?

Prediction, forecast, it's really all in how you choose to frame it. When Tom Skilling gets the forecast wrong, guy catches hell.

Individual pollsters vary widely in quality, but a lot of those people aren't skilled statisticians with methodologies of their own? Certainly, it's easy to look better when you're an aggregator of all those numbers.
Last edited by P.J. Craven on Tue Oct 30, 2012 2:02 pm, edited 1 time in total.
ty-lot wrote:Sick bro
P.J. Craven
Humankind's Greatest Musical Genius
Humankind's Greatest Musical Genius
 
Posts: 4049
Joined: Tue Sep 14, 2010 10:55 am
Location: sedentary af

Re: Dude: Nate Silver

Postby MWilke on Tue Oct 30, 2012 2:01 pm

Here's my first post in this thread.

MWilke wrote:
Other than the criticism on Rasmussen, which is over-simplified, most of the criticism is outcome oriented. As far as process goes, he's the best going right now. His methods are more rigorous than anything else I know of right now and they are constantly a work in progress as far as weighting and sifting out noise. He gives people the most accurate possibility of seeing an uncertain future outcome that is heavily dependent on aggregate human psychology. But the key is that it's still forecasting. So to have the expectation that he has to be correct 100% of the time is a poor expectation.


I've already admitted I've mis-remembered the outcome on Indiana in 2008. But it's a little tiresome to re-frame the reality of the situation so you understand it. Are you just really bored today or are you really bad at admitting you're on the weak side of the argument?
Last edited by MWilke on Tue Oct 30, 2012 2:02 pm, edited 2 times in total.
User avatar
MWilke
Lode Star of the Twenty-First Century
Lode Star of the Twenty-First Century
 
Posts: 2269
Joined: Mon May 01, 2006 2:48 pm
Location: West Burbs

Re: Dude: Nate Silver

Postby MWilke on Tue Oct 30, 2012 2:02 pm

DP
User avatar
MWilke
Lode Star of the Twenty-First Century
Lode Star of the Twenty-First Century
 
Posts: 2269
Joined: Mon May 01, 2006 2:48 pm
Location: West Burbs

Re: Dude: Nate Silver

Postby P.J. Craven on Tue Oct 30, 2012 2:04 pm

How is it weak to examine more information than the information that comes through the lens of one guy who has been more than fallible in his forecasts/predictions?
ty-lot wrote:Sick bro
P.J. Craven
Humankind's Greatest Musical Genius
Humankind's Greatest Musical Genius
 
Posts: 4049
Joined: Tue Sep 14, 2010 10:55 am
Location: sedentary af

Re: Dude: Nate Silver

Postby MWilke on Tue Oct 30, 2012 2:07 pm

I'm about finished here. I guess we'll pick this up in 2016.
User avatar
MWilke
Lode Star of the Twenty-First Century
Lode Star of the Twenty-First Century
 
Posts: 2269
Joined: Mon May 01, 2006 2:48 pm
Location: West Burbs

Re: Dude: Nate Silver

Postby Wood Goblin on Tue Oct 30, 2012 2:15 pm

P.J. Craven wrote:How is it weak to examine more information than the information that comes through the lens of one guy who has been more than fallible in his forecasts/predictions?


Obviously, it's not. The problem, however, is that most of the other information is worthless. Silver's work, as well as the work of other poll aggregators and statisticians, demonstrates this.

An example: Silver apparently assessed the outcomes of a thousands predictions made on The MacLaughlan Group. He found that the so-called experts' predictions fared no better than a coin flip. So right there: the experts provided "more information" that I can completely disregard based on actual, rigorous, non-anecdotal evidence.
User avatar
Wood Goblin
Heaven-Sent Hero
Heaven-Sent Hero
 
Posts: 6065
Joined: Tue Jun 06, 2006 12:43 pm
Location: South Loop, Chicago

Re: Dude: Nate Silver

Postby P.J. Craven on Tue Oct 30, 2012 2:24 pm

Silver undervalues GOTV to the point of leaving it completely out of his methodology (you're seriously telling me there's not a way to incorporate this?). He often heavily weights pollsters known for having a serious Republican lean (Gravis, Rasmussen). And sometimes, he'll heavily weight relatively amateur polls that come from colleges stocked with cheap labor. And he went against the grain in Britain only to get completely hosed.

I think his forecasts/predictions/whatever are the most fair and thoughtful. But it's just one more perspective. I tend to give his site the most "weight" (as he does with his polls) when I read the news. But I do not think he should be taken that much more seriously than other number crunchers.

Wood Goblin wrote:He found that the so-called experts' predictions fared no better than a coin flip.


I read this (or saw it on The Daily Show?). I am referring more to other pollsters and polling aggregators, but I think that even the opinions of talking heads are not entirely without merit.
ty-lot wrote:Sick bro
P.J. Craven
Humankind's Greatest Musical Genius
Humankind's Greatest Musical Genius
 
Posts: 4049
Joined: Tue Sep 14, 2010 10:55 am
Location: sedentary af

Re: Dude: Nate Silver

Postby Wood Goblin on Wed Nov 07, 2012 12:39 am

I'd say Silver did fairly well tonight, no?
User avatar
Wood Goblin
Heaven-Sent Hero
Heaven-Sent Hero
 
Posts: 6065
Joined: Tue Jun 06, 2006 12:43 pm
Location: South Loop, Chicago

Re: Dude: Nate Silver

Postby Justin Foley on Wed Nov 07, 2012 1:31 am

Yes he did.

= Justin
Justin Foley
Perfect Picture of Wisdom and Boldness
Perfect Picture of Wisdom and Boldness
 
Posts: 3049
Joined: Fri Apr 29, 2011 1:16 pm
Location: NYC

Re: Dude: Nate Silver

Postby Clyde on Wed Nov 07, 2012 1:34 am

His finally tally was 332-206, with Virginia, Florida and Colorado going to Obama, right? Looks like he might've called it. Not sure what his final call for Senate, House and popular vote were.
NewDarkAge wrote:Dialectics, mate. No easy answers.
Clyde
Lode Star of the Twenty-First Century
Lode Star of the Twenty-First Century
 
Posts: 2384
Joined: Thu Aug 18, 2011 4:12 pm
Location: Life

Re: Dude: Nate Silver

Postby MWilke on Wed Nov 07, 2012 10:06 am

DP
Last edited by MWilke on Wed Nov 07, 2012 10:07 am, edited 1 time in total.
User avatar
MWilke
Lode Star of the Twenty-First Century
Lode Star of the Twenty-First Century
 
Posts: 2269
Joined: Mon May 01, 2006 2:48 pm
Location: West Burbs

Re: Dude: Nate Silver

Postby MWilke on Wed Nov 07, 2012 10:07 am

Here was me in the 'Predict the Election' thread basically plagiarizing Nate Silver.

MWilke wrote:I expect Obama to win the popular vote by about a 1.5-2.0%. I expect Nate Silver's probabilities to be correct on every state, with an exception of an upset in Florida because of how good early voting turnout has been.

Obama: 334
Romney: 204


I don't know where I made the counting error to 334 instead of 332, but thanks for rewarding the faith and making me look good, Nate.
User avatar
MWilke
Lode Star of the Twenty-First Century
Lode Star of the Twenty-First Century
 
Posts: 2269
Joined: Mon May 01, 2006 2:48 pm
Location: West Burbs

Re: Dude: Nate Silver

Postby ::: on Wed Nov 07, 2012 10:11 am

:::
man forced to eat beard
man forced to eat beard
 
Posts: 1890
Joined: Tue Mar 20, 2012 12:37 am

Re: Dude: Nate Silver

Postby tbone on Wed Nov 07, 2012 10:11 am

Dude nailed it. Best part of last night was not having a pit of anxiety burning in my gut waiting for the results to come in. Every single state that came in, I was like "yup, no surprise there."
I play guitar in Mayor For Life and Jackhammer Hearts
User avatar
tbone
Master Of The Computer
Master Of The Computer
 
Posts: 4757
Joined: Mon Apr 19, 2004 2:53 pm
Location: Chicago Proper

PreviousNext

Return to Crap / Not Crap

Who is online

Users browsing this forum: Google [Bot] and 7 guests