Nate Silver has made a career out of predicting things • Hacker News Board

Re: Nate Silver has made a career out of predicting things

#21

Post by waterlesscloud » Sun, Oct 07, 2012, 10:01 PM UTC

tptacek wrote: post #18

sxcurry wrote: post #11

Earlier quoted context omitted.

I don't think he has a leaning one way or the other - he tries to be unbiased, and arrive at data-driven conclusions. What have you seen from Nate that makes him seem a "biased liberal" to you? I'd be very interested to know. Also, we'll see in about a month how good his meta-analysis is.

He leans Democratic and has said so on more than one occasion. He supported Obama in '08. One gets the impression that his professional reputation is more important than his ideological affiliation, and so I find him pretty credible.

He states straight out in his FAQ that he supports Obama.

"What is your political affiliation? My state has non-partisan registration, so I am not registered as anything. I vote for Democratic candidates the majority of the time (though by no means always). This year, I have been a supporter of Barack Obama."

And the next question -

"Are your results biased toward your preferred candidates? I hope not, but that is for you to decide. I have tried to disclose as much about my methodology as possible."

Re: Nate Silver has made a career out of predicting things

#22

Post by tptacek » Sun, Oct 07, 2012, 10:01 PM UTC

textminer wrote: post #5

ithcy2 wrote: post #4

"Predictions are hard—especially about the future." And here I was under the impression that ALL predictions are about the future.

Today's the day you learn about Yogiisms.

Sure it's Yogi? It's also attributed to Samuel Goldwyn and Mark Twain. (I doubt the Twain association).

Re: Nate Silver has made a career out of predicting things

#23

Post by tptacek » Sun, Oct 07, 2012, 10:02 PM UTC

waterlesscloud wrote: post #21

tptacek wrote: post #18

Earlier quoted context omitted.

He leans Democratic and has said so on more than one occasion. He supported Obama in '08. One gets the impression that his professional reputation is more important than his ideological affiliation, and so I find him pretty credible.

He states straight out in his FAQ that he supports Obama. "What is your political affiliation? My state has non-partisan registration, so I am not registered as anything. I vote for Democratic candidates the majority of the time (though by no means always). This year, I have been a supporter of Barack Obama." And the next question - "Are your results biased toward your preferred candidates? I hope not, but that is fo…

That's from 2008, right? FiveThirtyEight.com isn't his site anymore.

Re: Nate Silver has made a career out of predicting things

#24

Post by wilfra » Sun, Oct 07, 2012, 10:03 PM UTC

tatsuke95 wrote: post #16

Nate Silver came from the fantastic world of the statistical analysis of baseball, specifically Baseball Prospectus. You can check out his body of work here (some articles require a subscription): http://www.baseballprospectus.com/author/nate_silver/ His big contribution was a model used to predict player statistics. When he first started writing about politics, I assumed the name was a coincidence.

Actually he came from the fantastic World of poker :)

He gravitated to Sabermetrics from there.

Re: Nate Silver has made a career out of predicting things

#25

Post by kmfrk » Sun, Oct 07, 2012, 10:05 PM UTC

sxcurry wrote: post #11

andrewljohnson wrote: post #7

Republicans say Nate Silver is a biased liberal, and reading his blog this election cycle makes it seem that way to me even. But I would like to know - is there a Republican analog to Nate Silver? Someone who predicts 90% of the races at the end, does the meta-math, and has a computer model? If there is no Republican analog, can you possibly say Nate Silver is liberal or biased? Does using a computer model and math,…

I don't think he has a leaning one way or the other - he tries to be unbiased, and arrive at data-driven conclusions. What have you seen from Nate that makes him seem a "biased liberal" to you? I'd be very interested to know. Also, we'll see in about a month how good his meta-analysis is.

I don't really see why it's wrong that Nate has political opinions. The alternative is Jim Lehrer who values his "objectivity" so much that he doesn't even vote at elections.

Sometimes, facts favour liberals, sometimes they don't.

Re: Nate Silver has made a career out of predicting things

#26

Post by waterlesscloud » Sun, Oct 07, 2012, 10:06 PM UTC

tptacek wrote: post #23

waterlesscloud wrote: post #21

Earlier quoted context omitted.

He states straight out in his FAQ that he supports Obama. "What is your political affiliation? My state has non-partisan registration, so I am not registered as anything. I vote for Democratic candidates the majority of the time (though by no means always). This year, I have been a supporter of Barack Obama." And the next question - "Are your results biased toward your preferred candidates? I hope not, but that is fo…

That's from 2008, right? FiveThirtyEight.com isn't his site anymore.

Sure, but I'm having a hard time imagining the answer to either question has changed. If it had, I'm sure he's capable of having that information removed from a post bearing his name.

Re: Nate Silver has made a career out of predicting things

#27

Post by mtgx » Sun, Oct 07, 2012, 10:13 PM UTC

sxcurry wrote: post #11

andrewljohnson wrote: post #7

Republicans say Nate Silver is a biased liberal, and reading his blog this election cycle makes it seem that way to me even. But I would like to know - is there a Republican analog to Nate Silver? Someone who predicts 90% of the races at the end, does the meta-math, and has a computer model? If there is no Republican analog, can you possibly say Nate Silver is liberal or biased? Does using a computer model and math,…

I don't think he has a leaning one way or the other - he tries to be unbiased, and arrive at data-driven conclusions. What have you seen from Nate that makes him seem a "biased liberal" to you? I'd be very interested to know. Also, we'll see in about a month how good his meta-analysis is.

Must be because "reality has a liberal bias".

Re: Nate Silver has made a career out of predicting things

#28

Post by andrewljohnson » Sun, Oct 07, 2012, 10:14 PM UTC

waterlesscloud wrote: post #17

andrewljohnson wrote: post #12

Earlier quoted context omitted.

It correlates! Global warming, 2/3 Republican deny it, more than 3/4 Democrats believe it: http://www.businessweek.com/news/2012-09-30/global-warming-l... Creationism is 38% for Democrats, 60% for Republicans: http://www.cultureofscience.com/2011/08/19/republicans-and-d... I think Republicans are more likely not to want to mention or distribute condoms to teens in sex education, but I don't regard this as quite the s…

The greatest trick political parties ever pulled was making this kind of thing part of their images. With only two choices, people are roped into supporting all kinds of things that make no sense for them. But you get "My party believes in science!" and "My party believes in God!" and it becomes part of people's political identities. And then all sorts of other issues are piled on top of it and you find yourself wond…

If it truly is science vs. religion, then the Republican party will die and be reborn as something new. Religion loses to science over the course of history.

Re: Nate Silver has made a career out of predicting things

#29

Post by tptacek » Sun, Oct 07, 2012, 10:15 PM UTC

waterlesscloud wrote: post #26

tptacek wrote: post #23

Earlier quoted context omitted.

That's from 2008, right? FiveThirtyEight.com isn't his site anymore.

Sure, but I'm having a hard time imagining the answer to either question has changed. If it had, I'm sure he's capable of having that information removed from a post bearing his name.

Lots of people who supported Obama in 2008 no longer do. Meanwhile, why are you so sure he'd update a FAQ entry on a site he no longer maintains?

Re: Nate Silver has made a career out of predicting things

#30

Post by arrrg » Sun, Oct 07, 2012, 10:19 PM UTC

andrewljohnson wrote: post #7

Republicans say Nate Silver is a biased liberal, and reading his blog this election cycle makes it seem that way to me even. But I would like to know - is there a Republican analog to Nate Silver? Someone who predicts 90% of the races at the end, does the meta-math, and has a computer model? If there is no Republican analog, can you possibly say Nate Silver is liberal or biased? Does using a computer model and math,…

Humans who are not biased probably do not exist. Or at least do not care or know about politics in the slightest. I think there is nothing wrong with that. That doesn't invalidate his work.

I know too much about human biases to even try to claim that his do not influence his interpretations and work at all – I just don't think that's possible – but that is general and in all likelihood inevitable problem. I think it's pretty clear, however, that it's in his best interest to make accurate predictions, not partisan predictions.

I do not think it's really helpful to start out with that question. It puts too much focus on what's actually important. Yeah, bias can be hard to detect, but a little bias does not make models useless and when there is a lot of bias you should actually be able to point at what's wrong with the model instead of having to resort to some diffuse claims of bias.

Writing this made me think of Elisabeth Noelle-Neumann (http://en.wikipedia.org/wiki/Elisabeth_Noelle-Neumann). I disagree with many of her views and I think it would be fair to call her biased after reading her most famous work (The Spiral of Silence), but I do respect her work. There is a difference.