Green Room

The beatification of Nate Silver

posted at 9:24 pm on November 8, 2012 by

Dan Foster doesn’t get it. Neither do I.

Though it had lots of different inputs, especially early on, by the end Silver’s famed model was an approximation of a simple Monte Carlo of the polling data; or, if you like, a different way of presenting a variety of polling averages. It was not sorcery. Silver deserves credit for analyzing the data he had in a plausible, straightforward, and fairly transparent way, and for communicating his assumptions at every turn.

But the crap that Silver took from a lot of conservatives being held up as evidence of the supposed mass delusion and anti-empiricism of the Right makes little sense. I think a lot of conservatives with humanities degrees (myself included) instinctively thought that all the decimal places were too cute by half – just like a lot of liberals with humanities degrees instinctively thought every additional decimal place meant MOAR SCIENCE. (Aside: This is a longstanding bugaboo of mine. The “I Swear to Science!” and “Science Bless You! Haha Lolz” crowd are often in the same epistemic position vis-à-vis statements they take on bald authority as are cult members. The fact that journalists writing dire warnings about global warming, or activists campaigning against intelligent design could learn about long-term climate modeling or the mechanisms of natural selection if they wanted to doesn’t change the fact most of them don’t.)

He deserves his victory lap for getting the result right, but so do Drew Linzer and Simon Jackman and Sam Wang, who appears to have nailed the popular vote precisely (51.1 percent). They were all correct (because they were all modeling the same state polls), yet it’s Silver who’s on “The Daily Show” while the other guys are basically as anonymous as they were six months ago. And they’re not the only ones being overlooked. I keep wondering how Tom Jensen of PPP or Lee Miringoff of Marist (NBC/WSJ) are feeling watching Silver be treated as some sort of oracle. They were the ones who gave him the raw data for his model; they were the ones whose reputations were on the line when more established pollsters like Gallup showed a break in the national vote towards Romney. Silver had cover from Linzer et al. to support his state-poll model but Jensen and Miringoff were out on a limb in seeing a deep-blue electorate — and they got it right. Not a perfect analogy here, but it’s as if Silver predicted that Josh Hamilton would hit 100 homers this season and then Hamilton did hit 100 homers, and the media responded by swarming around Silver, not Hamilton. Note to liberal math wizards: Both accomplishments are impressive, but one’s more impressive than the other.

The reason Silver’s being worshiped, I think, is because (obviously) he’s got a vastly larger platform than the other modelers and because he writes engagingly on his subject, for which he also deserves credit. He was the left’s security blanket for six months; whenever they feared that Mitt Romney, the new Republican Hitler, was on the verge of destroying America as we know it, Silver was there to reassure them that there was still a 71.4935 percent chance that it wouldn’t happen. Looking forward to seeing how the model does in an election where nearly every major state poll isn’t breaking for the same candidate in the last few weeks.

Recently in the Green Room:

Blowback

Note from Hot Air management: This section is for comments from Hot Air's community of registered readers. Please don't assume that Hot Air management agrees with or otherwise endorses any particular comment just because we let it stand. A reminder: Anyone who fails to comply with our terms of use may lose their posting privilege.

Trackbacks/Pings

Trackback URL

Comments

Comment pages: 1 2

No, you weren’t being “lied to,” for cripes sake. We made an assumption that proved incorrect.

Allahpundit on November 8, 2012 at 11:55 PM

AP, is this all you all have to say for yourselves? Oops, sorry, we were wrong? As far as I can tell, you all based your assumptions based your assumptions on a gut feel rather than quantitive data. I bring this up because currently Hotair.com is feeding to its readers a steady stream if war propaganda. As far as i can tell, your war advocacy is based on your gut feel that Iran is feverishly pursuing nuclear weapons and will immediately wipe Israel off the face of the earth. You haven’t based your war advocacy on facts and data because according to the National Intelligence Estimate, Iran stopped its nuclear weapons program in 2003 and that they have not made the decision to restart their program. In fact, there has been no evidence that has been made public that Iran enriched beyond the limits allowed by the Non Proliferation Treaty. After your war has started and a lot of people are killed, what will you say if no nuclear weapons program is found? Oops, sorry, we were wrong?

antifederalist on November 9, 2012 at 9:23 AM

They weren’t “gut” feelings. They were based off the 2010 turnout and the Wisconsin election turnout. Also, when did AP advocate going to war with Iran? Or when did EM for that matter?

GiantOrb on November 9, 2012 at 9:46 AM

Ha! Red_Herring’s at it again in this thread. “OMG! How could Hot Air have LIED to us!?!? They’re so not experts at anything!”

This isn’t difficult folks. PPP made an assumption about the electorate that proved to be correct, i.e. despite 2010, presidential history, unemployment numbers, etc., the electorate would still be more heavily Democrat due to high turnout by some of the least reliable voters (young, Latino, African American). Conservative pundits and other pollsters made an assumption that there’s no way based on history, unemployment numbers, enthusiasm, etc. that those historically unreliable voters would turn out in numbers high enough to win. That assumption ended up being incorrect. Simply put, one assumption was correct, the other was not, like most things.

The whole Ed and AP lied and/or misled us so they must apologize is getting boring and is infantile. Don’t like it? Don’t read it.

change is for suckers on November 9, 2012 at 10:59 AM

No, you weren’t being “lied to,” for cripes sake. We made an assumption that proved incorrect.

Allahpundit on November 8, 2012 at 11:55 PM

And what was the justification for your rosy assumption, apart from vague stuff such as number of Romney/Ryan yard signs or the amount of noise produced by Republicans at Romney rallies in PA..? To liberals, it was obvious that Obama was in a stronger position than anticipated simply because he was the incumbent and also because of demographic changes. If your job approval is in the 49-50% range (and 90%+ among Democrats!), if you are spending record amounts on money on good GOTV operation plus early attack ads to discredit your opponent, then you are highly likely to win. Period. There is EMPIRICAL evidence for this — just look at the 2004 election! And the Dems’ get-out-the-vote operation was excellent as far back as in 2000 despite relatively low enthusiasm for Al Gore. That is why I always assumed the state polls most likely represented Obama’s floor, and that his reelection was almost guaranteed despite the state of the economy.

mlindroo on November 9, 2012 at 11:02 AM

Ed and Allah,

Thanks for unskewing the polls. You guys predicted the razor thin Obama victory. If it was not for your yeomans effort we would all have been expecting an obama blow out. (I guess the only upside3 would have been that Mitts narrow loss would have been a nice surprise).

sheikh of thornton on November 9, 2012 at 11:51 AM

They weren’t “gut” feelings. They were based off the 2010 turnout and the Wisconsin election turnout. Also, when did AP advocate going to war with Iran? Or when did EM for that matter?

GiantOrb on November 9, 2012 at 9:46 AM

How accurate would Nate Silver, a person whose track record with his predictions is not 100%, have been if the election had been before Sandy hit?

Anyone who feels that Ed or AllahPundit somehow misled them about the polling data is an emotional retard.

Anti-Control on November 9, 2012 at 11:55 AM

the thinnk that bugged me the most about silver’s “model” is the whole “83% chance of re-election”. its so friggin bogus. in an election w/ 2 candidates ( yeah, i know but shut up) the chance of winning is always 50%. same as a coin flip.

chasdal on November 9, 2012 at 11:56 AM

Good lord people can whine.

It’s not HA’s fault we lost the election, they were wrong, most conservatives were wrong.

And even if they did make it up the whole time, what difference is it to you? Are you so dumb as to think Ed or AP’s prediction dictated the election? Don’t understand it. Yeah, I’m mad we lost too, and I’m upset at how wrong more or less all of conservative media had it, but I thought the assumptions were rational.

Would you really have seen higher Dem turnout with 8% unemployment for four years and a generally dismal economy, plus healthcare nationalization among the myriad of other problems caused by the administration? Seemed insane to me, particularly given the 2010 elections and the massive opposition to Obamacare.

Yet here we are. You can stop being butthurt and do something productive or you can keep whining, but I’m not going to stop reading HA (nor will most readers) because your feelings are hurt.

John_Locke on November 9, 2012 at 12:02 PM

the thinnk that bugged me the most about silver’s “model” is the whole “83% chance of re-election”. its so friggin bogus. in an election w/ 2 candidates ( yeah, i know but shut up) the chance of winning is always 50%. same as a coin flip.

Yeah…that’s completely wrong. God, I really hope you don’t work in politics….

If a Democratic candidate is running in Utah against a Republican candidate, the odds of either candidate winning are not 50%. They are like 90% for the Republican.

So, you see, it’s not at all like a coin toss… That’s why we have these things called “polls”. Because, generally, one candidate is favored over the other.

agirlacamera on November 9, 2012 at 12:07 PM

I’m not going to ban people for disagreeing with me, even when they’re trolling, as red_herring now is. As long as they’re civil, they can comment.

Allahpundit on November 9, 2012 at 12:32 AM

Thank you for this enlightened attitude.

I couldn’t stand how disruptive gumbyandpokey was, but it wasn’t your job to counter him – it was ours, the commenters, who did a pretty good job of that in my opinion with the Trollcott, which needs to be applied to any poster who shows no interest in being civil or intellectually honest.

Anti-Control on November 9, 2012 at 12:12 PM

We knew someone should be getting out of the polling business after Tuesday. Now we know who it is.

Moesart on November 9, 2012 at 12:12 PM

That’s really not true. If Romney was ahead, he would have argued just as strongly for THAT prediction. I am sure of it. In fact, Nate called 7 of out of 8 Senate races correctly and the one that he got wrong was South Dakota. He predicted the REPUBLICAN would win and the Democrat won instead. Also, in 2010, he predicted Angle would win because that is what the polls said, and he was wrong there too when Reid won out of nowhere. So he just responds to what the polls suggest, not to what he wants to happen.

Which means that when there’s A LOT of polling data, like for a presidential race, his predictions will be pretty accurate because there is a lot of data.

agirlacamera on November 9, 2012 at 12:17 AM

What I highlighted is why I called you out on your stupid insult the last time we interacted.

Sandy helped save Silver’s reputation, with some people at least.

Anti-Control on November 9, 2012 at 12:19 PM

Fun to watch geniuses criticize Hot Air for doubting the “infallible” data, might want to go look at how Rasmussen and Gallup ended up faring with their “infallible” data (hint, not good). It’s always good to question and polls sometimes are right and sometimes they’re wrong, if you want to go with the notion that they’re always right, well, good luck with that.

clearbluesky on November 9, 2012 at 1:13 PM

What I highlighted is why I called you out on your stupid insult the last time we interacted.

Sandy helped save Silver’s reputation, with some people at least.

Soooo, what is your point? That polls can be wrong? Well, sure they can, but the sensible thing isn’t to assume that they are wrong when it’s a presidential race and there are 20+ different polling firms, polling different swing states, and pointing to an Obama victory. The sensible isn’t to assume they are wrong to the point that you predicting Romney will win a landslide and are then completely shocked that he loses. Doing that is completely nuts.

And my point stands. That Nate will predict a Republican victory if that is what polls suggest. And his two high-profile misses were both in favor of the Republican.

agirlacamera on November 9, 2012 at 1:27 PM

Yeah…that’s completely wrong. God, I really hope you don’t work in politics….

If a Democratic candidate is running in Utah against a Republican candidate, the odds of either candidate winning are not 50%. They are like 90% for the Republican.

So, you see, it’s not at all like a coin toss… That’s why we have these things called “polls”. Because, generally, one candidate is favored over the other.

agirlacamera on November 9, 2012 at 12:07 PM

And how scientifically sound is it to call an election four months out, ala Drew Linzer? He’s now crowing that his poll called the election all the way back in June. You think that’s sound methodology, when we don’t even know how the election would have turned out in the absence of Hurricane Sandy, let alone the million other things that happen in 4 months of an election cycle?

When Silver and Linzer and the rest of the “analyterati” class are calling it 85.56428736% for Hillary Clinton in June of 2016, are you going to be telling us that we can’t ‘ignore math’ and we should just accept the loss?

Good Solid B-Plus on November 9, 2012 at 1:27 PM

If a Democratic candidate is running in Utah against a Republican candidate, the odds of either candidate winning are not 50%. They are like 90% for the Republican.

agirlacamera on November 9, 2012 at 12:07 PM

Unless it’s Mia Love, of course. Can’t let any of those Black Republicans win a seat, it might kill the narrative that the GOP is a cabal of rich white men.

Good Solid B-Plus on November 9, 2012 at 1:30 PM

When Silver and Linzer and the rest of the “analyterati” class are calling it 85.56428736% for Hillary Clinton in June of 2016, are you going to be telling us that we can’t ‘ignore math’ and we should just accept the loss?

Of course not! Work for your candidate. Donate, make calls, GOTV, make your case against Hillary, speak out for the Republican candidate’s positions etc, etc. By all means, do all of that stuff. Just don’t be unskewing polls. Don’t be saying PPP and Nate Silver are lying. And don’t be predicting landslides based on who-knows-what! Basically, don’t bury your head in the sand. There is an extent to which ALL political junkies will do this, because that’s a human tendency. But when 94% of HA readers have no clue what is going to happen, that is a lot of unnecessary heartache. You are ALWAYS better off in knowing the facts.

If you look at Nate Silver’s trendline you see that Romney came REALLY close to overtaking Obama after the first debate. Just because it was predicted in June doesn’t mean it was over. He could have won. He just didn’t.

agirlacamera on November 9, 2012 at 1:36 PM

If you look at Nate Silver’s trendline you see that Romney came REALLY close to overtaking Obama after the first debate. Just because it was predicted in June doesn’t mean it was over. He could have won. He just didn’t.

agirlacamera on November 9, 2012 at 1:36 PM

So the glorious victory for the modelers is that they can call a victory with just about absolute certainty when we’re less than a week out and they have a preponderance of data from state polling, and this is supposed to represent some sort of fantastic evolution of political blogging from “fact-free punditry” to “empirical analysis”?

Face it, this isn’t about Nate Silver, it’s about stereotypes. Even though the exact same thing happened in 2004 re: Kerry and sites like Daily Kos, you want to press the lie that Conservatives are “low, low, low info voters,” that they’re “anti-science,” that they don’t understand statistics. Nate Silver is a shibboleth of that; if you do puja to him, then you understand his brilliance as a shining aleph of truth, and you are accepted into the grand priesthood of SCIENCE. If you were one of the GOP Mandarins who had the gall the question the divine tractates of Nate the Magnificent, you are everything that is wrong with America, a luddite who is afraid of math, science and rationality.

Good Solid B-Plus on November 9, 2012 at 1:46 PM

Unless it’s Mia Love, of course. Can’t let any of those Black Republicans win a seat, it might kill the narrative that the GOP is a cabal of rich white men.

I am not that familiar with this race. But if she lost, since this is Utah and all, it was because Republicans and Independents didn’t vote for her enough. The Democrats will obviously vote for the Democratic candidate.

agirlacamera on November 9, 2012 at 1:49 PM

So the glorious victory for the modelers is that they can call a victory with just about absolute certainty when we’re less than a week out and they have a preponderance of data from state polling, and this is supposed to represent some sort of fantastic evolution of political blogging from “fact-free punditry” to “empirical analysis”?

It must not be that easy to do this! Karl Rove, Michael Barone, Dick Morris, 94% of Hot Air readers, most of the conservative media, etc, etc, etc failed to do it.

2004 actually was a close race. Bush won by 286 electoral votes vs. Obama winning by 332 and being ahead in every single swing state’s polling…. Despite 2004 being a closer race there was no such thing as the entire liberal media system unskewing polls and predicting Kerry landslides.

agirlacamera on November 9, 2012 at 1:56 PM

Soooo, what is your point? That polls can be wrong? Well, sure they can, but the sensible thing isn’t to assume that they are wrong when it’s a presidential race and there are 20+ different polling firms, polling different swing states, and pointing to an Obama victory. The sensible isn’t to assume they are wrong to the point that you predicting Romney will win a landslide and are then completely shocked that he loses. Doing that is completely nuts.

And my point stands. That Nate will predict a Republican victory if that is what polls suggest. And his two high-profile misses were both in favor of the Republican.

agirlacamera on November 9, 2012 at 1:27 PM

I can see that you are not very good at connecting the BIG dot between your prejudice against Conservatives and your partiality towards Silver.

By your own statement, “he [Silver] predicted Angle would win because that is what the polls said, and he was wrong there too when Reid won out of nowhere” – why, then, do you continue to act so incredulously about Conservatives who didn’t have as much faith in the polls predicting an 0dumba win as you and Silver did? Doing so makes you look like a dumb partisan hack who is not very self-aware.

(I’ll insert here that if Sandy hadn’t hit, I believe we wouldn’t even be having this conversation, because Silver would have been about as wrong as we thought he was going to be, like he was, “out of nowhere“, about Angle vs. Reid)

Anti-Control on November 9, 2012 at 3:00 PM

It must not be that easy to do this! Karl Rove, Michael Barone, Dick Morris, 94% of Hot Air readers, most of the conservative media, etc, etc, etc failed to do it.

2004 actually was a close race. Bush won by 286 electoral votes vs. Obama winning by 332 and being ahead in every single swing state’s polling…. Despite 2004 being a closer race there was no such thing as the entire liberal media system unskewing polls and predicting Kerry landslides.

agirlacamera on November 9, 2012 at 1:56 PM

Were Karl Rove and Dick Morris actually predicting a Romney landslide based on the data, or, like the internal polling the Romney campaign released, was it a head-fake to try and stop crushing despair from overtaking the GOP GOTV effort?

Good Solid B-Plus on November 9, 2012 at 3:51 PM

Comment pages: 1 2