Post-election, Nate Silver's book sales soar

Play Video

Statisticians vs. pundits

The significance of Nate Silver's numbers

Nate Silver, the New York Times prognosticator who became something of a controversial figure in the final days of the 2012 presidential campaign, has seen an 850 percent spike of his book sales since nailing his electoral predictions yesterday, according to CNN Money.

The book, "The Signal and the Noise: Why So Many Predictions Fail-but Some Don't," is currently the number two bestseller book on Amazon.com, behind only the children's book "Diary of a Wimpy Kid: The Third Wheel."

Silver writes the blog FiveThirtyEight for the New York Times, and he, in addition to a number of other pundits, appears to have correctly predicted which states would go for President Obama and which would go for Romney in last night's contest. (The winner in Florida has not yet been called.) But in the days leading up to the campaign, his blog, on which he predicted the odds of an Obama victory at over 90 percent, took on a uniquely polarizing status: The right targeted Silver as a liberal apologist who favored skewed polls; the left, meanwhile, revered him as a purveyor of data-driven rationality - particularly in the days following Mr. Obama's widely panned first debate performance, during which Silver continued to bet on an Obama victory.

Amid backlash over his popularity, and the inevitable backlash to that backlash, Silver engaged on Twitter with his detractors, attempting to bet Sandy relief money with MSNBC's Joe Scarborough before being rebuked by the Times public editor in the pages of the paper.

He later backtracked on that bet, donated $2,538 to the Red Cross, and has since taken a somewhat more circumspect approach on Twitter in recent days. After disproving those who doubted his approach, however, he did pause for a moment of self-promotion: "This is probably a good time to link to my book," he wrote on Twitter. Apparently, he was right.

[lawyertom1]

Way to go Nate. You are definitely the man.

[Gwen Kraft]

I would just like to add that since Nate relies on data derived from Polls, and the Polls as a rule leave out many other names, such as Ron Paul or Gary Johnson, Nate can only deal with the results and demographics the media & corporations want in the public eye. You may conclude incorrectly that the Polls reflect the public completely fairly. They do not. The candidates who are denied exposure or given diminished exposure are not thus not provided with a venue to become known. We are still the slaves of those who decide for us whom we shall select among.

[_yourstruly_]

Nate:
Kudos to you for sticking to your guns and to the results of your data.
Your work is yet another facet of how scientific knowledge and arithmetic will knock the right wing zealots out of their dreams.

Just like Sandy knocked out David & Charlie Koch, Sheldon Adelson, Trump the Mump, and many millionaires and billionaires who had their mansions on the coastline of NJ and NC and who, just weeks before Sandy, were funding and passing State laws to ban scientific predictions of global warming. Sad Irony: Now what they have left is just the gate pillars which show where their mansions were..

[nanc12]

He's a genius, pure and simple. In the past couple of weeks, the right was out to destroy him. They said he was a fraud and was just posting nonsense to prop up the base. He had no logic, they said. Even though his site specifically explained how he came up with his data. Math and science works, repubs! You should try it sometime.

[GeorgeH429]

The video that accompanies this story plays like satire. When they sit down and talk to the head of research for CBS, Sarah Dutton, and she describes how CBS's polls are designed to get "inside the head of the electorate", any half-way decent journalist would have asked why they don't focus on predicting a winner instead. I'm not big on conspiracies, but I would imagine the ratings for CBS's election coverage last night would have been substantially lower if they had focused there research team on predicting a winner instead of making the election look as close as possible to keep you watching. To wrap up the story by comparing statistical analysis to a bake-off is an insult to everyone's intelligence and a fitting end to this sham of a story.

[bobnjersey]

[CBS's election coverage last night would have been substantially lower if they had focused there research team on predicting a winner instead of making the election look as close as possible to keep you watching.]
-----------------------------------------------
the numbers i'm looking at show that the election was in fact pretty close. a percentage point or two in one or two states would have resulted in a completely different outcome.

this was also true of the 2008 election as well.

[dtschuck]

Nate Silver should send an authographed copy of his book to Karl "Landslide" Rove. LMAO.

[matt6052]

When Silver tried to bet Scarborough, Silver's model predicted an 80% chance of an Obama win, but he only offered Joe only even money odds. It should have been 4:1, where Silver put up $800 and Joe put up $200 and the winner of the bet won the $1,000.

One record that fell this time is that no candidate who was above 50% in the Gallup Poll a month out has failed to win the presidency-- until now.

What was so different this time?

If it's because of an exposure bounce that Obama received during Hurricane Sandy, then that throws into question whether Silver (and Obama) just experienced some luck.

[SeriouslyD]

Sandy just gave Obama an "exposure" bump? Hmm? What about: 1. Obama gave help unstintingly and without condescension to a major, and quite insulting, political opponent whose state was steamrollered by a disaster. 2. People had to stop and think about Romney as a candidate who was inclined to constantly shrug off major government functions as being more appopriate to the individual states - whether or not the states could feasibly handle or afford those functions. That's not luck, that's voters giving due consideration to presentation of the facts in a new light.

[Beastmode_Crush]

"One record that fell this time is that no candidate who was above 50% in the Gallup Poll a month out has failed to win the presidency-- until now. What was so different this time?"

For starters, the Gallup voter model was way off. I believe Silver devoted an entire column on Gallup's inaccuracy when being an outlier.

This comment illustrates the problem when pundits use secondary data such as indicators and historical precedent to determine the likelihood of an event occurring. If there is no hard data or metrics, then fine, that's all you have. Data-driven analysis is superior.

Case in point: Last week, every anti-gay marriage measure brought before voters passed. Last night there were four measures, and all failed.

[JoshuaPerry]

Why do Republicans hate science and math?

[Kristin789]

Thank you Nate, you kept us sane and rational.
You did not, however, keep us from bring considered bonkers by our Fox News watching friends.

[filmguy107]

Nate Silver: YOU ROCK!!! AWESOME!!!!