"When quants tell stories," or why Nate Silver didn't win the day with awesome math alone

Felix Salmon at Reuters: "If you think that the value of Nate Silver is in the model, you’re missing the most important part: there are lots of people with models, and most of those models are pretty similar to each other. The thing which sets Silver apart from the rest is that he can write: he can take a model and turn it into a narrative, walking his readers through to his conclusions."


  1. I work in the applied sciences, and turning science into english for clients is a huge challenge. Sure, you can do good work, but can you tell someone with a finance background what they just paid you to do, and how it turned out?Because that is the key to repeat clients.

    1. Many scientists, engineers, and other people who work with numbers can do the math.  Being able to explain the process and results concisely and simply to a layperson audience is a gift. Which is why Nate Silver is a gawd.

      1. I’m a technical editor and I will second that. When I get a report that’s well written from the start, I’m pleasantly surprised. It’s not really the usual thing. The two skill sets are quite different, and few people have both.

        The Nate Silver blog was one of the things that kept me sane(ish) when the media kept talking about Romney’s supposed momentum. Loved seeing him on The Daily Show the other night; seems a right nice and humble chap.

  2. Well, Sam Wang has a quirky style that’s laced with humor. Silver made a notorious wager offer to Joe Scarborough, but Dr. Sam was the only one who offered to eat a bug if he was wrong about certain states (and later linked to a catalogue of edible bugs, just in case). But yeah, Silver is a pretty good communicator – that’s a big reason why his famous baseball quant model, PECOTA, became so popular so quickly. 

  3. Unfortunately, the corollary would seem to be that people who can skip the wonk stuff and just start with a nice red-meat story will obtain an even greater storyteller advantage without any of the tedious ‘knowing things’ work.

    See, for instance, the reality vacuum in which much of network TV(where, after all, a close-race-to-the-last-moment fills much more air time and is rather better for ratings) managed to live despite being easily capable of, and probably actually possessing, the same modelling expertise…

    1. But the good thing is that the people that can run the numbers can now gain an audience just with that ability. And if you’ve got a way with words as well, you can attain “Witch” status – see #natesilverisawitch on Twitter.

      Actually, the best numbers guy may be Drew Linzer at Votamatic – his economic-based prior + polls model predicted 332 EV for Obama literally MONTHS ago. But he’s not the communicator that Silver (or Wang) is.

      1. Don’t get me wrong: TV+internet beats the hell out of TV. It’s just that, unfortunately, the storytelling edge appears to continue to provide an advantage even among people who have abandoned even the pretense of telling stories in order to convey models of reality and extends merrily out to people who build models of reality in order to sustain a good story.

    2. See, for instance, the reality vacuum in which much of network TV (where, after all, a close-race-to-the-last-moment fills much more air time and is rather better for ratings) managed to live despite being easily capable of, and probably actually possessing, the same modelling expertise…

      Could some of that also be about getting people out to vote? If I’m told that the person I’m planning on voting for has a big lead, I might be discouraged from standing in line for 7 hours to vote for him. And maybe he’d still win after that, but some other races might be a lot closer.

  4. Fortunately, liberals have Nate Silver, while conservatives are stuck with people like Dick Morris, who is out with another crazy prediction: http://hollywoodandswine.com/following-prediction-of-romney-landslide-fox-news-dick-morris-predicts-oscar-sweep-for-thats-my-boy/

    1.  There actually was a right-leaning statistician that had accurate numbers.  I can’t remember his name or his site’s name, but Andrew Tanenbaum spoke of him on electoral-vote.com, which is the poll aggregation/prediction site that I used most of the time.  Dr. Tanenbaum’s predictions were also pretty accurate, which isn’t surprising since his methods appear to be similar to Silver’s.  Certainly his discussion probably falls slightly to the left of Silver, as he is an American ex-pat living in Amsterdam, but his methods were politics-less in the sense that he wanted to compare polls to each other, evaluate the accuracy of the polls themselves, and to find outliers and bias, which he did find plenty of.

      As far as I’m concerned, the statistics people have it, if the polls are conducted properly.  I’m a lot more inclined to listen to data-driven perspectives than the derp that we normally get.

      1. i think you’re talking about Scott Elliott, aka the “Blogging Caesar” at http://www.electionprojection.com . Elliott had doubts that the polls were telling the whole story, and even did an unofficial “unskewing” exercise of his own. But all of his official projections, including the election morning call of 303 EV for Obama (missing Florida, which is just as understandable as Silver missing Indiana in 2008), were made using the statistical model that had served him so well in the past.

    2. I think Silver called the GOP midterm sweep as well. I wouldn’t really put it in terms of right vs. left as maybe RBI vs. WAR… :)

      1. Silver and Wang both predicted GOP takeovers in the 2010 House elections, but underestimated quite a bit – Silver missed by 9 seats, Wang by 12. The winner of the 2010 Quant Count would be Scott Elliott, the conservative quant – he overprojected the GOP wave by a single seat. All three nailed the 2010 Senate races save for “Comeback” Harry Reid’s race – no one saw that one coming.

      2. try WARP instead, Navin!

        I remember a campaign we did with some other websites after Nate took over at BP, and one of the components was a piece of original writing from each participating website that the clients would run as part of their promotional package.  Now I might be biased, but Nate’s article was so much more fun to read than any of the other ones submitted…

Comments are closed.