Photo: Getty
Nate Silver’s Not As Foxy As He Thinks He Is

Nate Silver’s Not As Foxy As He Thinks He Is

By Photo: Getty

Last week, Nate Silver’s FiveThirtyEight website—which relaunched Monday at its new ESPN home—unveiled a new logo: a crafty orange fox. Why a fox? Silver cited Sir Isaiah Berlin’s essay on “The Hedgehog and the Fox,” which, in turn, takes its title from a quote attributed to ancient Greek poet Archilochus: “The fox knows many things, but the hedgehog knows one big thing.”

Silver esteems himself a fox. “Plenty of pundits have really high IQs, but they don’t have any discipline in how they look at the world, and so it leads to a lot of bullshit, basically,” he told New York magazine's Joe Coscarelli. Asked for examples of hedgehogs, Silver offered his usual villains, the op-ed columnists: “They don’t permit a lot of complexity in their thinking. They pull threads together from very weak evidence and draw grand conclusions based on them.…It’s people who have very strong ideological priors … They’re not really evaluating the data as it comes in.” Foxes, he wrote in his 2012 bestseller The Signal and the Noise, "are scrappy creatures who believe in a plethora of little ideas and in taking a multitude of approaches toward a problem. They tend to be more tolerant of nuance, uncertainty, complexity, and dissenting opinions."

Another way of putting it would be that pundits/hedgehogs think deductively, starting from grand premises and applying them to isolated facts; Silver/foxes think inductively, amassing facts and seeing what conclusion those facts suggest.

Here’s the thing, though. Foxes don’t write 3500-word introduction-cum-manifestoes, as Silver did Monday for his new site, even in support of something as foxy as data journalism. Foxes are extremely suspicious of self-justifying systems, and thus would never write (as Silver does in his book), of a proposition they subscribe to, “Bayes’s theorem predicts that the Bayesians will win.” And they aren’t dogmatically anti-hedgehog, because they aren’t dogmatically anything. (They’re foxes, not dogmas.)

There’s a touch of the hedgehog about Silver. Silver argues in his book that viewing the world without pre-existing biases and with a strong sense of probability is superior. “When we construct these stories,” he writes in The Signal and the Noise of hedgehogs’ premises, “we can lose the ability to think about the evidence critically.” He was similarly wary about the use of anecdotes in his introduction, although he acknowledged that FiveThirtyEight will use them occasionally. Many great thinkers would disagree with Silver and insist on the importance of such stories; religious thinkers are one example. It doesn’t make them right, but it doesn’t make Silver right, either. What it suggests is that the belief that human-imposed conceits get in the way of what the facts tell us is itself a hedgehog-like insight.

There’s nothing inherently wrong with Silver’s hedgehog status, although I imagine Silver would be rather prickly about the charge. But by denying his own “ideological priors,” Silver risks obscuring the limits of his approach.

That limit isn’t his fealty to numbers and data, a fealty that anyway he partially disavows in Monday’s article. Rather, it’s the importance he places on outcomes. (His emphasis on outcomes is why he thinks it so important not to let ideological premises interfere with observable facts.) Silver is all about outcomes, and in a way that suggests he may not fully own up to it. When he criticized Politico, a favorite bête noire, for trying to cover politics “like it’s sports, but not in an intelligent way at all,” he left unsaid that he himself covers politics like it’s sports—focused on who is likelier to win and who is likelier to lose. No wonder Politico’s Dylan Byers is really excited for the new FiveThirtyEight. Those two sparring partners are likely more similar than Silver would prefer to admit. “By adding probabilities to analysis,” Byers writes, “Silver is putting his neck on the line every time he hits publish—and it’s always fun to watch someone put their neck on the line.” Yes, particularly if you’re Politico’s media reporter!

Silver’s outcome-intensive approach risks obscuring the processes and the personalities, the ideas and the ideologies, which in politics matter also. A myopic focus on outcomes can also prove fatal to storytelling, which Silver announced was a prime goal of his site: though data journalism and narrative journalism need not be mutually exclusive, a data-based story still needs to tell a good story; and stories have beginnings and middles, too.

True, in certain areas, all we really do care about are outcomes. “Just W’s and L’s,” as Boston Red Sox second baseman Dustin Pedroia told Silver for Silver’s book. It is no coincidence that when Silver wrote, “It is forecasting’s original sin to put politics, personal glory, or economic benefit before the truth of the forecast. Sometimes it is done with good intentions, but it always makes the forecast worse,” it was in his chapter on weather—the best example of a field where outcomes basically are the only important thing. (It might be fun to understand pressure systems, but all anybody needs to know is whether or not it’s going to rain.)

But FiveThirtyEight’s purview has expanded to far more than sports and the weather to include politics, science, even “Life.” Outcomes aren’t always the most important thing about “Life.” And when you combine Silver’s insistence on focusing on outcomes with his emphasis on data, you can get something like Monday’s “Life” section post on the merits and demerits of toilet-seat covers by one of the site’s authors. The piece succeeded in making me interested in the cultural and social implications of toilet-seat covers, but failed to do what it sought to do, namely, tell me whether I should use them.

Outcomes aren’t even always the most important thing about sports! FiveThirtyEight’s NCAA tournament predictions will give you the likelihood of various outcomes and may help you win your pool. But they won’t tell you which games to watch, or which teams to root for or against. Other sites will, so that’s fine. FiveThirtyEight need not provide all the answers. It just needs to avoid deluding itself into believing it does.
Loading Related Articles...
The Plank
Article Tools