THE ENVELOPE PLEASE JULY 22, 2013
Nate Silver, the stat guru moving soon from The New York Times to ESPN/ABC, will cover sports, as he did early in his career. According to Politico’s Mike Allen, Silver will continue do politics, where he achieved fame and renown. He will also do weather, to which he devoted a chapter in his recent book, The Signal and the Noise, and which seems a natural field for a “forecaster.” And he apparently has developed an interest in an equally wonky topic, education.
Then there are the Oscars. “The battle for data whiz Nate Silver,” Allen reported, “was won by ESPN and ABC News (both part of The Walt Disney Company) after the 35-year-old was promised extensive air time, a role in the Oscars (airing on ABC through at least 2020), and a digital empire.” Silver’s apparently significant desire to do extensive Oscars coverage seems like the one curveball here. But actually, understanding why Silver wants to work the Oscars—specifically, as he has done for several years for the Times, to forecast the winner of each award—is crucial to understanding Silver’s larger ambitions and to appreciating the extent to which Silver’s worldview has permeated the culture.
It’s also crucial to understanding Silver's limits.
Like sports, the Oscars are not Important. But people pay attention to sports because they are frequently surprising and rarely predictable. And like politics, the Oscars are somewhat predictable. But people pay attention to politics because they are Important. If Silver broadcasts his Oscar forecasts on the actual Oscar telecast and turns out to be usually right, ABC runs the genuine risk of sucking the fun out of the Oscars, whose entertainment derives—along with the dresses, the bad jokes, and Jack Nicholson—from the unpredictability of the outcome.
For the Oscars and politics share something that sports doesn’t: a heightened magnitude of predictability. Political scientist Robert W. Erikson explained this to me about politics when I was writing about Silver last year: “If you look at the national polls and the state polls, they tend to converge on the correct answer as Election Day approaches,” he said. “If you look at polls the last week or two, why shouldn't they be accurate?” Put another way: A poll taken three days before an election is measuring something very close to the thing you are trying to forecast. Barring an October Surprise-type event, the poll and the vote should match pretty closely. This is why Silver articulated on Election Day a greater than 90 percent chance of a Barack Obama victory despite also (correctly) believing the election would be relatively close.
By contrast, in sports you can deploy forecasting techniques to predict which team is more likely to win a head-to-head contest, but there is no equivalent in politics—especially at the level of frequency—to a bad outing by a starting pitcher, a fluky fumble, a bad shooting day, an unlucky power play, a penalty kick, and all the other things that make sports exciting and unpredictable (things that would be outrageous if they took place in areas of human activity that actually matter). In sports, you cannot come nearly as close to measuring the thing you are trying to forecast. (As Erikson explained: “It’s a crapshoot in the playoffs. You have five games. Who would have predicted the Giants winning the World Series?”) This is why Silver’s record of forecasting sports outcomes is easily deemed “shaky.” There are only very few people who can actually predict sports outcomes with any kind of regularity—you can find most of them in Las Vegas—because sports resists such analysis.
In terms of predictability, the Oscars are more like politics than sports. By looking at the outcomes of prior awards—which are selected by many of the same people who vote on the Oscars—you are measuring something very close to the thing you are trying to predict. A quick example would be that in nine of the past eleven years, the Directors Guild’s pick for best picture was the film that won the Oscar.
The Atlantic Wire’s Connor Simpson noted that Silver’s Oscar projections have actually been less than stellar. He has guessed the six major categories three times, going four-for-six twice and five-for-six once. Not bad, but not great. We can expect them to get better as Silver invests more time in them (at the Times, they amounted to a blog post or two; presumably in his new job, he will be going on television to be viewed by millions). And, again, almost nobody can predict sports correctly 62.5 percent of the time; if many could, then we would have many more millionaires.
So imagine next year’s telecast. Millions tune in. Maybe many of them know who is taking home hardware for Best Actress, but a heavy perfume of suspense hangs over the evening nonetheless. Then during the red carpet pre-show, sandwiched between an interview with a beautiful actress and an attempted interview with a solipsistic director, Nate Silver, in his flat, awkward tone, appears and tells us all what is about to happen, and is right about most, if not all, of it. At best, the evening becomes rooting for or against (most likely against) this 35-year-old pundit getting things right; at worst, we shrug, flip the channel, and check Twitter to make sure we don't miss any Kanye/Taylor Swift-esque moments.
In other words, Silver himself would require a spoiler alert. That ABC would do this to one of its most important telecasts—last year’s Oscars scored 39 million viewers—tells you something about the direction in which the culture is heading. ABC is apparently betting that a savvier, more meta, and altogether more sophisticated Oscars experience is more appropriate than the old way of doing business, in which the drumroll that accompanies the opening of the envelope conveyed genuine suspense as opposed five seconds of filler preceding the announcement that Daniel Day-Lewis won, of course.
As for Silver, why would he want to apply his skills not only to important topics or to topics where his skills do not really ruin the suspense, but to something as irrelevant and hum-drum as the Oscars? The answer, I presume, is that he genuinely believes (and I do not mean to mock this belief) that he can make the world a better place by getting people to think about the future in terms of forecasting probabilities, and that the Oscars are a prime venue in which to further this goal.
Silver lays out his ambitions with a striking grandeur in the introduction to his book. It does not touch on on-base percentage or battleground state polls; it begins with the invention of the printing press and concludes with the era of Big Data. “There are entire disciplines in which predictions have been failing, often at great cost to society,” he writes. He argues that “information overload” absent widely accepted and effective ways of processing that information has led to many ills, from partisan deadlock to financial collapse to, at various times in the past several centuries, mass bloodshed.
His proposed solution? “Regulation is one approach to solving these problems,” he replies. “But I am suspicious that it is an excuse to avoid looking within ourselves for answers. We need to stop, and admit it: We have a prediction problem. We love to predict things—and we aren’t very good at it.”
The Signal and the Noise is suffused with an unstated but unmistakable urgency. It is, one senses, a part of his self-appointed project to make everyone, from the experts on down, better at forecasting, and at thinking about the future. The Oscars seem like a ripe platform from which to continue that fight.
Unless they aren’t. Silver should tread cautiously. Viewers who never cared about who wins an Oscar will not be affected—except, perhaps, to wonder why this nerdy guy is talking about prediction models when he's surrounded by beautiful celebrities. And what about those viewers who care more about the winners than the red carpet? Well, they may just stop watching, lest Silver ruin the fun.