The pollsters got it wrong. Math did not predict the 2016 election. Nate Silver fell from grace. So how do we explain all that? Look no further than the mighty burrito.

By Andrew J. Pridgen

It did not take long this election season for me to figure out there was a fatal flaw in boy wonder Nate Silver’s once impenetrable fortress of numbers and analytics.

Silver, the cherubic data guru for The New York Times, signed a contract with ESPN shortly after he called the 2012 election for Obama with pinpoint accuracy and cheekiness. A vertical dedicated to divining sense and reason from the chaos of human error then sprouted up in the form of fivethirtyeight—the name representing the total number of Electoral College votes.

The site touts itself as a user of  “analysis—hard numbers—to tell compelling stories about elections, politics, sports, science, economics” and, well…burritos.

That’s right, one of Silver and co.’s first projects post-launch was a full-scale, mind-numbing, gut-bursting breakdown of the best burritos in the land. The problems, immediately, were glaring based mostly on the methodology of judging.

The criteria in fivethirtyeight’s words with my commentary in bold italics below.

  1. Step 1: Data mining. Analyze the Yelp data to create an overall rating called Value Over Replacement Burrito (VORB) and provide guidance for the next stages of the project. (This step is already done, and I’ll be describing the process in some detail in this article.) Yelp isn’t as much data as it is a place to mostly anonymously pursue vendettas against local businesses online. Yelp should be used, at most, as a cross-referencing tool or a repository of who’s open latest should one be forced to layover.
  2. Step 2: Burrito Selection Committee. Convene a group of burrito experts from around the country, who will use the VORB scores and other resources to scout for the nation’s best burritos and vote the most promising candidates into a 64-restaurant bracket — 16 contenders in each of four regions: California, West, South and Northeast. (The committee has already met, and we’ll reveal the 64 entrants in a series of articles later this week and this weekend.) Copying a NCAA-style basketball tournament bracket is so very ESPN of them, but there really is such a disproportionate bend toward excellence in the West not represented in this case.
  3. Step 3: Taste test. Have Anna Maria Barry-Jester (the chief burrito taster) visit each of the 64 competitors, eat their burritos, rate and document her experiences, and eventually choose one winner in a multi-round tournament. (Anna will be posting her first reviews early next week. She’s worked as a documentary photographer and multimedia journalist, and as a producer at ABC News and Univision, where she’s spent years reporting on Hispanic-American culture.) This may have been the project’s ultimate undoing. Silver’s burrito-tasting team was a bunch of waspy Duke, Middlebury and Vassar grads, most of whom hailed from the Northeast and Midwest. The most insulting factoid of all, and the most damning to the endeavor, is the throwaway line that judge Anna spent “years reporting on Hispanic-American culture.“ Ouch. I’m not saying the talliers needed to be line cooks at Filiberto’s, but for crying out loud, just one person on the panel who grew up in board shorts and flip flops riding beach cruisers to Mama Testa’s would have done well better that rubes from Poughkeepsie (no offense to El Bracero lovers.)

These aforementioned flaws, plus the mildly offensive use of the Spanish language when points needed to be made (crowning the winner, “El Gran Campeon”—ugh) basically revealed the whiteness not to mention the lack of objectivity and proliferation of regional bias in the endeavor.

Silver’s crew ultimately selected La Taqueria from the Mission in SF as the winner. “I watched Nate in his Brooks Brothers oxford take his first bite with that snively Niedermeyer look, and I swear he achieved nirvana before my eyes as if he’d just returned from a TA’s office hours turning a C+ to an A,” BRTOLUVR Anna wrote.

Granted La Taqueria, which has been in the Mission since 1990, is a reasonable choice, but it is also relatively negligible and +/- with any number of neighboring spots: Taqueria Cancun, Papalote Mexican Grill, Pancho Villa and even Taqueria Cazadores, a taco truck.

Fivethirtyeight staff, however, showed their East Coast sensibility (and naivety) praising La Taqueria’s annoying and obvious use of sour cream. There aren’t many native California burrito connoisseurs who choose to glob the catsup of spoiled dairy on their concoctions when given the choice.

While I didn’t see fivethirtyeight’s site numbers in the wake of the effort, the burrito undertaking seemed to end more like a vanity project gone wrong than a tone-setting way to vet small restaurants and rate the most delicious delicacy that ever graced the inside of tin foil. So be it. Not all ideas are good ones. No real harm done.

But the project was something that I thought about multiple times when the Silver team seemed to be so assured of their numbers in what was otherwise a random mess of unpredictability and uncertainty during the unprecedented run up to Election Day. Something about Silver and co.’s ham-fisted belief in what they were selling—especially as I reached out personally to would-be Clinton voters in swing states—just seemed very, very off.

…Or at least it was a very far cry from the relative calm and sensibility of the voting public in 2012 when Silver and his team called 49 of 50 states correctly.

And it was.

Whether hubris is to blame or because the fivethirtyeight team really didn’t have the proper way of seeing analytics this time around, Silver and co. were dead-ass wrong, like the rest of the media, at every step, about nearly every region.

The problem I have is that that he won’t yet admit he kept selling super subjective information—like how a burrito tastes—as gospel. Last September, he told CNN that Trump had a 5-percent chance of getting the Republican nomination and doubled down shortly after with a “Donald Trump’s Six Stages of Doom” piece. In it, Silver actually laid out a schedule for the candidate’s inevitable collapse.

He also did a mea culpa after the nomination which is worth reading here—admitting in May that his own ego may have played a part in his inaccuracies up to that point. He even speculated whether there was too much “#datajournalist self-flagellation” in this election. And yet, after that, he and his team seemed to return to their old ways, relying on what turned out to be very skewed, very wrong polling systems.

In the pukey burp backwash of the election, Silver defended his final tallies saying as of Tuesday morning his site gave Trump a 29 percent chance to win by a narrow Electoral Collage margin. He said that percentage was more “for Trump” than just about any other prediction including oddsmakers who kept a Trump victory around 18 percent.

Even if that is true, he was way, way, way off—by a margin of more than 70 percent if you look at it strictly by the numbers.

For his errors, he cited a number of factors including late- and undecided voters breaking for Trump, which I personally can attest to after doing a little last-minute phone banking on Monday. Nobody on the fence among registered Democrats seemed energized to go out and punch a chad for Hillary, and it showed in the final numbers. In spite of a campaign that captivated a nation and made the world hold its collective breath, netting bonanza ratings and clicks for networks and news agencies, the voter turnout on Election Day was the lowest in two decades.

In other words, people who could not bring themselves to vote for either candidate simply stayed home in protest.

I recall a prescient statement on Nov. 2 that The New York Times columnist Thomas Friedman made to voters in his final appeal to reject Trump. Reading it in hindsight it almost feels as if Friedman knew the inevitable was coming. He wrote, “No one knows for certain how we deal with this new race with and against machines, but I can assure you it’s not Trump’s way—build walls, restrict trade, give huge tax cuts to the rich. The best jobs in the future are going to be what I call “STEMpathy jobs — jobs that blend STEM skills (science, technology, engineering, math) with human empathy. We don’t know what many of them will look like yet.”

I believe this is completely true. The future is mixing science and technology with the very human, very unpredictable qualities we bring (including sometimes taking into account that we are assholes who don’t do the right thing just to see how big a hole we can dig.) In the case of Silver and his team, they are on the right track using math and straight data and logic to determine a course, but they haven’t really nailed the whole being able to read what people’s tastes are and how that factors in yet.

Not everyone, after all, likes sour cream.

Andrew J. Pridgen is the author of “Burgundy Upholstery Sky”.

NO COMMENTS