On Topic: Our best guesstimates as told by forecaster Nate Silver
One of the ways you can tell something is a great idea is when, in hindsight, we smack our foreheads and declare, gosh, why didn’t someone else see that coming?
Nate Silver says lots of forehead-smack-worthy things in “The Signal and the Noise: Why So Many Predictions Fail — But Some Don’t.” His book offers shrewd insights for those interested in predicting the markets, unemployment, political horse races, terrorist attacks, earthquakes or the World Series.
You probably heard of Silver at the tail end of last year’s presidential election — he’s the clever guy who correctly forecast the red-or-blue outcome of every state. (He was right about 49 states in 2008.)
A one-time statistician for KPMG, Silver’s turned his hand to calculating the performance of baseball teams, poker — where he did exceptionally well, thank you — and, more recently, political campaigns on his FiveThirtyEight.com site. His book’s premise is that most talk-show pundits, scholars and — sigh, I saw this one coming — journalists don’t really have the slightest notion what the heck they’re taking about.
Reporters, Silver remarked during a November 2012 interview for Authors @ Google (you can look it up on YouTube), tend to write as if the poll cited in their story is “the only poll in the world” and rarely mention any of the dozen other samplings that might conclude entirely different propositions.
And pollsters, especially for political campaigns, “when they talk to the press put things in the most favorable light,” he commented. Narrative — like TV sports coverage — is what they’re peddling.
In his book, he examines, among other things, how the Federal Reserve was blindsided by the housing bubble and the subsequent recession. In fact, a quarterly poll released in November 2007 by the Fed in Philadelphia anticipated economic growth in 2008, at a tad under 2.8 percent
Here’s why so many predictions are worthless:
If I asked you to forecast the total that will be produced when you roll a pair of six-sided dice, the correct answer is not any single number but an enumeration of possible outcomes … . Although you will roll 7 more often than any other number, it is not intrinsically any more or any less consistent with your forecast than a roll of 2 or 12 … over the long run.
That is to say, the range of possible outcomes over the course of time can be a pretty wide swath, and telling the future isn’t as simple as picking a solitary magic number.
How often have these economic tea-leaf readings been wrong? Forecasts from 1968 to 2010 by “experts” on the actual value for GDP, Silver demonstrates, “fell outside the prediction interval” almost half the time.
And he has charts to prove it.
The National Weather Service, he contends, has become far better at recognizing “the importance of communicating uncertainty” than economists.
Silver surely is right about how much salt we need to ingest with these prognoses, for politics or the future of our economy. The more numbers involved, the broader spectrum of probabilities we need to bring on board.
When I was about eight or so, my Uncle Scuppy — yes, that really was his name — used to take me with him into his favorite corner tavern so he could place bets on my being able to correctly guess coin tosses in exchange for free shots. (He’d drink the shots, I’d keep the pennies. I had an interesting childhood.)
My success was down to one naïve principle: I’d determined the odds each time were 50-50, so I called “heads” repeatedly, only occasionally opting for “tails.”
We’d do reasonably well, until some bleary soul would figure out my method was not so much sophisticated statistics or psychic skill but more like tenacious optimism. (Or until one of my parents would track us down.)
What Silver seems to want us to learn most of all, in his oft-stated goal to make us more data literate, is that we foolishly “think we are better at prediction than we really are.” This new millennium, after all, has been rocked “with one unpredicted disaster after another.”
It’s just not as easy as guessing heads or tails, up or down, yes or no. In the real, adult world of economics — or politics, sports or climate science — we need to consider the much bigger, more complex picture.More is better, right?