Go to download electronic copy of book: $19.99
See book available for shipping from same merchant: $14.99
Check their physical store for same book: $27.99
After taking all this in, I did end up doing the ebook version of Nate Silver’s “The Signal and the Noise,” though largely because I wanted to start reading without waiting for it to be shipped.
Silver progresses through a series of topics building a case for improving predictions and models by largely being as honest as possible with the process. He highlights the need for good input data and especially in expressing results with degrees of confidence. As he argues, it may get more headlines to give an emphatic yes/no kind of pick, but everyone is better served if you honestly say there’s an 85 percent yes/15 percent no chance of whatever happening.
He focuses one chapter on economists’ forecasts for Gross Domestic Product, those periodic releases of data on how the economy is doing. So many of the picks come out as just a number, like 3 percent growth next quarter. But Silver says those picks tend to have what is basically a 3.2 percent margin of error, meaning a 3 percent target could in reality turn out to be 6.2 percent or -0.2 percent, which is a pretty significant difference.
To get the best use of a GDP forecast, Silver argues that perhaps we should be reporting them with margins of error just as we do with political polls.
“Danger lurks, in the economy and elsewhere,” he writes, “when we discourage forecasters from making a full and explicit account of the risks inherent in the world around us.”
The most fascinating chapter of the book for me is about weather forecasting. It’s no secret that people love to make fun of the profession, but perhaps because of my personal relationship with some meteorologists, I find myself being more of a defender. Silver points out that weather forecasts have gotten steadily better every year, and have dramatically improved in the past 10 or so.
But he brings up one thing that will truly make me see forecasts differently, and that’s how various outlets will talk to you about the chance of rain. Silver says a National Weather Service forecast of 20 percent chance of rain really does play out that often, while the Weather Channel will say 20 when it actually only rains 5 percent of that time.
Why?
“In fact, this is deliberate and is something the Weather Channel is will to admit to,” Silver writes. “It has to do with their economic incentives. People notice one type of mistake — the failure to predict rain — more than another kind, false alarms. If it rains when it isn’t supposed to, they curse the weatherman for ruining their picnic, whereas an unexpectedly sunny day is taken as a serendipitous bonus.”
Silver also talks about the challenges and risks of judging forecasts that may be what he calls “self-defeating.” That is, a forecast that end up affecting itself and thus not coming true.
“The most effective flu prediction might not be one that fails to come to fruition because it motivates people toward more healthful choices.”
And yet, how often do we see people throw up a prediction about something like flu season and say “SEE! SEE HOW WRONG YOU WERE!” More need to talk about ranges of outcomes and think about why things turn out the way they do.
Another chapter on his baseball model, called PECOTA, made me laugh and drop into a deep baseball-less depression. Silver really became first known for developing PECOTA, and among other things he used it to project how minor league players would perform. He says in the book that his model was optimistic about future stars like Ian Kinsler and Matt Kemp.
“But have you ever heard of Joel Guzman? Donald Murphy? Yusemiro Petit? Unless you are a baseball junkie, probably not. PECOTA liked those players as well.”
Yes, Nate, I HAVE HEARD OF YUSEMIRO PETIT. Granted, this book came out two years ago, but just last month I sat freezing in Nationals Park as the San Francisco Giants outlasted my beloved Nats 2-1 in an 18-inning game that was the longest in MLB postseason history.
Petit pitched six innings in relief for the Giants that night, allowing only one hit and earning the win as the Giants grabbed a commanding 2-0 lead in the best-of-five series.
I guess I can’t hold Silver responsible for the emotional effects of his forecast coming true. This is a great book for those interesting in modeling, data or just thinking about how we talk about the world around us.