“It’s hard to make predictions, especially about the future.”- Yogi Berra
Like so many people who blog about elections, Nate Silver has a penchant for making predictions. What separates Silver from many of his colleagues and competitors is that he is right. A lot.
In the 2008 presidential election, Silver successfully predicted the victor in 49 states. In the 2012 presidential contest, he picked the eventual winner in all 50 states. Silver did this without access to a crystal ball or the services of a high quality fortuneteller. What he does have access to is data, and he knows how to use it.
That so many other prognosticators have little idea how to effectively use data becomes obvious when reading Silver’s book, “The Signal and the Noise: Why So Many Predictions Fail – But Some Don’t.” In the book, Silver branches out from politics and describes how predictions can go bad in everything from sports to earthquake forecasting.
Unsurprisingly, the culprit in bad predictions is rarely the data. Instead it is often what accident investigators refer to as “operator error.” According to Silver, all too human foibles such as bias, hubris, and good old-fashioned mathematical mistakes, often sink predictions.
While Silver doesn’t provide a paint-by-the-numbers framework for how to make a good prediction, he effectively argues that people often predict what they want to happen, not what the data says is most likely to happen. Of course, anyone who watched Karl Rove continue to predict that Mitt Romney could still win Ohio late on election night already knows that. Still, it remains enlightening to hear it from a writer of Silver’s stature and talent.
It is also good to hear that we are not doomed to a future full of bad predictions. For readers ready to lose faith in the power of statistics, Silver provides a number of examples of how our predictive powers have improved with the advent of new models. We’re indisputably better at predicting the weather than we used to be, and baseball teams are better able to predict which players will succeed in the big leagues.
This may be little comfort to people who wince at the thought of all the cable TV pundits who continue to make incorrect predictions and face few consequences for it. But once we accept that such predictions are simply background noise that should be tuned out in the quest for accuracy, we may finally get on the right track. According to Silver, the old Doris Day song is wrong: The future is ours to see, we just need to know where to look.