Dan Gardner is the New York Times best-selling author of Risk, Future Babble, Superforecasting (co-authored with Philip E. Tetlock), and How Big Things Get Done (co-authored with Bent Flyvbjerg). His books have been published in 26 countries and 20 languages. Prior to becoming an author, Gardner was an award-winning investigative journalist. More >

Some Predictions About Predictions

At the time of writing, the results of the American midterm elections are not known. And yet, I can predict three outcomes with near-metaphysical certainty. One, some pundits will have predicted the results. Two, all pundits will explain the results. Three, most of those predictions and explanations will be -- to use John Nance Garner's famous description of the vice-presidency -- "not worth a bucket of warm piss." Now, these predictions of mine may seem a little odd given that I have just published a book about the futility of prediction. (That's Future Babble, available at a low, low price wherever fine books are sold.) Still, I am confident I am right. Why? Because some things -- the sun rising in the East, leaves turning colour in the autumn -- happen with fixed regularity and thus are perfectly predictable. Among them are the items on my list. Start with Prediction No. 1: Some pundits will have predicted the results. You might think this is a contradiction of the basic point of my book, which is that the average expert who attempts to predict the future is about as accurate as a dart-throwing chimpanzee. It's not. I am certain that some pundits will have predicted the outcome not because pundits are good at predicting elections but because so many pundits make so many predictions that all possible outcomes have been covered. At one extreme, some pundits predicted the Republicans would sweep both chambers of Congress. At the other, some said the Democrats would retain control of both the House and Senate. In between, lots more pundits covered off all possible permutations. Short of a landslide for write-in candidates who form a third party and seize control of Congress, there is literally no outcome that has not been predicted. Hence, no matter what happens, someone will have predicted it. The only difference between this and a lottery is that sensible people don't think someone who wins a lottery "predicted" the winning numbers (Incidentally, if the election actually does produce a landslide for write-in candidates who form a third party and seize control of Congress, I will claim to have predicted it. Do you have any idea how much CNN pays its talking heads?) Now, on to Prediction Number Two: all pundits will explain the result. This is odd, when you think about it. After all, most pundits will not be able to say they successfully predicted the outcome. They analysed the situation. They told a story that explained how things would turn out and why. And they were wrong. Which would suggest their analysis was wrong. They should be flummoxed. But pundits are never flummoxed. They're pundits. They always know why things turned out the way they did. No need to reconsider or conduct further research. The answer is obvious! And cable news will be flooded with their wisdom. Why are pundits never at a loss for an explanation? It's seldom their penetrating insight. If their insight was so penetrating, it wouldn't be surprised by events so often. No, what's really at work is what neuroscientist Michael Gazzaniga calls "the Interpreter," a neural system in the left hemisphere of the brain which automatically and compulsively churns out stories to explain observed facts -- even when the person has no clue what the real explanation is. The experiments in which Gazzaniga demonstrated how easily people fool themselves with bogus explanations are deeply disturbing. But pundits should worry more than others. With masses of facts stuffed in their brains, and lots of practice generating hypotheses on the fly, they are even more adept than the average person at whipping up explanations that are clear, compelling, and wrong. And they do. Constantly. Don't believe me? Ask a political scientist. Methodical research routinely reveals the clever stories told by pundits are bunk. That's especially true of the TV commentators who sometimes literally say the first thing that pops into their heads. Hence, Prediction Number Three: Most predictions and explanations will be worthless. Consider Dick Morris. Watch cable news for any length of time and you're bound to bump into Bill Clinton's former pollster explaining events and predicting what is to come.TVloves him. He's a clear and compelling speaker. And he's supremely confident. Remember the epic 2008 battle for the presidency between Hillary Clinton and Condoleezza Rice? Morris called it. John McCain defeating Barack Obama? Morris again. And when Hurricane Katrina struck it was Morris who saw that it would send George W. Bush's approval rating soaring upward. And yet, when you turn on cable news, there's Dick Morris, as sure of himself as ever. And who can blame him? John Kenneth Galbraith said of economists that they forecast "not because they know but because they are asked." It's the same for political pundits. We ask. Dick Morris answers. Who's the fool in that relationship? Unfortunately, we aren't likely to wise up. As I explain in my book, we listen to expert predictions not because they're accurate but because they satisfy our psychological need for certainty. And so we will keep asking the pundits for their predictions and explanations, no matter how demonstrably worthless they may be. And yes, that's a prediction.