Office For Anticipating Surprise [1]

After twenty years of research, Tetlock's first conclusion was that it is impossible to make long-term predictions. Indeed it is he who discovered that a chimpanzee at a dartboard is as successful at picking stocks as experts on Wall Street. But in the summer of 2010 he received a visit from IARPA. It wanted to hold "forecasting tournaments," but for short-term predictions. Victory followed soon thereafter as Tetlock's forecasters beat the competition by margins of over 60%—without the help of computers. Tetlock had finally found the kind of people he'd long been seeking: the "super forecasters."

During the lunch break the time finally seems right to pose the question: is it ethically admissible for scientists to work for an institute like IARPA? It soon becomes clear that a European journalist alone is apt to find that problematic. It turns out that many here have also worked for the Defense Advanced Research Projects Agency (DARPA), which is the research arm of the defense department and the model for IARPA. IARPA is run by the Office of the Director of National Intelligence.

According to Peter Lee [2], now head of research at Microsoft, IARPA is only the best-known attempt to imitate the success of DARPA. The latter was founded in 1958, after the Soviet launching of Sputnik, as the Advanced Research Projects Agency (ARPA) in order to give the U.S. an edge in the technological race between the superpowers. Its successes have made scientific history. This is where the rockets were developed that later flew to the moon, as well as the first version of the Internet, of GPS, the first drones and the first self-driving cars. For American scientists, according to Lee, DARPA has always been the institute of unlimited possibilities.

Nobel Laureate Daniel Kahneman also worked for the military, but in Israel. After the shock of the Yom Kippur War of 1973 he established a military prediction team. This is why Tetlock's work interests him: it provides an opportunity to introduce scientific standards into secret-service work. Until now, he says, secret-service agencies drafted their reports as essays. You couldn't ask for a less precise approach.

Edge_Master_Class.jpg
In the circle of clairvoyants: At a vineyard north of San Francisco, Philip Tetlock of the University of Pennsylvania (left) presented his findings. Initially skeptical was Nobel Laureate Kahneman (third from left). Photo: John Brockman / edge.org

Can science ensure that reason prevails?

Understanding this is fundamental to understanding the "super forecasters." Their most significant characteristic is the capacity to approach any conceivable topic without ideological or emotional baggage: scientifically. That is not easy in a polarized society such as America or Europe. In America, after all, the years after 9/11 were marked by a series of irrational feats such as the Iraq war, the societal control involved in the Patriot Act, and the surveillance society under the watchful eye of the NSA. But in Europe too the largest crises are now discussed, even at the highest levels of government, with a great deal of emotion and ideology—from the conflict in Ukraine to the economic collapse of Greece to the deficiencies of the European Union. Predictions, however, are not concerned with political or ethical classifications. Rather, their objective is to determine the facts so precisely that they allow for a glimpse into the future and, above all, clear-eyed decisions.

At the end of the weekend Philip Tetlock explains once more in detail why there is also a civilian application for the work of the "super forecasters," as there once was for the Internet or self-driving cars. Over the long term, he says, their methods and mindset could take the sting out of public discourse. If the traditional opinion leaders, the commentators and columnists, were faced with "forecasting tournaments," he believes their supposed expertise would soon be exposed as opinion. In any case, he finds, an analytical approach would constitute an improvement.

Daniel Kahneman, too, is convinced in the end. If we were able to teach "super forecasting," he says, it would gain a foothold in the scientific community. He only sees one obstacle: there's a long way to go from attaining scientific to attaining political relevance. A very long way. But Tetlock's methods will only be effective, he believes, when they improve the quality of political decisions. Yet perhaps that isn't so essential. Perhaps it's already an important step that science is forging the ideal of a post-ideological era in which reason can prevail.

[First published in German by the Süddeutsche Zeitung in the weekend Feuilleton, August 8, 2015.