This article is taken from the June 2025 issue of The Critic. To get the full magazine why not subscribe? Right now we’re offering five issues for just £25.
Is divining the future a skill that can be taught? That’s what a University of Pennsylvania psychology professor, Philip Tetlock, proposed in his 2016 book Superforecasting: The Art and Science of Prediction, co-authored with journalist Dan Gardner.
Nine years later, more people around the world are making predictions than ever, on topics ranging from stock prices to geopolitical events — but Tetlock’s book is now largely forgotten. The world has moved on.
With its promise that it “can make you smarter”, Superforecasting is a descendent of the specifically American genre of business self-help books that marry career advice to personal development, an unfashionable formula revived in a much more respectable form in the 2000s by Malcolm Gladwell.
The most notorious of the self-helps is Dale Carnegie’s 1935 bestseller How To Win Friends And Influence People, a manipulation manual which enjoyed an 80-year print run and gathered devotees including the serial killer Charles Manson.
Tetlock emerged just as the reputation of professional soothsayers began to decline. His book Expert Political Judgment: How Good Is It? How Can We Know? (2005) argued that professional soothsayers performed no better than “dart-throwing monkeys”. Similarly, James Surowecki’s The Wisdom of Crowds, published a year earlier, had argued that groups made better guesses than an individual.
Faith in experts fell further with the 2008 financial crash. Prediction markets, in which participants bet on a forecast, were beginning to emerge online. Superforecasting promised much, but the steak never really matched the sizzle. “I believe it is possible to see into the future, at least in some situations and to some extent, and that any intelligent, open-minded and hardworking person can cultivate the requisite skills,” wrote Tetlock. 
If this was hardly earth-shattering, the book’s subtitle is also misleading. Tetlock was not revealing a new algorithm or mathematical technique. No new “science” was imparted.
Superforecasting was instead praxis: a generalisable approach to studying information and creating an environment in which clever people do more reliable work. The banality of his advice left established practitioners scratching their heads.
Asked why his superforecasters apparently yielded superior results, Tetlock explained they “viewed forecasting as a skill that required deliberate practice, sustained effort and constant monitoring of current affairs”.
The occult-like powers attributed to Tetlock now seem faintly ludicrous
So be open-minded, read widely and try harder. Isn’t that just what every competent strategic planner and intelligence specialist had been doing? Nonetheless, he gained some high profile endorsements, including one from Dominic Cummings, Tetlock’s very own Charles Manson. But in reality, Tetlock’s emphasis on practice and diligence was already being overtaken, and this accounts for why his turn in the spotlight has not been an enduring one.
“Superforecasting has had its day,” says Alan Patrick, an award-winning former McKinsey consultant turned data science forecaster who correctly called the result of the EU Referendum and Trump’s first presidential election victory with the pioneering use of social media analysis. “It’s as old as the Ides of March and about as effective.” Superforecasting, he explains, has been overtaken by both people and machines.
“Tetlock mainly comments on human-centric systems,” Patrick explains. “The best pattern-spotter is still the mark one human eyeball. Machines don’t have the ability to see the big picture — but humans are not scalable; you can’t run them 24 hours a day, 7 days a week”.
So most prediction systems now are a hybrid of the two.
Google Trends shows that interest in forecast hybrids emerged in 2022 and Superforecasting disappeared entirely by 2024. In the background lurks artificial intelligence and the underlying analytical potential of machine learning. The latter has been incorporated into models ranging from sales forecasts to climate predictions.
Tetlock also found himself rubbing up against the big beast of the jungle (and former collaborator) Nicholas Nassim Taleb, whose Black Swan (2007) had turned the subject of probability into a global bestseller.
Both Taleb and Tetlock agreed experts could become clouded by their biases and needed “skin in the game”. In other words, self-interest led to better results. But Taleb’s speciality was warning of low probability high impact events, such as a pandemic. After the world experienced one of these, Taleb seemed the better bet.
Human judgement-based forecasting is really no better or worse today, but the occult-like powers that a figure like Dominic Cummings attributed to Tetlock now seem faintly ludicrous. Cummings has moved on too. His obsession with finding “a colonial government of Dark Wonks to sort out the complicated stuff backstage”, as the J’accuse Substack describes it, has found new candidates to promote.
Ultimately, Superforecasting was undone by its blandness — or if you prefer, commendable modesty — and in a world of buzz and hype, that’s not enough.
