Making iffy predictions about our future: it’s been a hallmark of human behavior for at least as long as we’ve been recording our present.
Perhaps the most famous early example of forecasting took place in Mount Parnassus in Delphi, Greece. From roughly the 8th century BCE to the 4th century BCE, for nine days each year, travelers could pay to pose their burning questions to the Pythia, a priestess said to channel the wisdom of Apollo.
To a modern reader, perhaps a bit skeptical about the whole Greek gods thing, it may be unsurprising to learn that the Pythia phrased her answers as riddles, arguably granting her some convenient wiggle room.
For instance, as the story goes, King Croesus once asked a Pythia if he should attack the neighboring Persian Empire. "Cross the border and a great empire will fall,” she told him. Encouraged, Croesus brought his troops over the border—and suffered a crushing defeat. Her reaction was to coolly point out that he had failed to make her specify which empire.
(Anyone who has ever spent an hour playing Simon Says with a pedantic seven-year-old can imagine how Croesus must’ve felt.)
Why would the ancient Greeks, the very people who brought us the root word for “logic”, buy into a riddle-based forecasting system? In a Guardian.com book review for Michael Scott’s Delphi: A History of the Centre of the Ancient World, writer James Davidson draws a direct parallel between the Pythias of yesteryear and a fund manager of today: “someone who gets paid vast sums for divining the future even though their well-informed bets produce slightly worse results than the stock-market average.”
Statistician Salil Mehta, who actually took the time to crunch the numbers, has emerged with an even grimmer picture of these modern day stock soothsayers. “It’s not easy to be as bad as they are,” Mehta told a New York Times reporter in 2016. “They are much worse than random chance alone would predict.”
And yet stock forecasting remains a lucrative career. TV news networks continue to employ the same political pundits who spectacularly failed to predict the 2016 election cycle. In a 1993 survey covering the 50 largest cities in the U.S., one third of police departments reported that they had accepted predictions from “psychic detectives.” And no matter how many times a group of wine experts fail the simplest of tests—for instance, even noticing that their “red wine” was simply a glass of white with dye in it —you still tend to believe your wine snob buddy when he insists he can detect notes of oak, moss, and autumn leaves in his Cabernet.
In 1980, writing for Technology Review, J. Scott Armstrong of the University of Pennsylvania Marketing Department coined a somewhat cynical term to describe this phenomenon of clinging to false prophets: the “seersucker effect.” As Armstrong puts it, “No matter how much evidence exists that seers do not exist, suckers will pay for the existence of seers.”
The 37-year-old paper is more than a little dated—Armstrong cites one study that measured the ability to predict if a given stranger was “heterosexual or homosexual,” and compared a layperson’s guesses to those of a “relevant expert” (Armstrong never explains what field of study this would even be.) He also had no way of knowing that in 2017, technology would allow for at least one type of hyper-accurate seer: as Nate Silver has noted, meteorologists continue to improve their predictions all the time.
Still, Armstrong provides some useful insight into why, millennia after the fall of Ancient Greece, we are still putting our faith in fortune-tellers: it’s not just about wanting to trust someone, he suggests, but the desire to offset some of the responsibility of a decision. Even when we know that third party could be fallible, we want to be able to say we went with the expert verdict. That way, if the results are disappointing, we can absolve ourselves. Sometimes, we want somebody else to take the wheel so badly, we don't pause to check whether or not they can drive.
If there’s a little King Croesus in all of us, maybe all we can do is remember to ask some key follow-up questions...