{"text":[[{"start":5.2,"text":"For most of the World Cup game on Wednesday, the wisdom of crowds looked absent. Prediction sites gave England an overwhelming chance of beating the Democratic Republic of Congo. Yet the men in white were still behind until 74 minutes in. Earlier in the week, Germany had been favourites before losing to Paraguay and the Netherlands were unexpectedly defeated by Morocco."}],[{"start":29.2,"text":"Away from the football (don’t worry, this won’t take long), the consensus was dead wrong about a weaker dollar this year, too. Oil forecasters are also at sea. Meanwhile, despite US equity valuations at bonkers levels, investors are sanguine."}],[{"start":46.25,"text":"Where does wisdom go sometimes? Why do markets that supposedly aggregate the knowledge of millions of people — or even a few clever ones — repeatedly go awry?"}],[{"start":56.75,"text":"The idea that crowds are intelligent dates back to 1906. Francis Galton, the Victorian statistician, attended a country fair where nearly 800 people paid to guess the weight of an ox. Individually, the results were mediocre. Collectively, though, the median estimate was within 0.8 per cent of the true weight. "}],[{"start":78.8,"text":"That simple observation spawned more than a century of research. From guessing the time to the number of jelly beans in jars, the wisdom of crowds really works. It also underpins today’s prediction markets, where millions of contracts are traded (don’t call it betting!) on everything from sporting events to elections."}],[{"start":98.1,"text":"But Galton’s experiment is widely misunderstood. The lesson was never that large numbers of people are always wise. It was that crowd-based forecasting is accurate only under specific conditions — the most crucial one being independence."}],[{"start":114.94999999999999,"text":"Imagine asking those same fairgoers to estimate the weight of an ox after they’ve all overheard the same expert, read the same reports and chatted among themselves. Suddenly the mathematics change. Errors no longer cancel out. Rather, they reinforce one another."}],[{"start":133.75,"text":"In one celebrated experiment by mathematician Jan Lorenz and colleagues in 2011, people first guessed quantities independently. They then saw one another’s answers before revising their own. As you might expect, confidence increased and estimates converged. But accuracy often failed to improve. "}],[{"start":151.05,"text":"Economists describe a related phenomenon as an information cascade. Once enough people appear to believe something, it becomes rational for everyone else to follow, even if their own views point elsewhere. John Maynard Keynes recognised the same dynamic nearly a century ago when he likened investing not to choosing the prettiest face, but to guessing which face everyone else would think everyone else would choose."}],[{"start":175.70000000000002,"text":"This helps explain why consensus forecasts often fail. Foreign-exchange strategists don’t independently discover that the dollar should weaken. They consume much of the same data, build similar models, attend the same conferences and read one another’s research. Sports bettors watch the same injuries, statistics and odds screens. Ditto equity investors and earnings calls. "}],[{"start":200.55,"text":"Still, England eventually won. Germany would probably beat Paraguay more often than not. Crowds remain extraordinarily good at aggregating information. The mistake is believing that a consensus forecast is necessarily the most informative one."}],[{"start":217.70000000000002,"text":"In fact, investors can learn a lot by looking in the opposite direction. Take those wayward forex strategists again. If 99 of them reckon the dollar will drop 5 per cent and one predicts a rise of 20 per cent, the consensus is for a modest decline. But which forecast should you analyse further? Almost certainly the outlier. "}],[{"start":239.60000000000002,"text":"We spend our whole lives having the love of dispersion beaten out of us — at school and as employees. But the wisdom of crowds only succeeds when lots of people make different mistakes. Evolution has the same playbook. Natural selection couldn’t improve a species if every organism was identical; variation is the raw material from which progress emerges. "}],[{"start":262.25,"text":"Galton’s crowd at that country fair would have looked homogenous. But it was accurate because farmers, butchers and doctors drew on different fragments of knowledge. Ten 50-year-old white men who have spent their lives in different industries can bring more independent information to a question than a group who may look more diverse but who all work in the same sector."}],[{"start":282.4,"text":"The wisdom of crowds has never lain in the consensus. It lies in the disagreement. Once everyone thinks alike, crowds are error-prone. Polymarket gives France an 80 per cent chance of winning its match on Saturday night. Just saying."}],[{"start":304.2,"text":""}]],"url":"https://audio.ftcn.net.cn/album/a_1783139950_7871.mp3"}