Predictions on what the world will look like, is as old as time itself.
According to the ever reliable Wikipedia, Michel de Nostredame (aka Nostradamus) is said to have made 6,338 predictions in his almanacs. Of those predictions, a number have been proven to be ‘correct’. A quick search on YouTube will often show that The Simpsons is equally good at predicting the future. For example, a video essay by ScreenRant shows that a lot of what has been shown on The Simpsons have ‘come to pass in 2020.
How accurate are people's predictive faculties?
One of the most famous studies on prediction was done by political scientist, Philip Tetlock in his 2006 book Expert Political Judgment. In his research, Tetlock asked a group of pundits and foreign affairs specialists to predict geopolitical events. Overall, the “experts” struggled to perform better than “dart-throwing chimps”. They were also consistently less accurate than simple statistical decision making models. This was regardless of creed or credentials.
Another experiment which seems to indicate that expert humans aren’t as great as their ego thinks on their predictive ability is by economist, Burton Malkiel in his book A Random Walk Down Wall Street. Testing the random walk hypothesis with his students flipping coins, Malkiel concluded that the future direction of the stock market could not be predicted by its past. That “a blindfolded monkey throwing darts at the stock listings could select a portfolio that would do just as well as one selected by the experts”.
Setu Mazamdar recreated Malkiel’s experiment by coin toss and compared the chart to that of the actual Dow Jones Industrial Average.
The Tools To Predict
Whether it be the hedge fund managers or Movie Studio Execs trying to predict and make a bet on a film’s likely success, data seems to suggest that ‘experts’ aren’t generally very good at predictions. In most attempts of predicting what will happen, with streaks of successes, the human guesses are not too different to statistical estimates of a stochastic process.
A larger amount of data and a new array of techniques plus technology may make us better at prediction. Cliodynamics and applying machine learning to large datasets (similar to DeepMind’s success with AlphaFold in protein folding) may of course change this.
🎯Predictive Hits & Misses.
With Tesla’s soaring stock prices and investors following the thesis of Tesla being more of a tech company than car company, we can look back into past predictions. The cheery GM Motorama Exhibit of 1956, shows the Firebird ‘self driving’ (“safe, cool and comfortable”) car as being coordinated with control towers (like airport terminal) and running on magnetic car strip roads.
The GM ad reflected the fascination of the 1900s with aviation and automation. Cars were influenced by turbines. It was imagined that in the 2000s, automobiles would be replaced by flying contraptions.
One leading German chocolate company, Hildebrands, made a series of commemorative postcodes and speculated that people would take to the skies. They predicted that people would travel by personal airships and gliders, that buildings would be movable (aka Howl’s moving castle) through trains, and that cities would be encased in glass to be protected from the weather elements.
In the 1920s, people were fascinated with skyscrapers and the term ‘robot’ was conceived. The dream was of a future of convenience and luxury.
The 1960s and 70s saw an even more futuristic prediction. With the US and USSR competing in a space race, people dreamt of humans as space colonisers. NASA’s founding in 1958 propelled the U.S. into being the first nation to land a person on the moon in 1969. The period saw the emergence of Googie building design in the U.S. and Hollywood imagined what an intergalactic species would look like through gripping films like Star Trek (1966) and Star Wars (1977).
In the 1980s and 1990s, computers became more prevalent and people predicted a digital and physical reality. Hollywood hits like Terminator (1984) and Back To The Future (1985) conveyed potentially existential technology risks to society. Capitalising on the hype for immersive digital experiences, Nintendo launched the Virtual Boy headset in 1995, an attempt to offer gamers a 3D experience. It was a flop - selling only 770,000 units and remains Nintendo’s biggest fail.
One of the best predictive bets
Whilst there have been some very good predictions and bets such as Masayoshi Son’s $20M investment in Alibaba which has returned SoftBank about 5000x (~$100B) There have also been a number of flops (by the same guy) and others.
Some of the worst predictive fails
Microsoft’s former CEO Steve Baller in 2007 thought that the iPhone
"[Apple's iPhone] is the most expensive phone in the world, and it doesn't appeal to business customers because it doesn't have a keyboard, which makes it not a very good email machine…"
John McAfee betting to eat his ‘d*ck’ in 2017 on what the price of Bitcoin will be in 2020