In 1976 J Doyne Farmer walked into a casino in Las Vegas and committed a robbery. Nobody noticed him doing it: to anyone else it would have seemed as though the lanky 24-year-old was simply playing a little roulette. Casinos watch carefully for cheating at card tables, but the spinning roulette wheel is effectively a random number generator for which the only strategy was to take a lucky guess. Farmer had found a way to beat the odds.
As the wheel span and the ball clattered, he began pressing switches hidden in his shoes, following a pattern he had spent hours practising at home. A click when the ball passed one reference point, a click when the wheel had rotated by a certain amount. Inside his shirt the world’s first wearable computer, built and programmed from scratch by Farmer and his friends, processed these inputs and buzzed his armpit with a code suggesting the most likely positions at which the ball would come to rest. The computer – which he called Harry, his best friend’s middle name – was able to beat the house, turning a regular profit from a game in which gamblers always lose.
Farmer smiled as he recalled the adventure from his office at Oxford University, where he is a professor of complex systems science. He is tall, with an athletic build; a high forehead and a long beard give him a slightly Darwinian look. He was, he tells me, “excited, scared, sweating a lot”, as he concentrated on the roulette system. It was not his first adventure at the edge of the law – as a student, he’d supplemented his income by digging a tunnel, more than 100 metres long, to smuggle motorbikes across the Mexican border – but the object was not to make money. Farmer and his friends were experimental mathematicians. They were demonstrating an understanding that he says has become more important than ever: that the world is more predictable than we think, and that the fundamental assumptions of modern economics (and therefore politics) are mistaken.
At his shoulder in the casino was the unseen demon conjured in 1814 by Pierre Simon Laplace, who wrote in his Philosophical Essay on Probabilities that for a sufficiently vast intellect, capable of knowing every fact about the universe and its implications, “nothing would be uncertain and the future just like the past would be the present to it”. Since then, philosophy students have wrestled with the idea that in a deterministic universe, the future is already set.
Does Farmer believe in free will? He has been thinking about this a lot recently; getting divorced has prompted him to consider how much people are really to blame for their actions. We are, he tells me, shaped by the “constitutive luck” of genetics and experience, and as a physicist he recognises that the universe runs on deterministic laws, but this does not mean the future is set, because the universe is also chaotic. “Chaotic means that there’s limits to how much we can predict. So in a sense, the fact that it’s deterministic is irrelevant; there’s uncertainty in the future, and that uncertainty is intrinsic.”
However, there is a difference between chaos and complexity. It may never be possible to see the future clearly, but complex systems can act in predictable ways. A human brain or an ant colony will display “emergent behaviour”, “emergent phenomena” will develop in a weather system. These are systems in which billions of things interact, and what emerges from those actions – a language, a storm – has entirely different properties to all the things that comprise it. Complexity science can recognise these patterns and simulate them. Farmer says the most important application of this view is in the field used to justify every political decision: economics.
His experiences with gambling and smuggling, he tells me, did not mean he had any love of money. He is fascinated by the economy and its markets as systems, but as he sees it, “the purpose of making money is to be able to ignore money”. This he did want, from an early age. He grew up poor, in a household he describes as “troubled”, in Silver City, New Mexico. “It was miserable to live at home,” he tells me. His parents moved to Peru when he was 14 and he went to live with his mentor, the physicist Tom Ingerson, whom he describes as “a genius, and a total non-comformist”; the book that he has been writing for the past decade, Making Sense of Chaos, is dedicated to Ingerson, who died in 2019.
After college, Farmer was reading a biography of J Robert Oppenheimer when a poster bearing the great physicist’s name caught his eye. It advertised fellowships at the Los Alamos National Laboratory, where Oppenheimer had led the project to develop the first atomic bomb. He applied, successfully, although he thought having a large nuclear arsenal was “crazy”, and never worked on weapons himself. The weapons scientists would come to his office or ask him questions in the cafeteria, then return to their work in the place Farmer and his civilian colleagues called “behind the fence”.
His thinking on the bomb would change over time. His father, who had landed at Omaha beach and fought across Europe, had just boarded a ship headed for Japan when Hiroshima and Nagasaki were bombed, bringing the war to an abrupt end and allowing him to return home; he still opposed the bomb, but without it, he might never have been born. Like everything else, the question was a little more complex than it had first appeared.
For several years, he was free to pursue his research into chaos and complexity at Los Alamos. Whenever he spoke at conferences, someone in the audience would ask why he hadn’t applied his understanding of complexity and emergent phenomena to the stock market, and he would tell them he had no interest in it. “I knew zero about finance,” he tells me. What he did know about economics seemed bizarre: the ideas that markets are efficient, or that people in markets act according to strict rules of rational self-interest seemed to him “so obviously wrong, that it was exciting to prove it wrong”.
He went to Wall Street – having bought his first ever suit – and found himself like an anthropologist studying an entirely different culture. “I liked the people there a lot better than I thought I would,” he tells me. In 1991, Farmer and his old friend Norman Harry Packard (after whom the roulette computer was named) set up Prediction Company, which used their understanding of complexity to make predictions in financial markets. The company did proprietary trading, meaning it traded on behalf of a bank, using the bank’s money and while it was relatively small, its performance was outstanding. Warren Buffett, widely regarded as the world’s most influential investor, produces a “return-to-risk ratio” about 20 per cent better than the wider (US) stock market; Farmer and his colleagues reached a ratio 500 per cent better. They would eventually sell the company to the Swiss bank UBS for $100m in 2005.
The markets fascinated him. He remembers visiting the trading floor of the Chicago commodities exchange in the late 1980s: “There were screens all over the place with what’s happening in the world, a thousand people yelling and screaming… people running around with messages”; he felt as if he could see “all the neurons of the world were flowing into there”, and the decisions of the bellowing traders spreading out across the globe.
For Farmer, an economy is humanity’s great emergent phenomenon, the means by which we as a group enjoy resources many thousands of times greater than if we were all going it alone. But this has also become our greatest risk, as the power it affords us leads to a more unstable world. This is what makes complexity economics necessary, he says – it can help us to make better predictions, and better decisions, when doing so has never been more important.
In 2020, Farmer and his colleagues at Oxford created models based on the principles of complexity economics – avoiding the standard assumptions of conventional economics and accounting for the more realistic details of how companies actually work – to understand how the UK economy would react as the pandemic arrived. Their prediction was considerably more accurate than that produced by the Bank of England. They are now using detailed, agent-based models to predict how the most difficult economic problem humanity has faced – the transition to renewable energy – will be made possible.
Never underestimate the power of good predictions, he tells me. Moore’s Law, which predicted that computing power would double every two years, became an assumption of the computing industry in the 1970s, giving companies the confidence to invest – because for every idea, the power would soon be provided. The same can be true of renewable energy, if it is effectively modelled, and if we accept an economics that predicts that it will, almost inevitably, become a lot cheaper than fossil fuels, and sooner than most people predict.
Farmer says the climate crisis, the increasing risk of other disasters such as pandemics and financial crises, makes it more important than ever that we use a better, more realistic economics in decisions that affect us all. “We do a poor job of making collective decisions at the political level, particularly at the higher level… as the world becomes more complicated, science becomes more and more necessary, not just for its own sake, but for helping us organise ourselves.”
[See also: The petit bourgeois insurrection]
Content from our partners
Related
