Do Anything You Like, Providing It Is Not Conventional
How a group of renegade scientists dropped the most sacred assumption in economics, and accidentally explained why trend following works
I have spent most of my career trying to answer a question that should not need answering.
Why do trends exist?
If markets are efficient, if prices already reflect all available information, if the collective wisdom of millions of participants has already been compressed into the number on the screen, then persistent directional movement should not happen. There should be no momentum to follow. There should be no signal in yesterday’s price that tells you anything about tomorrow’s. Every model taught in every finance programme in the world says so.
And yet.
Trends persist. They appear in equities, bonds, commodities, currencies, and volatility. They appear across timeframes and geographies. They appear with such regularity that an entire class of diversified systematic strategies has been built to harvest them, and has delivered positive returns through wars, pandemics, crashes, and rate cycles spanning more than four decades. This is not an anomaly that can be waved away. This is a signal that the theory is wrong.
In my recently published book The Fractals of Finance, I showed that markets carry the geometry of self-similarity and memory. In my forthcoming book titled Complex Adaptive Markets, I show that the living systems inside markets, the feedback loops, the cascades, the ecology of competing strategies, produce the patterns we observe. And in my coming book Carved by Impossibility, also due in 2026, I argue that structure itself is what remains when everything else is eliminated.
This series goes to the source.
Nearly four decades ago, a small group of physicists and economists gathered at the Santa Fe Institute in the New Mexico desert and did something remarkable. They dropped the foundational assumption of modern economics. They abandoned equilibrium. And what they built in its place turns out to be, I believe, the deepest theoretical validation that diversified systematic trend following has ever received.
This is the story of how that happened.
In the autumn of 1987, an Irish-born economist named W. Brian Arthur picked up his telephone in a cramped office at Stanford University and placed a call to Kenneth Arrow, one of the most decorated economists alive, a Nobel laureate, and a man not easily rattled. Arthur had a question that felt, even as he was dialling, slightly absurd.
The Santa Fe Institute had asked Arthur to organize an extraordinary experiment. The plan was to bring ten physicists and ten economists into a room for ten days and have them rethink economics from scratch. The physicists would include Philip Anderson, a Nobel Prize winner from Princeton. The economists would include Arrow himself. The sponsor was John Reed, the chairman of Citibank, who was putting up the money because he suspected that something was deeply wrong with the way his own industry understood risk.
Arthur’s problem was scope. How far were they allowed to go?
“Ken,” Arthur said, “we’re not quite sure what to do or how daring we can be here. What do you think?”
Arrow called Anderson at Princeton. Anderson called Reed at Citibank. The word came back through the chain: Do anything you like, providing it is at the foundations of economics and is not conventional.
Arthur would later compare this moment to Martin Luther receiving a message from the Pope in 1520 saying, “Redo theology any way you like, just not the way we’ve been doing it.” It was the kind of permission that changes everything. Not because it grants new capabilities, but because it removes the invisible constraint that everyone had been pretending wasn’t there.
The first decision the group made was the most radical: they dropped the entire notion of equilibrium.
If you are a trend follower, that sentence should stop you cold. Because equilibrium is the reason the textbooks say your strategy shouldn’t work.
The standard story goes something like this. The economy is populated by rational agents (consumers, firms, investors) who face well-defined problems and make optimal decisions. They know what they want. They know the rules. They process all available information and act accordingly. When all these rational decisions interact, they produce an outcome that is consistent with the decisions that produced it. Supply meets demand. Prices reflect value. The system is in balance.
This is equilibrium. It is elegant. It is mathematically tractable. And it has provided economics with the authority of seeming precision for two centuries.
It is also, I believe, the single most dangerous idea in finance.
Because if you accept equilibrium, you must accept its consequences. Markets cannot trend, because any predictable pattern would be arbitraged away instantly. Crashes cannot cascade, because rational agents would see them coming and correct the mispricing. Diversification across dozens of uncorrelated markets cannot produce persistent edge, because there is no edge to be found. Every price is already correct. The entire philosophical foundation of systematic trend following becomes a logical impossibility.
And yet there I was, year after year, watching trends emerge from markets that were supposed to be in balance. Watching energy stored behind compressed boundaries release into moves that the models called impossible. Watching the same patterns repeat across grains, metals, rates, and equity indices, not because someone designed them, but because the system was alive. I did not have a word for what I was observing. The scientists in Santa Fe did.
Consider what the equilibrium framework required you to believe about financial markets in October 1987, the very month Arthur was making his phone call.
On Monday, October 19th, the Dow Jones Industrial Average fell 22.6 percent in a single session. Nearly a quarter of the total value of American equities vanished in six hours. No war had broken out. No pandemic had struck. No fundamental economic variable had shifted by anything close to 22.6 percent.
Under equilibrium theory, this was essentially impossible. Prices were supposed to be “right.” If they moved, they moved because the underlying reality had changed. But reality hadn’t changed on Black Monday. What had changed was the behaviour of the agents inside the system. Portfolio insurance programmes, designed to automatically sell stocks as they fell, triggered cascading waves of selling. Each round of automated sales drove prices lower, which triggered more automated sales. The system fed on itself.
If you follow my work on complex adaptive markets, you recognise this immediately. A boundary compressed. A gradient accumulated. The boundary broke, and the stored energy released. This was not an anomaly. This was a complex adaptive system doing exactly what complex adaptive systems do.
The economists called it a “market anomaly.” The physicists arriving in Santa Fe called it “positive feedback.” And they knew exactly how to think about it, because positive feedback was central to their understanding of everything from laser physics to the formation of galaxies.
The economists treated self-reinforcing dynamics as exceptions to be explained away. The physicists treated them as the main event.
Diversified systematic trend followers treat them as the source of returns.
What emerged from that first Santa Fe workshop, and from the years of research that followed, was not just a critique. It was a framework. Brian Arthur would eventually call it complexity economics.
The core ideas can be stated simply. Standard economics assumes agents are identical, rational, and all-knowing. Complexity economics assumes agents are different from one another, limited in what they can know, and forced to figure things out as they go. Standard economics assumes the system reaches a resting point. Complexity economics assumes the system may never rest.
To make this concrete, Arthur invented a thought experiment in 1994 that has since become one of the most cited problems in complexity science. He called it the El Farol Bar problem, after a real bar in Santa Fe.
One hundred people must decide each Thursday whether to go to El Farol for live music. The bar is fun if fewer than sixty show up, and miserable if more than sixty show up. Each person decides independently, based only on past attendance patterns.
There is no correct solution. Every strategy in the El Farol problem is a prediction about a specific future state: the bar will be full, the bar will be empty, attendance will follow last week’s pattern. If everyone uses the same prediction, they all act together, and the prediction defeats itself. In a reflexive system, prediction at scale is inherently self-destroying.
Read that sentence again. It is the most important sentence in this article for anyone who runs a systematic strategy.
What Arthur discovered when he simulated the problem was that agents evolved an ecology of different predictive strategies. Some used simple rules. Others used complex pattern recognition. Others went randomly. The overall attendance fluctuated around sixty but never settled. The system never reached equilibrium. The ecology of predictions perpetually adapted.
This is the market. This is what you trade every day. Value predictions become popular and compress their own premium. Volatility predictions become consensus and the floor drops out. Short-selling predictions crowd until the exit cannot accommodate the crowd. The ecology of predictions shifts. It never resolves.
And here is the connection that matters most: a diversified systematic trend following programme, spread across dozens of uncorrelated markets, is not a prediction competing inside the ecology. It is a responsive strategy that harvests the behaviour of the ecology itself. It does not forecast which prediction will crowd and fail. It observes the directional moves that prediction failure produces, and follows them. When one ecology collapses somewhere in the world, a trend begins somewhere in the world. The programme captures it. Not because it predicted it, but because it was positioned to respond.
The portfolio does not need any single market to behave. It needs the ecology of predictions to keep forming and failing. And the ecology always does, because Arthur proved that in a reflexive system, it must.
Arthur had encountered these dynamics in real economies long before he named them.
In the early 1980s, while still at Stanford, he became interested in a question that standard economics couldn’t answer: why do certain technologies win not because they are the best, but because they got lucky early? He called this increasing returns: in some markets, the more market share a product gains, the easier it becomes to gain more.
The classic example is the QWERTY keyboard. Sholes designed it in the 1870s to prevent typewriter jams. By the time jamming was no longer an issue, millions had already learned the layout. Every new typist learned QWERTY because that was what everyone else used. The layout locked in. Not because it was optimal, but because history gave it an early advantage that compounded.
Arthur tried to publish this in economics journals in 1983. The rejection letters were brutal. One reviewer wrote that increasing returns “cannot be an equilibrium.” That was the point, of course. But the reviewers couldn’t see it. Their training had made the most important feature of the system invisible.
Arthur was eventually vindicated. The technology economy of the 1990s ran on increasing returns. Windows dominated not because it was best, but because early adoption created a network effect that compounded. Google, Facebook, Amazon, Apple: the same dynamics, the same positive feedback, the same lock-in.
In financial markets, the same mechanism operates every day. When BlackRock’s iShares launched the first wave of exchange-traded funds, increasing returns kicked in immediately. The largest ETFs attracted the most volume, which meant tighter spreads, which attracted more investors, which made the funds larger. By 2024, the top three S&P 500 ETFs held combined assets exceeding $1.5 trillion. Their dominance was not the result of being fundamentally better. It was the result of positive feedback.
Under equilibrium theory, prices reflect fundamental value. Under increasing returns, prices reflect network dynamics. A stock rises not only because its earnings improve but because its inclusion in a popular index fund guarantees automatic buying from every new dollar of passive investment. The more it rises, the more passive money it attracts.
Trend followers know this pattern intimately. We don’t call it increasing returns. We call it a trend. And we don’t try to predict when the lock-in will break. We simply follow the direction of the feedback, cut when it reverses, and move to the next market where the ecology is building pressure. The science and the practice say the same thing. They always have.
There is one more idea from Arthur’s work that deserves attention here.
In a paper titled “All Systems Will Be Gamed,” Arthur argued that any economic system built on rules will eventually be exploited by the agents inside it. Not through corruption. Through adaptation. If you have diverse agents operating in a system governed by fixed rules, some will discover ways to use those rules to their advantage in ways the designers never intended. The exploitation emerges from the interaction between adaptive behaviour and fixed structure, the same way water finds cracks in a dam.
The 2008 financial crisis was a textbook case. The rules of the system (capital requirements, credit ratings, securitisation structures) were transparently published. Within that framework, an entire ecosystem of adaptive agents discovered how to manufacture AAA-rated securities from pools of mortgages that were, individually, terrible bets. No single actor caused the crisis. No one decided to destroy the global financial system. What happened was emergent. Mortgage originators, investment banks, rating agencies, and yield-hungry investors each made locally rational predictions about risk. The global outcome, a cascading collapse that wiped out $2 trillion in bank capital, was a property of the system, not of any individual within it.
Arthur’s point was not merely that crises happen. His point was that equilibrium thinking, by its very structure, cannot see them coming. If you assume the system is in balance, then by definition, cascading failures cannot happen. The framework that was supposed to describe reality had made the most dangerous feature of reality mathematically invisible.
This is why diversified systematic trend following programmes survived 2008. Not because they predicted the crisis. No one predicted it. But because trend following does not predict at all. It responds. It assumes the opposite of equilibrium: that the system is alive, that feedback loops will amplify, that boundaries will break, and that when they break, the stored energy will release into a trend. The programme doesn’t need to know which boundary will break. It needs to be positioned across enough markets that when one does, the signal is captured.
In 2008, while the equilibrium models were saying the world was fine, systematic trend followers across commodities, bonds, and currencies captured some of the largest moves in a generation. They didn’t predict. They responded. They were aligned with the reality of the system, not with the fiction of balance.
There is a phrase Arthur uses that, once you hear it, is difficult to forget.
Standard economics, he says, views the economy as something “mechanistic, static, timeless, and perfect.” Complexity economics views it as something “organic, always creating itself, alive and full of messy vitality.”
Messy vitality.
That phrase describes what I have spent my career trying to articulate. Markets are not calm pools that occasionally get disturbed. They are churning ecosystems where predictive strategies compete, adapt, die, and are replaced. Continuously. Without pause. Without resolution. I wrote about murmurations of starlings to describe this. Arthur wrote about ecologies of beliefs and forecasts. The language differs. The observation is the same.
On February 12, 2026, the Santa Fe Institute Press published The Economy as an Evolving Complex System IV: two volumes, thirty-one chapters, representing the state of the art in complexity economics. The editors include J. Doyne Farmer, a physicist who went from beating casinos with hidden computers to building one of the first quantitative trading firms to modelling entire economies at Oxford. The contributors include researchers whose agent-based models are now used by central banks, financial regulators, and policymakers around the world.
The earlier volumes, the editors note, proved that complexity ideas could work in principle. This volume shows that they are working in practice.
For anyone who manages money through a diversified systematic framework, this matters. If markets are complex adaptive systems (and they are), if they are ecologies of competing predictions that never reach equilibrium (and they don’t), then the entire conceptual framework of modern finance needs to be rethought. Crashes are not acts of God; they are phase transitions in a system pushed too far from its critical point. And the strategies that survive are not the ones with superior predictions. They are the ones that have stopped predicting altogether and started responding: following the trend when prediction failure generates directional persistence, fading the extreme when the move overshoots. Trend following and mean reversion are not anomalies that equilibrium theory needs to explain away. They are structural features of a reflexive market, as permanent as the market itself.
Trend following is not a bet on direction. It is alignment with the living dynamics of a non-equilibrium system.
When Arthur and his colleagues gathered in Santa Fe in 1987, they could not have known what they were starting. The science of complexity was barely named. Agent-based modelling required computing power that didn’t yet exist. The idea that the most sacred assumptions in economics could simply be dropped was considered professionally dangerous.
Nearly four decades later, the revolution they began has arrived. And the practitioners who have been operating in alignment with it all along, the ones who respond to the system rather than forecast it, who diversify across dozens of uncorrelated markets rather than concentrating in a few, who cut losses short and let winners run because that is what adaptive survival requires, are finally seeing the theoretical framework catch up with what they have always known to be true.
The economy is not a machine that returns to balance. It is organic. It is alive. It is full of messy vitality.
The traders who thrive are the ones who understand that the mess is the signal.
The eight articles in this series will take you inside the revolution. We will meet the physicist who beat casinos, built a trading firm, and then went on to simulate entire economies. We will watch a computer build an artificial stock market from scratch and produce, spontaneously, every pattern that responsive strategies are built to capture: fat tails, clustered volatility, speculative bubbles, and sudden crashes. We will see how central banks are using these tools today. And we will trace the line from the desert workshop in 1987 to the diversified systematic portfolio you run right now.
The science confirms what the markets have been telling us all along.
There is no equilibrium. There is only adaptation.
Follow the trend.
Want the theoretical foundation for why trend following works?
The Fractals of Finance: Determinism, Adaptation and the Geometry of Markets
The book explores the full architecture of feedback, fat tails, and fractal structure in financial markets, and what it means for how we trade, invest, and understand risk.
Available now on Amazon in paperback, hardcover, and Kindle.
Want a practical field manual for trading trends and capturing outliers?
The Aussie Turtles Trend Following Guide: A Field Manual for Hunting Outliers adapts the timeless principles of the original Turtle traders into a systematic, rules-based approach for modern markets. Co-authored with Adam Havryliv.
Available now on Amazon in paperback, hardcover, and Kindle.