What Complexity Economics Means for Your Trading
Eight principles from thirty-eight years of Santa Fe science, translated into the language of portfolio construction, risk management, and the practice of following trends
I began this series with a question. Why do trends exist?
It is the most basic question a trend follower can ask, and for decades it was the one that mainstream finance could not answer. The efficient market hypothesis said trends should not exist. The random walk theory said price movements are unpredictable. The equilibrium framework said that any exploitable pattern would be arbitraged away before a systematic strategy could profit from it.
Seven articles later, the answer is in front of us. It has been assembled piece by piece from the work of Brian Arthur, Doyne Farmer, John Holland, and the community of physicists, mathematicians, biologists, and economists who have spent thirty-eight years building an alternative to equilibrium thinking at the Santa Fe Institute.
Trends exist because markets are not in equilibrium. They are complex adaptive systems, populated by agents who form predictions, test them, and update. The predictions crowd. The crowding generates instability. The instability produces directional moves that persist long enough for a responsive strategy to capture. This is not an anomaly. It is a structural feature of any market populated by adaptive agents. It will not be arbitraged away. It cannot be, because the mechanism that produces it is the same mechanism that produces the market itself.
That is the science. Now let me translate it into practice.
Principle One: The market is not a machine. It is an ecology.
This is the foundational insight and everything else follows from it.
An ecology is a system of interacting species that co-evolve, compete, and adapt. No single species controls it. No central planner designs it. The patterns it produces (population cycles, migration, predator-prey dynamics, the periodic reshuffling of which species dominates) arise from the interaction of the species themselves.
A market is an ecology of trading strategies. Value investors, momentum traders, arbitrageurs, market makers, passive index funds, algorithmic systems, retail speculators, central banks: each is a species in the ecology. Each feeds on a particular niche. Each is affected by the behaviour of the others. The aggregate dynamics of the market (trends, reversals, volatility clusters, fat tails, bubbles, crashes) arise from the interaction of these species.
If you accept this, then the way you think about your portfolio changes fundamentally. You are not trying to predict the future state of a machine. You are trying to survive and thrive in an ecology. The relevant question is not “what will the market do?” It is “what is the ecology doing, and how is my strategy positioned within it?”
Principle Two: Predictions are food. Strategies that consume them are the ones that endure.
This is the lesson of the El Farol bar, applied directly to portfolio management.
Every prediction in a market is a species of strategy. When many participants converge on the same prediction (“housing will not fall,” “volatility will stay low,” “passive will keep flowing in”), the prediction crowds. The crowding changes the market in ways that eventually invalidate the prediction. The prediction fails, and the failure releases energy: a directional move, a volatility spike, a repricing.
Trend following does not make predictions. It consumes the failures of other people’s predictions. When a consensus breaks (housing collapses, the Swiss franc de-pegs, oil crashes, bonds sell off), the energy released takes the form of a trend. The trend follower captures that energy. Not by knowing in advance what would break, but by being present across enough markets that whenever something breaks, the programme is there to respond.
This is why I wrote in Complex Adaptive Markets that predictive strategies are food for responsive strategies. They exist in a feeding relationship. The more predictions are made, the more predictions will crowd. The more predictions crowd, the more spectacularly they will eventually fail. And the more spectacularly they fail, the greater the energy available to a responsive programme. Trend following does not compete with predictive strategies. It feeds on their lifecycle.
Principle Three: Trends are increasing returns in the domain of price.
Arthur’s theory of increasing returns explains why small advantages compound into dominant outcomes. The VHS standard triumphs not because it is better but because early adoption triggers positive feedback: more users, more content, more users. The dynamic is self-reinforcing until something breaks the cycle.
A trend in a market is the same dynamic playing out in price. A small initial move attracts attention. The attention generates buying (or selling). The buying pushes the price further in the same direction. The further move attracts more attention. The cycle is self-reinforcing. This is not momentum as a statistical artefact. It is momentum as positive feedback, the same positive feedback that locks in technologies, concentrates market share, and creates the power law distributions that characterise complex systems.
The practical implication is precise. You do not need to predict which market will trend. You need to be present in enough markets that whenever positive feedback begins, you participate in it. And you need a mechanism for recognising when the feedback has reversed (when diminishing returns have replaced increasing returns), so that you exit before the reversal consumes your gains.
This is what a trend following system does. It is an increasing-returns detector. It identifies positive feedback, rides it, and exits when the feedback structure changes. The entry is a bet on increasing returns. The exit is the recognition that they have ended.
Principle Four: All systems will be gamed. All boundaries will break.
Arthur’s principle, explored in detail in Article 4, carries a direct instruction for portfolio construction: do not assume that any boundary will hold.
Currency pegs will break. Volatility floors will break. Interest rate regimes will break. Trade agreements, sanctions regimes, central bank forward guidance, regulatory frameworks: every fixed structure in the global economy is a boundary, and every boundary stores energy. When the boundary breaks, the energy releases as a directional move. The Swiss National Bank abandons its floor in January 2015 and the franc moves twenty percent in minutes. The Bank of Japan adjusts its yield curve control in 2022 and the yen trends for months. The Federal Reserve reverses decades of falling rates and bonds enter their worst bear market in a generation.
A programme that takes Arthur seriously does not predict which boundary will break. It distributes its exposure across as many boundaries as possible. It monitors all of them continuously. And when one breaks, it responds. Not instantly (that would be a predictive strategy, betting on the direction of the break). But promptly, following the direction of the energy release once it becomes visible in price.
This is the architecture described in Carved by Impossibility. You build the portfolio not around what you predict will happen, but around the elimination of everything that would destroy you if you were wrong. You do not concentrate in the boundary you think most likely to break. You distribute across all of them, because you cannot know which one will go first, and the one that goes first might be the one you never considered.
Principle Five: Fat tails are not rare events. They are the system’s normal output.
The artificial stock market demonstrated that extreme price moves arise spontaneously from the interaction of adaptive agents. Farmer’s market ecology models explain why: when a large cluster of similar strategies unwinds simultaneously, the resulting price move is disproportionately large. The move is not caused by an external shock. It is caused by the internal dynamics of the strategy ecology.
This changes everything about how you think about risk.
If fat tails are caused by external shocks (earthquakes, pandemics, wars), then they are rare and unpredictable. You can insure against them, or you can accept the risk. But if fat tails are caused by the internal dynamics of the strategy ecology (and they are), then they are not rare at all. They are the system’s normal mode of adjusting from one ecological state to another. They will happen regularly, in every market, in every era. The question is not whether they will happen. It is whether your portfolio is positioned to survive them and profit from them.
Trend following is designed for fat tails. Its asymmetric return profile (small losses cut quickly, large gains allowed to run) means that the programme benefits from the very events that destroy strategies built on the assumption of normal distributions. The fat tail is not a risk to be managed. It is the return to be captured. This is the deepest alignment between the science and the strategy: complexity economics says fat tails are structural, and trend following is the strategy structurally designed to harvest them.
Principle Six: Simple rules outperform in complex environments.
This is perhaps the most counterintuitive lesson of complexity science, and one of the most important for practitioners.
In a complex adaptive system, the environment is non-stationary. The rules that govern the system’s behaviour are changing, because the agents within the system are adapting and the ecology is evolving. In such an environment, a complex model with many parameters will overfit to the recent past. It will mistake the current state of the ecology for a permanent law. When the ecology shifts (and it will), the complex model will fail.
A simple model with few parameters will not capture the recent past as precisely. But it will be robust to shifts in the ecology. It will not mistake a temporary pattern for a permanent structure. It will continue to function when the environment changes, because it was not optimised for conditions that no longer obtain.
This is why the most enduring trend following systems are simple. Moving average crossovers. Breakout channels. Momentum filters. These are not sophisticated. They are not intellectually impressive. They do not win awards for financial engineering. But they work, and they continue to work, across decades and across every market that has ever been tested, precisely because they are simple enough to survive the shifting ecology of a complex adaptive market.
Complexity economics provides the theoretical justification for simplicity. In a system that never reaches equilibrium, the optimal strategy is not the one that captures the most information. It is the one that is robust to the information being wrong.
Principle Seven: Diversification is not risk management. It is the structural response to an ecology.
In the equilibrium framework, diversification reduces risk by spreading exposure across uncorrelated assets. This is the Markowitz logic, and it is correct as far as it goes. But complexity economics reveals a deeper reason for diversification.
If the market is an ecology of strategies, and if each market is a separate ecology with its own species composition, then diversifying across many markets is not just spreading risk. It is increasing the number of ecologies you participate in. Each ecology will produce its own dynamics: its own trends, its own bubbles, its own fat tails, its own cycles of prediction formation and failure. By being present in all of them, you increase the probability that at any given time, at least one ecology is producing the dynamics your programme is designed to capture.
This is why the best trend following programmes trade everything: equity indices, government bonds, currencies, energies, metals, grains, soft commodities, interest rates. Not because any single market is likely to trend at any given time. But because the ecology of each market is independent enough that when one is quiet, another is moving. The programme harvests the aggregate output of dozens of ecologies, and the aggregate is far more stable and far more profitable than any single market could provide.
In the Trend Following Manifesto with Niels Kaastrup-Larsen, we describe this as “the breadth principle.” The returns of a diversified systematic programme come not from the quality of any single trade, but from the breadth of the programme’s exposure to the world’s ecologies. More markets, more boundaries, more opportunities for the programme to capture the energy that complexity generates.
Principle Eight: The programme is the strategy. Not you.
The final lesson of complexity economics is perhaps the hardest one for a discretionary trader to accept.
If the market is a complex adaptive system that no individual agent can fully understand, then the correct response is not to try harder to understand it. It is to build a system that responds to it. A system that detects trends without needing to understand why they formed. A system that cuts losses without needing to know whether the reversal is temporary or permanent. A system that scales position sizes to the volatility of each market without needing to predict whether volatility will rise or fall.
The programme is the strategy. The rules are the rules. The human’s role is not to override the system when the ecology produces outcomes that feel uncomfortable. The human’s role is to build the system correctly, to maintain it, and to let it operate. Arthur showed that no agent in a complex adaptive system can deduce the correct strategy from first principles. Holland showed that agents who evolve their strategies through simple rules outperform agents who try to compute optimal solutions. Farmer showed that the dynamics of market ecologies are structural and persistent. The science is unanimous: in a complex adaptive system, the systematic response outperforms the discretionary one.
This is not an argument against human judgment. It is an argument for encoding human judgment into a system and then trusting the system to execute it. The greatest discretionary act in trend following is the decision to build the programme. Everything after that is discipline.
Let me draw the through-line one final time.
In 1987, Brian Arthur and Kenneth Arrow gathered a group at the Santa Fe Institute and asked what economics would look like if you dropped the assumption of equilibrium. Thirty-eight years later, the answer is clear. It looks like complexity. It looks like an ecology of adaptive agents, producing emergent dynamics that no individual agent controls or predicts. It looks like fat tails, clustered volatility, speculative bubbles, cascading failures, and persistent trends. It looks like the world that trend followers have been trading in for decades.
The artificial stock market proved it in simulation. Farmer’s market ecology models explained the mechanism. The Bank of England and the European Central Bank have begun incorporating these insights into their analytical frameworks. The fourth volume of The Economy as an Evolving Complex System, published in February 2026, documents a field that has arrived.
And the strategy that aligns most naturally with this science is the one that makes no predictions. The one that responds to dynamics rather than forecasting them. The one that distributes across many markets and many boundaries. The one that uses simple rules, cuts losses quickly, and lets profits run. The one that treats fat tails not as risks but as opportunities. The one that understands the market as a living ecology and positions itself to harvest whatever that ecology produces.
That strategy has a name. It has had a name for a very long time. And now, for the first time, it has a science.
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.