The dominant tradition in Western science is reductionist. Break a system into its components, study the parts, and reassemble your understanding from the bottom up. It is a powerful method. It built modern physics, chemistry, and medicine. But it has a structural limitation: it struggles with complexity. And markets, like most things that matter, are complex.
The alternative is not mysticism. It is a shift in emphasis, from components to processes, from objects to relationships, from static structure to dynamic interaction. That shift has profound implications for how we think about financial markets, and for why systematic trend following works.
What Complex Systems Actually Are
A complex system is not simply a complicated one. Complexity is a specific property: it arises when a large number of interacting components produce behaviours that cannot be predicted or explained by examining any component in isolation. The system is defined by its relationships, not its parts.
Financial markets are a complex adaptive system (CAS). So is the climate. So is a living organism. What these systems share is a set of structural features that distinguish them from merely complicated mechanisms.
They harness locally available resources to fuel development over time. In a market, the injection of capital creates pressure gradients that drive the formation of sub-systems: sectors, asset classes, correlated trading strategies, institutional flows. These sub-systems interact, nest within each other, and give rise to emergent behaviours that no individual participant designed or intended.
They are self-sustaining without central control. No single agent governs the price of crude oil or the yield curve. The system coordinates through the exchange of internal resources and the feedback loops between its participants. Over time, this process of continuous interaction tends to make the system more efficient, more robust, and more stable, until a regime shift breaks the equilibrium and the process begins again.
They exhibit memory and path dependence. What happens now is causally linked to what happened before. A price series is not a sequence of independent draws from a distribution. It is a record of a system evolving through time, shaped by the accumulated interactions of its participants. This is not a minor technical point. It is the foundation on which the entire case for systematic trend following rests.
And they produce non-linear dynamics. When components are tightly coupled and multi-functional, small perturbations can cascade into large effects. The same coupling that makes the system efficient in stable conditions makes it vulnerable to cascading failure when critical relationships break down. Fat-tail events, the extreme tail-risk events that Outlier Hunters are designed to capture, are a structural consequence of this non-linearity, not a statistical anomaly to be normalised away.
Complexity Within Us
The temptation is to treat complexity as a property of things out there: markets, ecosystems, economies. But we are ourselves complex systems, and recognising this is useful for understanding why reductionist thinking is so deeply embedded, and so persistently misleading.
We tend to perceive ourselves as stable, discrete, enduring entities. The closer examination suggests otherwise. We are multi-cellular processes, nested collections of sub-systems within sub-systems: organelles within cells, cells within tissues, tissues within organs, organs within the organism. Zoom further in and the apparent solidity dissolves entirely. At the quantum level, what we encounter are not objects but interactions, probabilistic events, trace signatures of processes rather than things.
Quantum mechanics shattered the Newtonian picture of a deterministic, clockwork universe. Wave-particle duality showed that the fundamental constituents of matter do not have fixed, observer-independent properties. What a particle is depends on how it is observed. Reality, at the deepest level we can probe, appears to manifest through the interaction between observer and system rather than existing independently of it. This is not philosophical speculation. It is the empirical finding of a century of experimental physics.
The philosopher’s version of the same puzzle is the Ship of Theseus. If every plank of a ship is gradually replaced, is it still the same ship? Our cells renew continuously throughout our lives. The atoms in our bodies cycle through the environment and return. What persists is not the material substrate but the process: the pattern of organisation, the web of relationships, the ongoing dynamic that we call a self. We are not assembled objects. We are sustained processes.
This matters because it reveals the depth of the bias we are working against. Reductionism is not just a scientific method. It is a perceptual habit, built into how we experience ourselves and the world. Overcoming it requires a genuine shift in perspective, not merely an intellectual acknowledgement that complex systems exist.
Beyond Reductionism
The Newtonian worldview gave us a universe of objects moving through a fixed arena of space and time according to deterministic laws. It was extraordinarily productive. It was also wrong, or more precisely, incomplete. Einstein’s theories of relativity replaced the fixed arena with a dynamic geometry: matter shapes space-time, and space-time guides matter. The relationship between observer and system became central to the description of physical reality in a way that pure mechanism could not accommodate.
The consequence for how we model complex systems is significant. A reductionist model of a market treats price as the output of a collection of independent, rational agents each processing information and optimising their utility. The interactions between agents are assumed to be linear and additive. The result is a tractable mathematical framework that produces elegant theories: the Efficient Market Hypothesis, mean-variance portfolio optimisation, and others built on similar foundations.
The problem is that these models are wrong in the ways that matter most. They cannot account for fat-tail events. They cannot explain volatility clustering. They cannot reproduce the serial correlation that makes trend following possible. The assumptions that make the models tractable, rationality, independence, linearity, are precisely the assumptions that fail during the regime shifts when the models are most needed.
The alternative is not to abandon modelling. It is to model differently. Think of a conventional jigsaw puzzle: each piece fits into a defined position, the solution is unique, and reductionist assembly works. Now think of a Rubik’s cube. Every move affects every other position simultaneously. The permutations are astronomical. No sequential, component-by-component analysis will solve it. You need heuristics, pattern recognition, and a tolerance for operating under uncertainty with an acceptable margin of error.
Markets are a Rubik’s cube, not a jigsaw puzzle. Statistical tools can describe the gross properties of a system, but they cannot fully capture the intricate dynamics of a CAS. What is required is a process-oriented approach that accepts irreducible uncertainty, uses robust heuristics rather than precise predictions, and is designed to remain functional across a wide range of regime states rather than being optimised for a single one.
Navigating Complexity as an Outlier Hunter
For a systematic trend follower, the recognition of complexity is not abstract philosophy. It is the operational foundation of the entire approach.
Trend following does not attempt to model the components of a market: the fundamentals, the balance sheets, the macroeconomic variables. It models the process. Price is the output of the complex adaptive system, the aggregate result of all the interactions, feedback loops, and regime dynamics discussed above. Following price means following the process rather than any individual component of it.
This is why simple rules outperform complex ones in trend following. A sophisticated fundamental model must be right about a large number of specific things simultaneously. A simple price-following rule need only be right about one thing: that the process is currently in a state that produces directional persistence. The model does not need to know why. It needs to be positioned to capture the move when it happens and to limit the loss when it does not.
The emergent nature of trends is central here. Trends are not the product of a single cause. They arise from the interaction of countless participants, feedback loops, and regime dynamics playing out simultaneously across a complex adaptive system. Trying to predict them from first principles is the wrong problem. The right problem is building a process that is structurally capable of capturing them when they emerge and surviving the periods when they do not.
This is also why diversification is not merely a risk management technique for Outlier Hunters. It is a consequence of taking complexity seriously. In a system where Outliers are unpredictable in their timing, origin, and magnitude, concentration in any single market, asset class, or timeframe is an act of false precision. Wide diversification is the honest acknowledgement that the system is complex, that we cannot know in advance where the next major regime shift will originate, and that the correct response to irreducible uncertainty is breadth of exposure rather than depth of conviction in any single forecast.
The path-dependent, non-ergodic nature of markets also has direct implications for position management. Because the history of a price series matters, because a market in drawdown is not the same as a market at a high watermark even if the current price is identical, the Outlier Hunter’s approach to building and reducing positions must be adaptive rather than fixed. The Cut Back Rule, which systematically reduces exposure as drawdown geometry worsens, is a direct application of this principle: it is a rule designed for a path-dependent world, not an ergodic one.
Markets are processes. The participants within them are processes. The trends that Outlier Hunters capture are the emergent outputs of those interacting processes operating far from equilibrium. Accepting this, and building a systematic approach that is designed for this reality rather than against it, is what separates a process-driven Outlier Hunter from every approach that depends on the market behaving the way a model assumes it should.
The reductionist tradition asks: what are the parts, and how do they fit together? The process tradition asks: what is happening, and how is it evolving? For navigating complex systems, the second question is the right one.