Why This Series Exists
Underneath every trading book, every backtest, every strategy sits a question most practitioners never think to ask. It is the kind of question you can work around for an entire career without noticing it is there.
The question is simple. What does it mean to act inside a system you cannot stand outside of?
You cannot observe a market without participating in it. Every price you study formed in your absence, and the moment you act on what you studied, you become one of the forces that moves the next price. There is no vantage point above the market, no hilltop with a clear view. You are in the valley, part of the weather, changing the conditions you are trying to read.
This is the condition the series calls embeddedness, and it shapes far more than most traders realise. It shapes how strategies decay, how intelligence gathers in a system that no single participant can hold, how belief hardens into the structure traders then try to read, and how thin the line really is between you and the market you think you are separate from.
The Embedded Agent is seven essays that take this condition seriously and follow it wherever it leads. Each one builds on the one before it. Together they make a single argument: the deepest edge available to any participant is not a better indicator or a faster feed. It is the recognition that you are inside the system, that you have always been inside it, and that your practice looks different once you stop pretending otherwise.
Here is where it goes.
Episode 1
The series opens with the gap every systematic trader knows and few can explain. The backtest is clean. The logic is sound. Then you turn the strategy on, and the fills come in worse, the drawdowns arrive sooner, and nothing in the code has changed.
The usual suspects, slippage and costs and overfitting, are all real, and none of them reaches the actual problem. A backtest is a photograph of a room you were not in. It records a market that formed without your participation, and the moment you deploy capital you change that market. You consume the liquidity the record assumed was there, your stops cluster with everyone else’s, your entries add pressure the history never felt. The map was drawn of a territory that did not include the mapmaker. Now the mapmaker is walking on it, and the ground shifts underfoot.
This reframes what traders call model degradation. The model is not slowly wearing out. It still describes its world with perfect accuracy. That world has simply moved, because you and everyone who studied the same history walked into it. What a backtest can give you is the character of a strategy, its relationship to certain conditions, rather than a forecast of what it will earn. You hold the map lightly and keep redrawing it as you walk, because you are inside the territory it describes.
Episode 2
The Predator Who Changes the Prey
In 1995, fourteen wolves were returned to Yellowstone. Within a decade they had changed the course of rivers, not by moving earth but by changing how the elk behaved, which changed the vegetation, which firmed the banks. A predator reshapes a landscape by reshaping what it feeds on.
A strategy is a species, and its returns are not a property it owns. They emerge from the relationship between the strategy and the ecology it feeds on, and that ecology changes the instant you enter it. When a signal is yours alone, your entries barely register. When the signal becomes crowded, the clustered entries of everyone running it amplify the move, which looks in the data like an even stronger signal, which draws more capital, until a breakout that used to come from genuine shifts in supply and demand now comes partly from the crowd hunting it. When the move turns, those same clustered positions unwind together, and the drawdown runs deeper than any backtest measured, because the backtest was taken in an ecology with fewer predators feeding on the same prey.
This is why alpha behaves like a relationship to be maintained rather than a treasure to be found, and why every niche has a carrying capacity that success eventually strains. The essay closes on why trend following persists through all of this. It does not depend on a specific inefficiency that can be arbitraged away. It feeds on a generic feature of complex adaptive systems, the tendency of large moves to run further than chance allows, and that feature is renewed by the very dynamics that create it.
Episode 3
The Intelligence That No One Has
No ant knows the architecture of the colony. No neuron understands the thought it helps to form. Intelligence appears at a level none of the parts can perceive, and markets do the same thing.
A single price integrates the inventory of market makers, the positioning of institutions, the trigger levels of systematic strategies, the hedging of dealers, the flows of pension funds, and a thousand other inputs no participant can see in full. The market computes that number through the interaction of everyone in it, including you. This is not the efficient market hypothesis, and it is not a claim that prices are right. A market computes, it does not optimise, and computation without purpose can produce order or catastrophe with equal ease.
The unsettling part for the participant is that you feed the computation but cannot read its output. The price on your screen is only the visible trace. The process that produced it is distributed across every active participant and is not happening anywhere you can look. More data does not solve this, because the intelligence lives in the process, not the record. The essay is careful about what follows. The market is not wise, and it has no interest in your survival, so the embedded agent stops trying to decode it and learns instead to read it, the way a surfer reads water they cannot model, navigating a system smarter than themselves with tools built for navigation rather than comprehension.
Episode 4
When George Soros bet against the pound in 1992, he did not only predict the devaluation. Ten billion dollars of selling helped cause it. His observation was an intervention.
Soros framed reflexivity as a feedback between perception and reality, with beliefs pulling prices away from fundamentals until reality reasserts itself. The essay pushes further. In a complex adaptive system, belief does not distort the structure, it becomes the structure. The clearest case is Value at Risk. It was built to measure risk and ended up manufacturing it. Falling volatility lowered VaR, which freed leverage, which enlarged positions, which suppressed volatility further, until the cycle reversed and ran the same machinery into a cascade. The tool for measuring the system had become a component of the system.
The same loop runs through narratives, credit, and consensus forecasts, each one producing the evidence that seems to confirm it. This is why a reflexive bubble is not an illusion. Its prices and flows are entirely real, which is exactly what makes it fragile, because a structure built from belief comes apart when belief withdraws. The embedded agent stops hunting for the true fundamentals beneath the loop and studies the loop itself, its direction, its intensity, and its fragility. Protection here comes from sizing and diversification robust enough to survive a reversal you cannot time, not from foresight.
Episode 5
You act upon the market. You read it, enter it, respond to it, as though you and it are separate things. The essay takes that boundary apart.
Your stop loss is not a private decision. It is a market order waiting to fire, and when it fires alongside the thousands of others set at the same levels, it becomes the cascade that triggered it. A cell wall is not a barrier but a living membrane, permeable and continuously negotiated, and your relationship to the market has the same quality. You take in prices and emit orders, your orders become the environment other participants read, and their responses become the prices you absorb. Even inaction is participation. Cash withheld during a rally and a position held through a crash are both configurations of your presence, with consequences for the system either way. There is no neutral position and no outside to stand in.
What follows is a different understanding of risk. You do not manage risk from beyond the system, you co-produce it from within, along with everyone who trades as you do. Your hedges can manufacture the very correlations they were meant to guard against. Your edge is not something you possess but something you maintain through the quality of your participation, and it dissolves the moment you stop attending to the feedback between your actions and their consequences.
Episode 6
You can absorb everything the first five essays teach and still not name tomorrow’s close. That is not a gap in your understanding; it is how the system is built.
There are two kinds of prediction problem. Chess is complicated but knowable, and given enough computation every position resolves. Weather is complex. We understand the physics completely and still cannot forecast far, because tiny differences in conditions we can never fully measure produce large differences in outcome. Markets are weather, not chess. Stephen Wolfram’s name for this is computational irreducibility. For some systems there is no shortcut to the future state, and the only way to know what they do is to let them run. Markets belong to that class, so no model and no volume of data will ever crack them, not as a matter of technology but as a matter of mathematics.
Understanding still matters, though not as a route to prediction. It grants orientation, so you are rarely blindsided even when you are wrong on the specifics. It grants structural alignment, so you build for fat tails and regime change rather than for a stationary world. It grants a survival architecture that prevents the fatal drawdown rather than every drawdown. And it grants the capacity to be wrong well, treating a surprise as information about the system rather than a verdict on yourself. The Polynesian navigators crossed oceans they could not predict, oriented by the stars and the swell, and survival, not foresight, was the only success available to them. It is the only kind available here too.
Episode 7
The final essay is a pivot rather than a conclusion, because the condition it describes does not resolve. You are embedded, you remain embedded, and the question is no longer whether you understand that but what you do about it.
The centre of the practice is sizing. Not sizing in the conventional sense of fitting risk to a distribution, but sizing against a distribution that includes your own participation and the participation of everyone who trades as you do. In practice this means holding your size below what the model suggests, because the model describes a world without you, and the surprises it omits tend to arrive at the worst possible moment. Around that sits a discipline the essay calls committed adaptability, acting with full conviction on a view you know to be incomplete, following your framework while knowing it may need to evolve, and developing the judgement to tell when to hold the line and when to adjust it.
The other half is consistency. You cannot time your embedding, you can only sustain it, and your presence through the flat and hostile periods is what earns your presence when the system finally moves your way. Underneath all of it is a shift in how you hold uncertainty, no longer an enemy to be reduced but the medium the whole practice works in, the same uncertainty that produces every trend worth catching. The essay closes on the image of the sculptor who is also the clay, shaping the market and being shaped by it, and on the master’s single brushstroke, which looks simple and carries a lifetime of practice behind it. The next trade, sized well and held lightly, is that stroke.
The Arc
Read in order, the seven essays trace one line of thought. They begin with the problem: you are inside the system, your models describe a world you are absent from, and your actions change the thing you are acting on. They move through the consequences, as strategies disrupt the ecologies they feed on, intelligence gathers at a level no participant can access, belief builds structure, and the boundary between trader and market dissolves. Then they reach the hard question. If you cannot predict, what can you know? And they arrive at a practice built for participants who have stopped pretending to be observers.
This is not a series about trading strategies. It is about the condition that every strategy must live inside. It was written for people who suspect that the most important things about markets are the things nobody discusses, and who want to follow that suspicion all the way down.
The Embedded Agent is a companion to my book Complex Adaptive Markets, and it covers ground the other series in this body of work have not. Where The Natural History of Markets catalogued the system’s behaviour and The Deep Structure of Markets examined its architecture, The Embedded Agent turns the question back on the one doing the looking and asks what happens to the agent trying to understand a system that includes them.
Episode Directory
Episode 1: The Map That Walks
Episode 2: The Predator Who Changes the Prey
Episode 3: The Intelligence That No One Has
Episode 4: The Reflexive Loop
Episode 5: The Boundary Problem
Episode 6: The Asymmetry of Knowing
Episode 7: The Practice of Embedding
Richard Brennan writes on systematic trading, complex adaptive markets, and the philosophical foundations of trend following at atstradingsolutions.com. His books include The Fractals of Finance and Complex Adaptive Markets. The forthcoming Carved by Impossibility completes the trilogy.
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.