Youtube X-twitter Linkedin-in

Foundations

The Compass · Foundations

The Foundations of Systematic Trading Under Uncertainty

Before the machinery, before the research, before the outliers, there are a few ideas that everything else on ATS rests upon.

Systematic trading is often presented as a collection of rules: an entry, an exit, a position size, a portfolio. But the rules make sense only when the assumptions beneath them are made explicit.

The ATS framework begins from a different place than prediction. Markets are uncertain, adaptive and path-dependent. The largest moves are rare. The future cannot be known in advance. So the task is not to construct a perfect forecast, but to build a process that remains useful when the forecast is wrong.

The principles below are the foundations on which the rest of the site is built.

Foundations are not the simple part of the framework. They are the load-bearing part. If these ideas are unclear, the more advanced work on Outlier Hunting, research, complex markets and system architecture becomes much harder to understand.

Markets are uncertain

The starting point is not that markets are random. It is that they are difficult to know in advance. They are populated by adaptive agents, shaped by feedback, changing incentives and events that were not contained in yesterday's model.

Foundation 01

Uncertainty changes the objective.

If the future cannot be known reliably, the job is not to eliminate uncertainty. It is to design a process that can operate inside it.

Trend following is a process

Trend following does not require a story about why a market should rise or fall. It observes what is happening and responds according to predefined rules. That is one of its great strengths. The process does not need to understand the cause of a move before participating in it.

Foundation 02

Rules replace the need for prediction.

A systematic process defines what will happen before emotion, narrative and hindsight have a chance to renegotiate the decision.

Outliers drive outcomes

Financial returns are not neatly distributed around an average. A relatively small number of large observations can dominate long-run results. For the trend follower, this has a profound consequence: many trades will contribute little, while a few exceptional trends can carry an extraordinary share of performance.

Foundation 03

The rare event is not a nuisance in the data.

For an Outlier Hunter, it is often the event the entire process exists to capture.

This principle is developed further in Outlier Hunting and The Lifting Power of Outliers.

Diversification is structural

If nobody can know which market will produce the next exceptional move, breadth is not merely a technique for smoothing a return stream. It is a way of increasing the number of opportunities available to encounter what cannot be identified beforehand.

Foundation 04

Hold many possibilities because you cannot know which one will matter.

Diversification is therefore connected directly to uncertainty and outlier capture, not simply to reducing portfolio volatility.

Path matters

Investors do not experience an average return. They experience a sequence. Gains and losses compound multiplicatively, drawdowns alter the capital available for future opportunity, and an absorbing barrier can end the game before the long-run expectation is ever realised.

Foundation 05

The order of outcomes can change the destination.

That is why geometric growth, survival and the equity path matter alongside conventional summary statistics.

For a deeper treatment, see The Compounder's Delight and The Fair Game That Kills You.

Volatility is not risk

Volatility tells us how widely returns are dispersed. That is useful information. But dispersion is not the same thing as the full economic meaning of risk. A positively skewed strategy can be volatile because the very observations it is designed to capture sit far from the centre of the distribution.

Foundation 06

Measurement must respect the objective.

Variance and Sharpe can describe aspects of a return stream, but they do not by themselves tell us whether a strategy is achieving the objective for which it was built.

The empirical argument is developed in Risk Is Not Variance.

Robustness over optimisation

A backtest can be made beautiful by fitting rules to the particular history in front of us. The danger is that the future is under no obligation to resemble that history closely enough for the fitted solution to survive.

Foundation 07

Build for many possible futures, not one perfect past.

Robust systems favour broad applicability, simple logic and behaviour that survives changes in market, parameter and environment.

Process over prediction

The preceding principles converge here. If markets are uncertain, outcomes are fat-tailed, opportunity is dispersed, paths matter and models are imperfect, then the durable advantage is not a superior forecast. It is a process that determines how we respond.

Foundation 08

The system is the commitment made before the uncertainty arrives.

Entries, exits, sizing, diversification and portfolio rules are not independent tricks. Together they form an architecture designed to keep the trader in the game.

The Foundations Series

The principles above are the conceptual map. The full Foundations Series develops the practical machinery in ten connected essays: the system itself, position sizing, diversification, edge, noise, systematic versus discretionary trading, drawdowns, backtesting, equity curves and the psychology of following a process.

Two ways to enter the series: begin with the Foundations Series Synopsis for the guided sequence, or use The Vault when you want to explore the wider ATS archive alongside it.

The purpose of a foundation is not to tell us what the market will do. It is to give us something solid enough to stand on while the market does whatever comes next.

← Start Here   |   Research & Evidence →

↑ Back to top