How Our Systems Work

A Systematic, Multi-Layered Approach

Our work is built around a single idea: robust decision-making emerges from systems, not individual components.

Rather than relying on any single market, data source, or technique, we design systems that combine breadth, discipline, and statistical understanding to operate reliably across changing conditions.

We are asset-class agnostic, and do not build systems for a single market or asset class. Our approach is deliberately broad, encompassing:

  • Multiple markets

  • Multiple asset classes

  • Multiple regimes and conditions

Breadth by Design

We are interested in how markets behave from many perspectives. Our systems are designed to work with:

  • Different types of market data

  • Multiple time horizons

  • Diverse information sources

Multiple Data Streams

Raw data is rarely useful on its own, which is why we apply a wide range of transformations to data streams to express behaviour in more meaningful, structured forms.

From Data to Signals

We use AI and machine learning as tools within a controlled framework. Every strategy is evaluated with a strong emphasis on behaviour across time and stability under different conditions.

Strategy Generation as a System

The system is designed to combine many independent components into larger portfolios where correlations are explicitly controlled.

Portfolio Construction

Continuous Re-Evaluation

Markets change, and systems must adapt.

Clear boundaries create robust systems. Building reliable systems requires knowing not only what to pursue, but what to deliberately avoid.

Over time, we’ve learned that many common approaches create fragility, hidden risk, or short-lived results. Our systems are designed with explicit boundaries to prevent this.

By being explicit about what we don’t do, we reduce hidden risk, avoid fragile complexity, and maintain long-term clarity.

What We Don’t Do

  • We don’t depend on any single strategy, signal, model, or insight. No individual component is considered essential. Every part of the system must justify its existence through behaviour over time and in combination with others.

  • We don’t chase peak results, headline metrics, or brief periods of exceptional performance. Short-term optimisation often produces systems that fail quietly when conditions change. We prioritise stability, consistency, and controlled behaviour instead.

  • We don’t deploy systems we do not understand. Complex tools are used carefully and within defined constraints. Outputs are evaluated, stress-tested, and monitored rather than blindly trusted.

  • We don’t assume that past conditions will persist. Relationships change. Behaviour degrades. Regimes shift. Static assumptions are a source of risk.

  • We don’t scale strategies that cannot demonstrate resilience outside their original context. Components must show acceptable behaviour across time, conditions, and combinations before they are allowed to contribute meaningfully.

  • We don’t build systems that look good on paper but behave poorly in production. Design, implementation, deployment, monitoring, and recovery are treated as a single lifecycle.

  • We don’t optimise for how quickly something can be built if it compromises clarity, safety, or long-term maintainability.

By being explicit about what we don’t do, we reduce hidden risk, avoid fragile complexity, and maintain long-term clarity.