The Returns We Want | Intro

If we could design a money machine to generate returns for us, what would we want it to look like?

A money machine runs a portfolio of strategies, producing a continuous stream of profit and loss. That stream can take many shapes: smooth or jagged, stable or volatile, predictable or surprising, with long drawdowns or rapid recoveries. Some of these features are outside our control, but others can be influenced by how we construct the system. This immediately raises the question: what kind of returns do we want?

A roadmap

This series sets out to answer that question in four parts:

1. Simulating returns. We begin by modelling return streams with progressively richer statistical structures, from simple independent daily draws to processes with volatility clustering, autocorrelation, fat tails, and skew. This allows us to see what kinds of P&L patterns emerge under different assumptions.
2. Uncertainty and learning. Next, we address the fact that the parameters of these models are never known with certainty. We operate under uncertainty about means, volatilities, and correlations, and we update our beliefs over time. This makes inference as important as return generation.
3. Preferences. We then ask how to formalise what we truly value in a return stream: smoothness, early resolution of uncertainty, limited drawdowns, positive skew, or other features. This requires utility functions that extend beyond the standard metrics of institutional finance.
4. Translating into practice. Finally, we compare preference-based evaluation with common industry metrics such as Sharpe and Sortino ratios, drawdown statistics, and Value-at-Risk. This lets us derive heuristics that connect explicit preferences with the established language of the industry.

The goal

The purpose of this series is not to find a single formula for the “right” return. Rather, it is to build a framework for understanding the trade-offs in different patterns of profit and loss, and for making deliberate, explicit choices about what kind of money machine we want to construct.

At the end of this series, we will have a structured way to connect our preferences to measurable return characteristics, and from there to practical design choices. This means that when we design a money machine, we will not be guided by inherited metrics or convention, but by a framework that reflects our own values and objectives.

Why this matters for Money Ex Machina

Institutional asset managers must design portfolios for a broad base of external investors. This forces them to adopt simple, standardised measures such as Sharpe ratios or maximum drawdowns; metrics that reflect compromises across many conflicting preferences.

As independent builders, we are not constrained in this way. A money machine does not need to conform to external mandates; it needs to serve our own objectives. This is a profound advantage. It means we can design for our own utility function, making explicit the trade-offs we are willing to accept and the features of the return stream that matter most to us.

To capitalise on this freedom, we need a framework that connects abstract preferences to measurable properties of returns, and then to design principles for strategies and portfolios. The Returns We Want is that framework.

In Part 1, we begin with the foundations: simulating return streams and understanding the only free lunch in finance: diversification.