02 — The Thesis
Labour is finite. Capital may produce returns, but only if the return can survive reality.
Capital in the Wild began with a practical constraint: income earned entirely from labour requires continuing labour. A financial asset or repeatable return process may change that relationship. But markets do not owe us a return, and replacing a wage with trading risk is not automatically an improvement.
We therefore investigate two connected propositions.
1. Can a repeatable return process be constructed?
Systematic trading attempts to express decisions as explicit rules or researchable processes. This makes assumptions visible and allows testing, replication and automation. It also creates many ways to deceive ourselves: selecting favourable histories, ignoring costs, fitting noise and mistaking simulation for evidence.
An attractive backtest is a hypothesis about what might be possible. It is not the same thing as reliable future performance.
2. Can the return actually be held?
A return that cannot be realised or retained is not useful to the purpose that motivated the investigation. We have to think about drawdowns, diversification, the cost of implementation, operational mistakes and the human difficulty of persisting with a process when it is losing money.
This is why the engineering grew beyond identifying signals. The data, portfolio, execution and monitoring systems are integral to the thesis, not optional technical flourishes.
The actual objective
Our target is not maximum theoretical performance. It is financial surplus that could reduce dependence on selling time, subject to the realities of capital risk, uncertainty and ongoing maintenance.
More automation does not mean zero work. More complexity does not mean a better system. No research result establishes that any particular reader should trade or that future returns are assured.
Evidence changes the thesis
We must distinguish candidate edges, in-sample and out-of-sample research, deployed machinery, live execution and realised results. Where results are unavailable or not yet independently verified, we should say so.
A useful thesis is one that can be disproved or revised by the investigation — not one protected from the evidence.