03 — The
Process

You are subscribing to how we actually build and risk the book — gold-led across GC, MGC, XAU/USD, GLD, and GDXU, plus other assets the desk studies for the quant fund — not to data dumps or trade calls. We show the sauce: problem framing, validation, rejection, and agentic research workflows we use to test ideas. Each week’s packet is the public translation of what Parallax structured, Sentinel validated, and PRISM scored: what we think matters, what would invalidate it, and how we would execute if it holds. When a hypothesis fails, we publish the postmortem anyway — because that archive constrains every future claim.

This person is not just posting market opinions. This system studies, validates, rejects, archives, and learns.

• Weekly briefing — research packet (gold-led / multi-asset), Market Codex week map, and technical analyst report, shipped together
• Named artifacts — validated notes, failed postmortems, feature notes, execution reviews, regime studies, agentic workflow notes — each with required sections
• Evidence bar — walk-forward, robustness, regime-conditioned results, costs and slippage included before anything publishes
• Process transparency — sensitivity surfaces, trust/distrust rationale, archive entries — how we solve problems, not just the conclusion
• Primary sources only — every claim traceable (see Sources)