Investment-thesis catalog
Investment theses, stress-tested.
Each entry is an economic framework, the portfolio it implies, and the evidence for how that allocation holds up under synthetic and historical stress.
13–14% median · 25–26% p95
Regime-Balanced Allocation with a Debasement Tilt, Conservative tier: deepest synthetic drawdown, range across 5 seeds. Every stress drawdown is shown with its tail, not just the typical path. The tails are model figures, not validated against real episodes.
tested → falsified
Hypotheses are preregistered before fitting. Annual seasonality terms improved the out-of-sample regime probabilities in 0 of 8 cells — so they are excluded, and we say so.
2022: the hedge broke
The stress model includes the regime where diversification fails — bonds falling with equities — not only the regimes where it works.
Regime-Balanced Allocation with a Debasement Tilt
Derived from the Systemic Money Model — credit-money, r < g erosion, the debasement/Fisher channel.
The portfolio assigns a different defender to each macro regime, with a deliberate but bounded tilt toward scarce real assets. Whether each defender carried is read from the episodes.
Free-Rider Aggregate Ownership
Derived from the Free-Rider Portfolio Framework — collective intelligence, aggregation ownership, relentless friction minimisation.
The framework holds the weighted-average outcome of everyone trying to pick the winner, with the market’s price discovery doing the work at near-zero cost.
Robotics & Physical AI
An editorial framework — AI value creation extends from software into the physical world; the exposure is built as a basket satellite, not a winner-pick.
The framework describes the robotics build-out as a diversified satellite on a world-market core. In real episodes the theme’s extra loss came in the rate-driven regimes; in the model its deepest regime is the crisis shock.
Each new entry adds a framework, the portfolio it implies, and a full stress run — in the same format.
How to read this
Frameworks are stated as falsifiable hypotheses with evidence grades; portfolios are illustrative, rule-based allocations; stress results are model-dependent and described with their caveats. Nothing here is a recommendation, a suitability statement, or a forecast — only whether the evidence supports, qualifies or refutes each claim.