Our strategies
How a hypothesis becomes capital at risk, and what stands between the two.
The research loop
01
Explain
A written, hard-to-vary account of why the return exists and who pays it. If the story can stretch to fit any outcome, it is rejected.
02
Predict
Falsifiable predictions written before testing. The live strategy carries six standing conjectures as its validation suite.
03
Measure
One measurement per concept: term structure, skew, realized versus implied, cross-asset stress. Each is admitted only after multi-year sign stability.
04
Test
Two-stage, walk-forward, deterministic: design frozen on training data, validation touched once, and every simulated trade uses only a model that could have existed that day.
05
Encompass
The final test: demonstrating the edge is structural premium plus sizing discipline, not a forecasting claim the market has already priced.
Risk, before return
Sized to danger
Exposure shrinks continuously as measured danger rises, reaching zero before extremes.
Mechanical exits
Full exit triggers on volatility term-structure inversion and institutional put-buying spikes, the signatures of modern volatility crises.
Hard caps
A fixed portfolio-at-risk ceiling that no conviction can renegotiate. Responses to loss are written before the loss.
Where we struggle
Every strategy has an environment that does not pay it. Ours is a prolonged volatility compression: not a crash, but a long grind in which the premium on offer stops compensating for the risk of selling it. In that regime the discipline does what it is built to do, which is to size down and wait rather than force trades.2022 was such a year.
A second research program, in progress, extends the same danger-state discipline across a wider set of instruments, and exists specifically to earn in the regime the live strategy declines. It is research, not a live capability.
A separate and narrower limit: a signal that reads the market once a day cannot catch an event that happens inside one. Those are bounded, not eliminated, by fixed maximum loss at entry and the portfolio-at-risk ceiling.
What we don't discuss
Not because secrecy is fashionable, but because that edge belongs to our investors. By policy, performance figures appear only in offering materials, where they carry the labeling and context regulation requires. They never appear on a website. What we will always discuss: process, structure, and posture.