A pre-registered event study: when the divergence regime steps toward crisis, does it carry information about what Hormuz-exposed instruments do next? Signals, instruments, and horizons were fixed before any data (registration PR-0002); outcomes accrue here in the open, one signal at a time. See the methodology.
Fixed at registration — nothing here can widen after the fact.
Instruments: TANKER_EW, ITA_RS, GLD, OVX · horizons 5d, 10d, 20d. Two-sided by registration; no inference below n=20 per cell (below that, descriptive only). The multiple-testing denominator counts every registered test ever — 12 cells across 1 test(s) — so it can't shrink by dropping the tries that failed.
Mean realized forward log-return; n = realized signals in the cell.
| instrument | 5d | 10d | 20d |
|---|---|---|---|
| TANKER_EW | — | — | — |
| ITA_RS | — | — | — |
| GLD | — | — | — |
| OVX | — | — | — |
The full ledger — each signal's instruments and horizons, realized as they mature.
Signals reference the published, hash-verified record; baselines and realizations use published closes only (no look-ahead). This is a research study, not a settlement benchmark and not investment advice. Machine-readable at /experiment.json.