Call #3622 · GH · Currency depreciation (FX reference rates)

GHS depreciates 1.95% or more against USD at any point within 14 days, from 11.7314.

Made 2026-10-02, resolving 2026-10-16 — a 14-day window, fixed before the outcome was known. Every field on this page was frozen when the call was made.

28%
what we said
stated on 2026-10-02, before the window opened
calls table · probability
26%
how often it happens anyway
GH's own base rate at issue — we sit +3pts from it
calls table · base_rate
MISS
resolved 2026-10-16
peak depreciation 0.69% (stored as basis points)
resolver: fx_rates (daily USD reference rates)
The criterion

Frozen at issue

Hit if the currency's peak rate against USD inside the window exceeds the issue-day reference (11.731372) by at least 1.95%. Peak, not endpoint: a devaluation that retraces still happened.

The resolver is external — fx_rates (daily USD reference rates) — and the daily snapshot of this call in ledger-snapshots/ carries GitHub's timestamps. Diff any two dated files to verify nothing above moved after 2026-10-02.

History

Every GH currency call

2026-10-24call #511225% · resolves 2026-11-07 2026-10-23call #504525% · resolves 2026-11-06 2026-10-22call #497825% · resolves 2026-11-05 2026-10-21call #491126% · resolves 2026-11-04 2026-10-20call #484426% · resolves 2026-11-03 2026-10-19call #477727% · resolves 2026-11-02 2026-10-18call #471027% · resolves 2026-11-01 2026-10-17call #464327% · resolves 2026-10-31 2026-10-16call #457628% · resolves 2026-10-30 2026-10-15call #450928% · resolves 2026-10-29 2026-10-14call #444229% · resolves 2026-10-28 2026-10-13call #437528% · resolves 2026-10-27 2026-10-12call #430828% · resolves 2026-10-26 2026-10-11call #424129% · resolves 2026-10-25 2026-10-10call #4160said 29% · MISS 2026-10-09call #4093said 28% · MISS 2026-10-08call #4026said 29% · MISS 2026-10-07call #3958said 29% · MISS 2026-10-06call #3890said 29% · MISS 2026-10-05call #3823said 28% · MISS

← The full ledger · the same data as JSON

Daily snapshots of this book are committed to the public repository — ledger-snapshots/ — so the stated probabilities and thresholds carry GitHub's timestamps, not ours. A reader can verify no call moved between issuance and resolution without trusting us.