Calibration in public
Accountability Ledger
Every regime call and radar name published on this site gets scored against what the market actually did — on a fixed schedule, with the misses kept on the page. This is a trader’s journal graded in public, not a performance record.
As of —
What gets scored
- Regime call alignment — each published regime read is checked against realized market behaviour at 21 trading days.
- Radar names — tickers flagged in the Pulse reports are re-checked at +5 and +21 sessions against the published observation.
- Watch conditions — “watch for X” statements are resolved against the next session.
What is not claimed
- No performance or returns claims. Nothing here is a track record of trades, P&L, or portfolio results.
- Observation language only. Entries score whether a published observation aligned with what followed — they were observations, never instructions to act.
- Misses stay published, with attribution, alongside the hits.
Current quarterscored entries · updated after each close
Open
—
Resolved
—
Hit
—
Invalidated
—
Unresolved
—
Hit rate +5d
—
Hit rate +21d
—
Regime alignment 21d
—
First scored rows publish this week — the engine went live 10 June 2026. Regime calls resolve at 21 trading days, radar names at +5 and +21 sessions, and watch conditions the next session, so the table below fills in as those windows close.
| Date | Type | Ticker | Statement | Horizon | Status | Resolved | Note |
|---|---|---|---|---|---|---|---|
| — | — | — | — | — | — | — | — |
How scoring works. Each row is a dated, published statement — a regime read, a radar name, or a watch condition — frozen at publication and never edited. Regime reads are checked 21 trading days later: did the market behave the way that regime label implies? Radar names are re-checked at +5 and +21 sessions against the specific observation that was published (not against any price target). Watch conditions resolve the next session: did the stated condition occur? A row scores hit when the published observation aligned with what followed, invalidated when it didn’t, and unresolved when the window closed without a clean read either way. Misses are published with attribution — the original post stays linked and unedited. Scoring is mechanical and runs after each close.
Personal observations of one trader. Not investment advice.