Open positions · fast book 34

Worst first, by the live mark. Every position is a SHORT: the book gains when the ticker falls behind SPY, so P&L and “ticker move” always carry opposite signs. Live columns are unbooked; booked is the last session the ledger has written, which is what every reported return uses.

Trade log · fast book 196 closed

⚠ 2 flagged (|ret| > 50%, P&L taken in full)

Newest exit first. Ticker and SPY are the two legs over the hold; the short earns their difference, and net is after costs. Cum is the book's realized total as of that exit (chronological), so it will not read monotonically here.

Slow book · +20 trading days

The same gated hazards the book above trades, on a four-times-longer hold. The two books share ~100% of their entries by construction, so this is NOT a horse race — the question is the paired difference per entry: what days 6–20 are worth. No verdict before 30 closed positions (~6 weeks), per-trade spread is roughly double the fast book's, and a red first month is noise.

Balance · 2026-09-14

$966.99

▼ -$1.65 overnight

The rule is the $1,000 start · 31 sessions · hover any session for its balance

64 open · $582 staked · cash $385 · params v2 · 9 closed to date · ⚠ 139 dead feed · ⚠ 1 flagged

Signal performance · by event type · +5d vs SPY

1829 alerts · 1561 graded at +5d vs SPY · unflagged, benchmarked rows only · positive excess = the signal was right, and for short rows that means the ticker UNDERperformed SPY · rate and excess stay grey until ≥5 graded

Scorecard · by source and kind · +5d vs SPY

Unflagged, benchmarked rows only. Hazards are graded short — they hit when the ticker underperforms — and are deliberately excluded from source trust, because feeding a short-side rate into a novelty trust prior would corrupt the alert path. Grey bars mean fewer than 5 graded. The tick marks a 50% coin flip.