โ€บNavigation

Iteration 11 โ€” PROPOSAL: The Campaign's First Status Declarations DRAFT ยท AWAITING OPERATOR REVIEW

What happened in this iteration, in plain words: the campaign did the thing it was built for. With both instruments in the decision rule certified and unexpired (the orthogonality gauge G0 and the newly-admitted regime counter k_v), the loop drafted its first feature status declarations. The rulebook discipline held: the declaration thresholds were derived from already-committed numbers before any feature was assigned (null rate 0.264 from the frozen campaign, the standard ฮฑ and FDR conventions โ€” the bar came out at "orthogonal in at least 6 of 10 regimes, panel-corrected"). The verdict table is strikingly clean: on the flagship pair, features are either stably redundant or stably orthogonal โ€” four features clear the CONDITIONAL bar with a perfect 10-of-10 record (and so also carry the STABLE-candidate flag), zero features land in FRAGILE, and twelve fail today's orthogonality test outright. Nothing becomes official until the operator reviews this page.

Preflight (A0)

2026-07-03 14:18 UTC โ€” load1 2.73 / 32 cores (โ‰ค24) ยท no competing nasimubd jobs ยท 21 GiB available ยท clickhouse-server active ยท sidecar active + /health healthy โ†’ ALL PASS. Hermetic run (sha-pinned committed cells, zero ClickHouse), <1 s inside the cap, zero generated values.

Pre-registered declaration rule (fixed before assignment; every constant committed-anchored):
p_base = 0.264 (committed e123_v3 crypto base orthogonality rate) ยท ฮฑ = 0.05 (G3) ยท k* = 6 = min k with P(Binomial(10, 0.264) โ‰ฅ k) โ‰ค ฮฑ (exact, computed not chosen) ยท BH q = 0.10 (G11) across the 16-feature panel ยท "today" = etf_bull_2024 (latest completed slice).
FAIL-today = not orthogonal today (G0's existing verdict) ยท FRAGILE = orthogonal today, k_v < k* ยท CONDITIONAL = orthogonal today, k_v โ‰ฅ k*, BH-significant, regime scope named ยท STABLE-candidate flag = k_v 10/10 (a flag, not a STABLE declaration โ€” that still requires the full stack + F014).

The draft declarations (Tier-A BTCUSDT@250 ยท expiry: next completed crypto regime slice)

Featurek_vbinomial p vs baseDRAFT statusRegime scopeSTABLE-candidate
aggression_ratio10/103.1e-06 (BH โœ“)CONDITIONAL (DRAFT)all 10 named slices (bear_2018 โ†’ etf_bull_2024)โœ“ flagged
bar_dispersion_entropy10/103.1e-06 (BH โœ“)CONDITIONAL (DRAFT)all 10 named slicesโœ“ flagged
bar_katz_fd10/103.1e-06 (BH โœ“)CONDITIONAL (DRAFT)all 10 named slicesโœ“ flagged
bar_petrosian_fd10/103.1e-06 (BH โœ“)CONDITIONAL (DRAFT)all 10 named slicesโœ“ flagged
FRAGILE drafts: none โ€” every feature orthogonal today has a perfect stability record; the verdict is bimodal on this grid (stably redundant OR stably orthogonal). FRAGILE may first appear on other combos or on forex. ยท FAIL-today: 12 features (volume family, OHLC, duration, turnover_imbalance, several bar_*) โ€” G0's existing verdict, not new declarations; full rows in loop2_first_declarations.json.

Technical record

ItemValue
Consumption rule (frozen ยง5.3)every matrix in the decision rule ADMITTED + unexpired on crypto: G0 (grounded stack) โœ“ ยท k_v (row 15, expires at next completed crypto slice) โœ“
Ledgerrow 16 (NOTE โ€” PROPOSAL, awaiting operator review)
Coverage16 features with complete 10-slice verdicts on BTCUSDT@250 (of 26 with any data; the JSON lists per-feature coverage)
Honesty notethe 4 CONDITIONAL drafts all have FULL 10/10 scope โ€” they are conservative assignments of features that are really STABLE-candidates; CONDITIONAL is the strongest status assignable today (STABLE needs F014's identifiability-sharpened falsifier, Frontier #3)
Harnessloop2_first_declarations.py ยท loop2_first_declarations.json โ€” hermetic, sha-pinned panel, <1 s
Compute / prod impactzero ClickHouse ยท zero writes outside audit folder + dashboard

โ–ถ Next iteration

Iteration 11 ยท 2026-07-03 ยท Loop-2 first consumption ยท PROPOSAL (draft, not official) ยท hermetic ยท capped ยท zero synthetic data ยท append-only