No forward alpha has been statistically proven yet. What is running: the point-in-time infrastructure is live, every daily signal set is anchored to a public, git-committed ledger, and every statistic on this page reproduces from published derived data with one command. The evidence-risk channel (C6) is the next out-of-sample test target — its in-sample sign is known and its forward test is defined below. Everything here is research and education, not investment advice; research classifications, not recommendations.

Continuous for display only: the dashed May-18 line marks the in-sample replay → forward out-of-sample regime boundary. All statistics are computed per-regime and never blended across the boundary. Windows are rebased to their start date.

Continuous for display only: the dashed May-18 line marks the in-sample replay → forward out-of-sample regime boundary. All statistics are computed per-regime and never blended across the boundary. Windows are rebased to their start date. Label cohorts have small, variable membership — see the archived replay for per-date n.

Research classifications, not recommendations. Membership is recomputed at every signal date; the cohort above is today's snapshot and changes daily. Evidence grades follow the public grading rubric (confidence × evidence depth); "Insufficient" appears when composite confidence < 0.30.

Qualification (point-in-time): grade A/B OR >=1 cited filing OR SourceLock-accepted event within 30d; grade 'Insufficient' excluded. Evidence-quality fields exist only from the v4.0 launch (2026-06-01), so this panel is structurally forward-only — no in-sample variant is possible without fabricating grades.

Honest result: over its first 72 periods the qualified pool trails the universe (+0.82% vs +2.53%). The cohort is too young and (at the decile level) too small for inference; it accrues as evidence coverage grows. Criteria will NOT be loosened to fatten n.

Evidence age = days since the last SourceLock-accepted filing event for the ticker (stale flag at >90d). Limitations: 72 periods; decile cohort of 2–5 names; insider-evidence stream live since 2026-07-16 (batch coverage ended 2026-05-15; gap backfilled with ingestion-time as-of) — evidence ages for some names reflect the coverage gap until live events accrue; qualification uses the public grade rubric and is recomputed point-in-time daily.

Honest reading: at the current sample sizes, no forward spread, IC, or alpha is statistically significant at the 5% level once overlap is corrected. The forward 20-day IC is positive on all observed dates but has too few independent blocks to test. "Not yet significant" is the finding — the statistics accrue daily and this panel recomputes with them.

Most platforms never test this. The comparison is pre-registered; results appear as measured. Registered primary: composite minus the top-scoring baseline per the registered argmax rule (persistence) at k=5 = -0.0230 CI (-0.0389, -0.0073) [PRELIMINARY] over 48 forward dates / 79 clustered tickers. At this sample the persistence baseline reads ahead of the composite at k=5 — printed exactly as measured; the window is young and the comparison recomputes as the record accrues. Interpretation note: yesterday's score and today's are highly overlapping (the composite moves slowly), so the persistence comparison is not between independent signals — a one-day-lagged copy of a slow-moving score can rank ahead at short horizons under reversal-like return structure without containing any information the composite lacks. Because it re-uses yesterday's own composite, this baseline measures day-over-day continuity and smoothing behavior, not an independent competitor strategy — the honest reading is that the daily composite is noisy at k=5, not that a rival wins. Investment implication: none established — no buy, sell, or alpha conclusion is supported by this page.

Registered primary: composite jointly-neutralized IC at k=5 = +0.0483 CI (-0.0085, +0.1017) [DESCRIPTIVE] — the association is not explained away by market beta, momentum, or volatility exposure, and it is also not statistically distinguishable from zero at this sample; both facts print. Component cells are exploratory: note c8's raw association largely dissolves under neutralization (factor exposure, reported as measured); blank rows are the masked (c2) and frozen (c4) components with no cross-sectional variation. not computable from owned config (no per-equity sector mapping) — stated, not proxied. Investment implication: none established — no buy, sell, or alpha conclusion is supported by this page.

Illustrative exhibits of how classifications trace to filings — never stock picks. Every event ID round-trips to the evidence store before render; full event exports are in the evidence packets. Investment implication: none established — no buy, sell, or alpha conclusion is supported by this page.

• no headline-statistic changes since the previous build

Known outage: 2026-06-26 to 2026-07-03 — snapshots continued point-in-time on frozen price inputs; zero retroactive edits (disclosed since it happened). Chain health: 30-minute monitor with alert file; push/deploy alarm: FAILURE MARKER PRESENT; deploy-verify runs after every push (byte-identity against the live site).

Investment implication: none established — no buy, sell, or alpha conclusion is supported by this page.

# packaged (v5.0+)
pip install yuclaw
yuclaw replay-lab

# or fully standalone (stdlib only, nothing to install)
curl -sO https://yuclawlab.github.io/yuclaw-brain/replay/lab_replay_bundle.json
curl -sO https://raw.githubusercontent.com/YuClawLab/yuclaw-brain/main/tools/replay_lab.py
python3 replay_lab.py lab_replay_bundle.json

The script (Python ≥3.10, standard library only; pip install yuclaw optionally adds the full SDK) rebuilds the decile cohorts from the bundled composite scores, re-derives every cohort period return, recomputes all Panel-3 statistics (same bootstrap seed), and — the tamper-evidence step — recomputes every forward snapshot's sha-256 leaf hash from disclosed derived inputs and rolls them into daily roots that must match the public yuclaw-trust ledger. It exits non-zero on any mismatch.

this build derives from: source commit ae79710b7338 · ledger block 2026-09-14 · evidence-ledger root 08c14e030cc07587… · 84 public ledger blocks

Compliant data path: the bundle contains YUCLAW-derived data only — scores, locked labels, component scores, content hashes, and derived period returns. No raw vendor OHLCV rows are exported (data-provider terms). Analyses requiring raw prices need the user's own licensed price feed.

Independent replication is invited. Run the three commands in "Reproduce this page"; the script exits non-zero on ANY mismatch between recomputed statistics and this page, or between recomputed hash roots and the public ledger. If you find a mismatch, the ledger is broken and we want to know: open an issue with the script output. Replications that confirm are equally welcome — independent verification is the point of publishing the bundle.

• · C6 out-of-sample test: the evidence/risk channel fires on ~25% of forward snapshots (rare by construction). Defined test: conditional 20-trading-day forward outcomes of C6-fired vs quiet snapshots, evaluable once the forward window exceeds the horizon with sufficient fired n.
• · Drawdown-vs-evidence-staleness study: does evidence age (days since last SourceLock-accepted filing) relate to subsequent drawdown? Evidence-age is measured and displayed today (Panel 4); the study runs when the qualified cohort has enough history — it is not fabricated now.
• · Live Form-4 ingestion: enabled 2026-07-16 — the insider-evidence stream is live in the forward era (batch coverage 2026-02-18 → 05-15; the 05-15 → 07-16 gap is backfilled with ingestion-time available_as_of and cannot affect past replays).

Definitions (exact internal rubric, deterministic verifier — no LLM in the loop):
Corpus grounding = points_grounded / points_total across the filing corpus, where an agent key-point is grounded iff it carries ≥1 citation that verifies as a verbatim (whitespace/case-normalized) span of the source filing AND every numeric token in the point appears within those verified quotes; ungrounded points are discarded with the reason recorded.
Citation fidelity = citations_verified / citations_total — the share of quoted spans an agent cites that locate as verbatim spans of the source filing after whitespace/case normalization. Verifier source: v5/swarm/grounding.py.