BL-A-006: Document Evidence Register
BIJECTIVE LABS
BL-A-006: Document Evidence Register
Status: RELEASED
Classification: PUBLIC
Revision: 1
Date: 2026-09-09
Thirty dated internal documents, carrying the figures that no machine-readable artifact survives to support, together with the pre-registration that fixed the acceptance gates for the single held-out opening.
1. Why this register is separate from BL-A-001
Two kinds of evidence sit behind the numbers in this record, and they are not equally strong.
An E-nnnn identifier resolves to a file an instrument wrote. A reader opens it
and reads the value. Nothing stands between the number in the document and the
number on disk.
A D-nnn identifier resolves to a dated report a person wrote. It records what
was measured, but the measurement itself is gone: the instrument ran, produced a
figure, the figure was written into a report, and the intermediate artifact was
not retained.
Both are evidence. The second is weaker, and a reader is entitled to see which one they are standing on. Merging them into a single register would hide that, so they are kept apart and cited with different prefixes.
Where a figure has both, the artifact governs. Two figures in this record were
corrected on exactly that principle, and BL-SD-000 section 7 records them.
2. What a document identifier does not promise
It does not promise the figure reproduces. It promises a dated document states it, and that the document was written by the people who ran the measurement, at the time they ran it.
That is a real standard and it is a lower one than E-nnnn. Documents in this
class share three weaknesses worth naming.
A figure can be transcribed wrongly into a report and nothing downstream will catch it.
A figure can be correct and lose its context, so that a reader cannot tell which configuration produced it.
A report can round. BL-A-001 section 2 gives values to the precision the
artifact carries; several of these documents give one fewer digit, and that lost
digit is not recoverable.
3. The register
Ordered by date. The Carries column names the Kill Ledger rows that rest on
each document.
| ID | Date | Document | Carries |
|---|---|---|---|
| D-001 | 2026-05-28 | Directional alpha pivot | VAC |
| D-002 | 2026-05-29 | Parallel shadow portfolio | DA-B |
| D-003 | 2026-05-30 | Final pivot, maker to taker wider edge | PS-A, PS-B, PS-C |
| D-004 | 2026-05-31 | Rolling universe pivot | T1 |
| D-005 | 2026-06-02 | High-conviction confluence | OPSI-H |
| D-006 | 2026-06-14 | Strategy research on the unused data substrate | XSEC-REV, XSEC-MOM, XSEC-CAR |
| D-007 | 2026-06-17 | Open-interest accumulation strategy | OI-ACC |
| D-008 | 2026-06-22 | Dislocation verdict and lessons | DISLOC |
| D-009 | 2026-06-25 | Cointegration retired, momentum lead recorded | COINT, SPR-MOM |
| D-010 | 2026-06-27 | Cash and carry evaluation and retirement | CARRY |
| D-011 | 2026-07-02 | Venue strategy designs | S5 |
| D-012 | 2026-07-03 | Cohort persistence report | S2 |
| D-013 | 2026-07-03 | Premium reversion screen | S3 |
| D-014 | 2026-07-03 | Premium characterisation | S3-C |
| D-015 | 2026-07-05 | Independent audit of prior venue work | S5 |
| D-016 | 2026-07-05 | Multi-confirmation doctrine | OPSI-H |
| D-017 | 2026-07-06 | Venue strategy post-mortem and successors | LEADLAG |
| D-018 | 2026-07-10 | Reversion edge pre-registration | S1-MC |
| D-019 | 2026-07-10 | Momentum close pre-registration | M3B |
| D-020 | 2026-07-10 | Identity maker pre-registration | S4-G1 |
| D-021 | 2026-07-10 | Replay pre-registration | S4-R |
| D-022 | 2026-07-14 | Identity burst pre-registration | IB1 |
| D-023 | 2026-07-14 | Wallet cartography characterisation | WC0 |
| D-024 | 2026-07-14 | Whale drift pre-registration | WC1 |
| D-025 | 2026-07-14 | Maker reversion pre-registration | BR1 |
| D-026 | 2026-07-17 | Session characterisation report | TOC1, TOC2 |
| D-027 | 2026-07-17 | Reversion by session pre-registration | TOC3 |
| D-028 | 2026-07-18 | Cascade premise gate, step one | B1 |
| D-029 | 2026-07-28 | Small-capital assessment across the six doors | D5b |
| D-030 | 2026-08-02 | Held-out gate pre-registration, fragility family | H-X4 |
4. What this register changes about the Kill Ledger
Before this pass, eighteen ledger rows resolved to an artifact and thirty-three
rested on prose that had not been checked against the ledger. That is why the
ledger carried status IN REVIEW rather than RELEASED.
Every row now resolves to one identifier or the other, and every figure in the
thirty-three has been read back against the document that states it. BL-SD-000
section 8 carries the map and section 7 carries what the check found.
The ledger's status follows from that check rather than from the register existing. Registering a document proves it exists; reading the figures back against it is what makes the ledger citable.
5. The six artifacts with no surviving producer
Separate from the documents above, six machine-readable artifacts in BL-A-001
have no producing instrument left in the repository. They are listed in section 5
of that register.
Those sit between the two classes. The artifact survives and can be opened, so the figure is checkable. The instrument that wrote it does not, so the figure cannot be reproduced.
R-25 was written partway through the first programme, after an acceptance
script that lived outside the repository reported a false result and was then
lost entirely. These six predate that rule or sit outside it. Their contents are
published and their numbers are usable, and no verdict in this record rests on
any of them alone.
6. Redaction
Documents in this register are cited rather than republished. They are internal
working records: they carry role attributions, internal system names, and
operational detail that the classification policy in BL-FM-004 section 4 keeps
out of the published record.
What is published is the figure, the date, and the identifier that locates it. A client under agreement can request a redacted copy of any document in this register.
Copyright (c) 2026 Bijective Labs. All rights reserved.