← orangekrakenlab · Corridors · Ledger · About

Verify the freeze — externally, with no trust in us

Our moat is a track record of predictions frozen BEFORE their outcomes are known. A freeze we verify ourselves proves nothing — so every freeze is anchored in two independent public systems: the Bitcoin blockchain (via OpenTimestamps) and this project's public git history. You can confirm any freeze existed at its stated time without trusting our servers.

Anchored, and pending — what each word covers right now

622 prediction rows are committed in a Bitcoin block — through manifest 2026-09-08, which anyone can verify today. Today's manifest (2026-09-09) holds 623 rows and is pending: submitted to the public calendar servers and waiting on a block. Until it lands, the rows it adds are git-anchored only — this page will not call them Bitcoin-anchored, and never describes a gap retroactively.

The unit is prediction rows inside a manifest — a finer unit than the immutability registry's artifacts (one artifact can hold many rows) and a different population from the ledger's rows. Manifest counts are per-day snapshots of the anchored set as then defined, so the history is not monotonic.

What that claim does and does not cover

Bitcoin leg: 622 prediction rows are committed in a Bitcoin block, through manifest 2026-09-08. The latest manifest (2026-09-09) holds 623 rows and is PENDING — submitted to the public calendar servers, awaiting a block; until it lands, the rows it adds are git-anchored only and are never described here as Bitcoin-anchored. A new manifest is stamped AUTOMATICALLY on any day a new claim is frozen (daily check since 2026-07-19; monthly-only before that).

Git leg: Each frozen record is committed and pushed to the public repo, normally the day it is frozen — GitHub's own record independently timestamps it. HONEST CAVEAT (audit-forge catch, 2026-07-21): before the daily anchor automation (2026-07-19), pushes were batched — records frozen 2026-07-10..07-18 first reached the public repo on 2026-07-19 (worst observed lag 8 days). During any such lag a record is self-attested only; the lag is disclosed here rather than hidden, and the daily automation now bounds it to ~1 day.

“Pending” means: otsStatus 'pending' = submitted to public calendar servers, awaiting aggregation into a Bitcoin block (~hours-days); 'confirmed' = provable against the Bitcoin blockchain. A freeze between stamps is git-anchored until its manifest lands — it is never retroactively described as Bitcoin-anchored for the gap period.

HONEST CAVEAT (audit-forge catch, 2026-08-01): manifests stamped BEFORE 2026-08-01 did not cover the whole record. The anchor was built from lab_ledger.json, which never read the impact boards, so 228 of the 346 registered freezes (65.9%) — including all 226 impact-board rows, every one still ungraded — were git-anchored ONLY, with no Bitcoin proof. Fixed the day it was found: the anchored set is now defined in one place (freeze_anchor.anchor_input_rows) and guard #27 (check_anchor_coverage) fails the publish gate if any registered freeze is missing from it. Earlier manifests stand as written — they are historical artifacts and are never rewritten — so a record frozen before 2026-08-01 has its Bitcoin proof from the 2026-08-01 manifest onward, and its git timestamp from the day it was frozen.

Independently verified this record and want to talk? orangekraken1@gmail.com (research/diligence/licensing inquiries; include which manifest you verified)

The first named reader’s seat is open

Nobody outside this lab has independently verified an anchored artifact yet — that sentence is the record’s biggest weakness, and replacing it is a frozen, dated commitment (rung R2 of the 2026-08-08 meta-contract: at least one named external reader by 2027-03-01, or the claim grades as a miss). Fifteen minutes makes you the first:

  1. Download the latest manifest and its .ots proof from this page.
  2. Run ots verify against it (or re-grade a mechanical claim with the public regrade.py recipe) — no account, no permission, no contact with us needed.
  3. Email what you found — match or mismatch — with the manifest hash you checked.

With your consent, your name (or handle) and your finding go on this page, dated — including, and especially, anything that does not match. The first verified mismatch report is worth more to this record than a year of praise.

Latest anchor · 2026-09-09

pending
Predictions frozen
623 prediction rows in this manifest, as of 2026-09-09 — pending, not yet Bitcoin-committed
Manifest sha256
c5df8483b7c9e5e6c7d8bdbf43b2b49a6d9b01b0bd741d82957b197ed19d6735
Bitcoin (OpenTimestamps)
submitted to public calendars, awaiting a Bitcoin block (~hours)
Git commit at freeze
743de427f5afb77b953c28dde1f49f8273df7462
git clone https://github.com/orangekraken1/omen-wealth pip install opentimestamps-client ots verify anchors/manifest_2026-09-09.json.ots

How to verify us — with no trust

  1. 1ONE CLICK — fork https://github.com/orangekraken1/omen-wealth, open the Actions tab, run 'verify-freeze'. It re-hashes every anchored manifest, recomputes every row's claimHash from its own freeze fields, walks each OpenTimestamps proof to its Bitcoin block, and cross-checks that block's merkle root on two independent public explorers. PASS/FAIL, ~2 minutes, no trust in us. (Locally: pip install opentimestamps-client && python3 scripts/external_verify.py)
  2. 2Or by hand — clone the public repo: git clone https://github.com/orangekraken1/omen-wealth
  3. 3Install the verifier: pip install opentimestamps-client
  4. 4Verify against Bitcoin: ots verify anchors/<manifest>.ots (it re-hashes the manifest and checks the hash is committed in a Bitcoin block — 'pending' means it is submitted to the public calendars and awaiting a block, ~hours).
  5. 5NOTE — `ots verify` needs a local Bitcoin node (it will report 'Could not connect to Bitcoin node' otherwise). That requirement is deliberate: checking the block header yourself is what makes this trustless. Without a node, run `ots info anchors/<manifest>.ots` instead — it prints the BitcoinBlockHeaderAttestation(<height>) the proof commits to, and you can look that block up in any explorer. That last step trusts the explorer, so it is corroboration rather than proof; a node is the only fully trustless path.
  6. 6Cross-check the hash: sha256sum anchors/<manifest> must equal the sha256 shown here and in anchor_ledger.json.
  7. 7Corroborate the git timestamp: the manifest is in the repo at the listed commit, so GitHub's own record independently dates it.

Anchor history

Each manifest is a per-day snapshot of the anchored set AS THEN DEFINED, so the counts are not monotonic across definition changes: on 2026-07-21 the manifest was partitioned to immutable predictions only (263→108 — the living registry moved out, zero frozen predictions mutated, hand-verified), and on 2026-08-01 guard #27 widened coverage to the impact boards (137→484). No prediction has ever been removed; earlier manifests stand as written.

DateAnchored setManifestsha256BitcoinCommit
2026-09-09623 predictionsmanifest_2026-09-09.jsonc5df8483b7c9e5e6…pending743de427f5
2026-09-08622 predictionsmanifest_2026-09-08.json21b247d94dde5d0b…confirmed9911281d7f
2026-09-07618 predictionsmanifest_2026-09-07.json6c8f86c201414115…confirmed5b76096310
2026-09-06614 predictionsmanifest_2026-09-06.jsonae05837c5e03e531…confirmed9809ae0b54
2026-09-05612 predictionsmanifest_2026-09-05.jsonaec1b790d24bd3a8…confirmed6435a36f19
2026-09-03597 predictionsmanifest_2026-09-03.jsoncf0a72f274a28edc…confirmed72b506a332
2026-09-02577 predictionsmanifest_2026-09-02.json4fb7b04ed1e72bee…confirmed351734550a
2026-08-30573 predictionsmanifest_2026-08-30.json41656980f657c6e5…confirmed2645859a80
2026-08-29571 predictionsmanifest_2026-08-29.json5b4bb184ae2b4025…confirmedecc17c024b
2026-08-28567 predictionsmanifest_2026-08-28.jsonef5ed51515074d56…confirmedf667a579f7
2026-08-27558 predictionsmanifest_2026-08-27.json7ae9fac5c34e3154…confirmed9ff19ed0cd
2026-08-26557 predictionsmanifest_2026-08-26.json9a9fb38406888660…confirmed02e5eb741a
2026-08-24547 predictionsmanifest_2026-08-24.json55674c2812946fd1…confirmed0cf7478dd9
2026-08-16519 predictionsmanifest_2026-08-16.jsone4d2c1634a7762dc…confirmed6e8a301a9e
2026-08-09518 predictionsmanifest_2026-08-09.jsondbdce4a8303ca86e…confirmed5274886495
2026-08-05517 predictionsmanifest_2026-08-05.json724126cfa050c8d8…confirmed62b910199d
2026-08-04489 predictionsmanifest_2026-08-04.json9cc2ab83d4e5f555…confirmed860799747a
2026-08-01484 predictionsmanifest_2026-08-01.jsoncdff4346402b8716…confirmed7f219c3d13
2026-07-29137 predictionsmanifest_2026-07-29.jsone09d7a536379a5f0…confirmeddc4861abc6
2026-07-24135 predictionsmanifest_2026-07-24.json5bc6dad6a0cb2a2e…confirmed1aee9e64c6
2026-07-234 source files (protocol freeze)world_protocol_v1.json34402264c3e20ad0…confirmedb1f5516d52
2026-07-22118 predictionsmanifest_2026-07-22.jsona66d32692285b56f…confirmed94af1fca16
2026-07-21108 predictionsmanifest_2026-07-21.jsondd9ff6da9512f590…confirmedce708be5db
2026-07-20263 predictionsmanifest_2026-07-20.jsond137228f03832a01…confirmed29a558c015
2026-07-19245 predictionsmanifest_2026-07-19.json98370a879a67645e…confirmedbbc671951c
2026-07-09134 predictionsmanifest_2026-07-09.jsona30d8c4dcea10661…confirmed69b69573d6
2026-07-07129 predictionsmanifest_2026-07-07.json31695bf9a7cfde12…confirmed2e55781937

The Direction — an open protocol, pointed at ourselves first

declared 2026-07-28

Everything on this page is the reference implementation of a larger, deliberately narrow idea: the Forward Grading Protocol v0.2 (draft) — a vendor-neutral way to measure one thing about any predictive system, human or AI: when it says 80%, does the thing happen 80% of the time? Claims are frozen before outcomes exist, anchored where no one can edit them, scored against pre-registered nulls, and graded mechanically — so the resulting number works like a credit score, not a speed test: one click to start being measured, months of un-editable elapsed time to have a score, and no way to buy, backfill, or fake one.

The instrument is aimed at this lab first, in public. As of 2026-09-09: 409 frozen claims in the lane freeze stores (440 sitewide including pre-registrations and protocol artifacts with their own grading rules), of which 102 (25%) grade through a purely mechanical, third-party-reproducible leg — a number published honestly low on purpose: since 2026-07-28, every new claim must declare its mechanical grading source at freeze time (enforced by an automated guard), and this share exists to climb. The first graded cohort — 10 claims across 6 independent thesis families, with blind pre-registered priors and decided-in-advance interpretation bands — matures 2026-08-12 through 2026-09-02. Win, lose, or mixed, the result publishes with every miss shown; at an effective sample of six families, no outcome will be called proof.

The protocol spec is versioned in the open and freezes as v1 only after that first cohort is scored and an outside re-grade of a mechanical claim matches ours. In plain words: before this machinery ever measures anyone else, it has to survive measuring us.

The manifest anchors the IMMUTABLE freeze content of each prediction (its claim, benchmark, and freeze date) — not the later grade, which is measured separately against sealed truth. So the hash only moves when a NEW claim is frozen; a silent edit to a past claim would change the hash and break the chain. That is the point.

Public proof metadata only. Not investment advice.