BUD SIGNAL — PRE-LAUNCH DUE-DILIGENCE CLAIM (Bitcoin timestamp-anchored)
=======================================================================
Sealed: 2026-07-13  |  Anchored to the Bitcoin blockchain via OpenTimestamps

THE CLAIM (as published)
------------------------
"Pre-launch due diligence (simulated via historical replay of our production
logic — not live performance): over 3 years / 6,040 4-hour candles / 288
simulated signal windows, Bitcoin moved in the signal's direction by at least
+0.5% in 84% of windows, +1% in 67%, +2% in 46%. Average best move +2.5% per
window; typical (median) +1.8%, against a typical worst drawdown of ~1.8%.
Our live public record is consistent with — and a touch more conservative
than — these long-run simulated base rates. It is not a lucky streak, and it
is not a promise of profit."

MACHINE-DERIVED VALUES (recomputed from the sealed result set)
-------------------------------------------------------------
{"windows":288,"ge05":84,"ge10":67,"ge20":46,"meanPeak":2.5,"medPeak":1.8,"medAdverse":1.8}
"Best move" = the furthest Bitcoin ran in the signal's direction during the
window (a maximum favourable excursion), NOT a captured return — exits are the
subscriber's own. Drawdown has a fat tail: ~1 in 9 windows drew down 5%+ before
resolving. Bitcoin is volatile.

WHAT THIS TIMESTAMP PROVES — AND WHAT IT DOES NOT
-------------------------------------------------
PROVES: the complete evidence bundle behind this claim — the historical input
data, our production engine, the replay harness, and the resulting 288-window
statistics — existed in this exact byte-for-byte form on or before the Bitcoin
block referenced by the accompanying .ots proofs, and has not been altered
since. This is tamper-evident and checkable by anyone against Bitcoin, with no
trust in us.

DOES NOT PROVE: that future performance will match; that a simulation equals
live trading (it does not — this is labelled "simulated, not live"); or, by
the timestamp alone, that the historical market data is authentic (that is
separately checkable by re-pulling the market-data provider's API). The live
public record (since 2026-05-20) is a separate, separately-labelled dataset.

Our engine — the exact weighting and logic — is proprietary and is NOT
published. Full line-by-line reproduction of these numbers from the engine is
available to a qualified third-party auditor under NDA; the engine's
cryptographic fingerprint is itself sealed to Bitcoin (see the live proof page)
so an auditor can confirm they are shown the same engine that produced this.

INTEGRITY CHAIN (public verification — no trust required)
--------------------------------------------------------
This file (CLAIM.txt) is anchored to Bitcoin by CLAIM.txt.ots. The three public
files in this bundle are hashed in PUBLIC-MANIFEST.sha256, which is itself
anchored by PUBLIC-MANIFEST.sha256.ots. Verify:

  1. Install the client:   pip3 install opentimestamps-client
  2. Confirm the anchor:   ots upgrade CLAIM.txt.ots  &&  ots verify CLAIM.txt.ots
  3. Confirm nothing changed:  shasum -a 256 -c PUBLIC-MANIFEST.sha256
     (expect: CLAIM.txt OK, PROOF.md OK, outcomes-summary.json OK)

The full private evidence bundle (data + engine + harness + reports) is
committed, opaquely, by this hash — revealed only to an NDA auditor:
  SHA-256(full evidence manifest) = 7318e4a18db2f98c1299a6daded1c5de2d2116a49108156d892b1f72fff43d3e

Nothing in this public bundle discloses the engine's inputs, weighting, or
logic. It discloses only the claim, the outcome statistics, and the proofs
needed to confirm neither was altered after the seal date.