# Don't take our word for it — verify it against Bitcoin

Before we ever sold a subscription, we ran our **exact production signal engine**
over **3 years of Bitcoin market history** (6,040 four-hour candles). The result —
labelled for exactly what it is, a simulation, not live trading:

> **288 simulated signal windows.** Bitcoin moved in the signal's direction by
> at least **+0.5% in 84%** of them, **+1% in 67%**, **+2% in 46%**. Average best
> move **+2.5%** per window (**+1.8%** on a typical/median one), against a typical
> worst drawdown of ~1.8% — with a fat tail: about 1 in 9 windows drew down 5%+
> before resolving. "Best move" is how far price ran in the signal's favour, not a
> captured return; you choose your exits. Bitcoin is volatile.

Our **live** signals since launch (in the track record above) land in line with —
and a touch more conservative than — this 3-year simulation. Not a lucky streak.

## Why this can't have been fudged after the fact

We sealed the whole evidence bundle into the **Bitcoin blockchain** using
[OpenTimestamps](https://opentimestamps.org). Two things follow that nobody can
fake or undo:

1. **The results can't have been edited afterward.** They're anchored to a
   Bitcoin block on the seal date. Change one digit and the fingerprint stops
   matching Bitcoin — obvious to anyone who checks.
2. **The engine can't have been tuned to flatter the backtest.** Its
   cryptographic fingerprint is sealed too, *before* the results were published,
   so it couldn't have been quietly reshaped to fit them.

## Every signal from here is on the record, too

Our engine's fingerprint is anchored to Bitcoin, and from launch onward **every
signal is written to a hash-chained ledger and anchored to Bitcoin the moment it
fires — stamped to the exact engine version that produced it.** If we ever
improve the engine, that new version starts its own sealed record and the old one
stays frozen and dated. It's a running, out-of-sample track record we cannot
revise, backdate, or cherry-pick.

## Verify it yourself (no trust required)
1. Install the free checker: `pip3 install opentimestamps-client`
2. Confirm the Bitcoin 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`)

## What this proves — and what it doesn't (straight talk)
- **It proves:** the results and the engine existed in this exact form on the
  seal date and haven't been altered since; and our live record is consistent
  with the simulation.
- **It does not prove:** future performance (nothing can), and it is **not a
  promise of profit**. Trading involves real risk; signals are informational.
- **The engine stays private** — its weighting and logic are our edge.
  Independent, line-by-line reproduction is open to a qualified third-party
  auditor under NDA, who can confirm — via the sealed fingerprint — that they're
  shown the same engine that produced these numbers.

Public files here: `CLAIM.txt` (the sealed statement) · `CLAIM.txt.ots` (its
Bitcoin proof) · `PUBLIC-MANIFEST.sha256` (+`.ots`) (fingerprints of these public
files) · `outcomes-summary.json` (the numbers, machine-readable).