THE LIVE RECORD

If this is your first visit: this page is a public logbook. Every week the method publishes its instruction before the week happens, seals it so it can't be changed, and records what the rule did. The chart compares that rule against simply buying the same amount every week. That's all this page is — you can check any line of it yourself.

One row per week since 2026-08-24, sealed before the outcome was known. Every number on this page comes from a stamped artifact you can verify yourself.

This record tracks what the engine instructed and what a rule-compliant follower would hold. It is not the operator's personal account and does not report the operator's fills.

The two lines

The follower line is a simulation of a rule, not a person or an account: a fixed weekly contribution, deployed exactly as the week's published instruction said, from the first sealed week. The benchmark buys the same contribution at the same weekly close (plain DCA). Both lines are in coins, and coins are one of two measures shown — the drawdown strip beneath carries the other, always.

The record

The criterion

A shutdown criterion was frozen on 2026-08-25, when the record was one week old. If the compliant follower's cumulative coins are more than 5% behind plain DCA at the annual evaluation, this page stops extending. The evaluation reads the last governed Monday on or before each anniversary: 2027-08-23 (published review — cannot trigger) and 2028-08-21 (armed). If that week is unverifiable, the evaluation falls back to the most recent verified Monday before it, both dates disclosed — never forward. Each evaluation also publishes whether mandatory correction campaigns fired and completed in the period. If the criterion fires, the page is frozen, not deleted: a dated notice replaces it, and the record stays reachable and verifiable at this URL.

What this number is. −59.03% is the largest decline in the value of the method's book during the calibration backtest — the four years from late 2018 to the bear-market bottom of 2022, one of the hardest windows in Bitcoin's history. The book means the Bitcoin held plus the cash reserve, with the weekly contributions inside it. The figure was measured, not chosen: it is what the method's rules actually went through on that path, and the operator signed it as accepted. For scale, Bitcoin's own price fell −75.16% over the same window — a decline of this size is nothing irregular for this asset; its volatility comes with every way of holding it. Two things this number is not. It is not a coins figure: a price fall does not take away a single coin — it changes what the coins are worth in money on that day, and this record's objective is coins. And it is not from the live record on this page: the live record draws its own drawdown, week by week, in the strip under the coins chart, from week one. It is published here because the method buys hardest when the price is lowest, and this figure is what that costs at its worst on the measured path — and because a page that shows its coins without its worst money-value week flatters, and this page does not. One more thing, so the number is sized correctly: the book here is the Bitcoin allocation and its buying reserve only — money to live on sits outside the method by design. A fixed weekly contribution buys more coins when the price is lower and fewer when it is higher; what that adds up to is what the chart above measures, week by week — nothing on this page will tell you in advance.

FINAL_CALIBRATION.md · sha256 1aa9d99f635760e8… · proof | PRICE_DRAWDOWN_2026-08-28.md · sha256 352f31fa7f690dba… · proof

Method note

The record begins at 2026-08-24, the ledger epoch. Weeks before it exist in the repository but were not sealed in advance, so they are not part of the record; they are listed separately as unsealed. The sealed record is 7 verified weeks long — the count is printed under the chart and grows by one each verified Monday. Each week's artifacts are hashed and anchored in Bitcoin via OpenTimestamps. The chart is generated from those artifacts: a week that fails re-verification is drawn as a gap and named — never interpolated, never skipped in silence. To check any week yourself, download its artifact and its .ots proof, drop both into opentimestamps.org, and compare. What the proofs establish: that nothing on this page was changed after it was stamped. They do not establish that the numbers were correct when written — the engine's published guards stand behind that. Two different assurances; the stamps are claimed only for the first. This record is live, not a backtest: every row was published before its week's outcome existed.

Markers

The method's constants change only by dated ruling. Every change is marked at its date and never back-applied — the line before a marker runs on the constants that were live then.