How PRIORAXIS is computed
PRIORAXIS is a rules-based reading of where Bitcoin stands in its market cycle. This page explains where the data comes from, what the numbers measure, and — just as plainly — what stays private and why. Everything here is meant to be checked.
The words on this site
Five terms carry everything. Here is what each one means, in plain language.
PRIORAXIS (PA). The whole system: the chart, the weekly reading, the graded history, and this site.
The PA Lens. The free public chart. It compares Bitcoin’s price with what holders actually paid for their coins, as of the last completed day. One measurement, visible whole, free forever.
The PA Engine. The instrument behind the weekly reading — formally the PRIORAXIS Value Read (PA-VR). It condenses five on-chain measures into one number and one zone. More below.
The five zones. Deep Deep Value, Value, Neutral, Elevated, Distribution — plain words for how cheap or expensive the market is compared with its own history. Five zones, three plain words: Deep Value and Value mean cheap; Neutral means fair; Elevated and Distribution mean expensive — always measured against Bitcoin’s own history, never a prediction. Both instruments speak in this vocabulary, but they are different instruments: the colored bands on the chart are the Lens’s zones; the graded episodes on The Record are the Engine’s.
The Record. The Engine’s graded history: every time it entered a decisive zone, dated, and judged only after the cycle resolved — including the reads that fell short. The Lens has no record page, because a single visible measurement needs no grading. A composite does, and gets one.
How they fit together, in one breath: the Lens shows you the terrain, free. The Engine reads five instruments and produces one output a week. The Record grades the Engine. The Engine’s current Monday reading goes to members; everything historical is public.
Where the data comes from
Every number PRIORAXIS publishes is computed from two sources we operate ourselves: the public Bitcoin blockchain, and our own multi-venue price index.
The on-chain measures are computed directly from the full transaction history of the Bitcoin network — public data anyone can verify. The price series is a volume-weighted daily index we build from the public market data of three exchanges, cross-checked daily, with a documented composition. Its history begins in October 2013.
No third-party paid feeds. No licensed data. Nothing in the published numbers can be repriced, withdrawn, or silently changed by a vendor — because there is no vendor.
One honest consequence of doing it this way: when a day’s data fails our internal checks, we publish nothing for that day rather than an estimate. The chart always shows the date of its last verified reading. A visible gap is the system working, not failing.
The PA Lens — the chart
The Lens compares Bitcoin’s market price with its on-chain cost basis — the realized price: the average price at which every coin in existence last moved. Think of it as the market’s collective purchase price. When today’s price is far above it, holders sit on large paper profits; far below it, the market as a whole is underwater.
Around the breakeven line, the Lens draws zones that describe how far price has stretched above or below what holders actually paid. In past cycles, deep readings below breakeven have coincided with capitulation and cycle floors; extreme readings above it with cycle peaks. The Lens does not predict either. It answers one question, the same way every time: where does price stand, as of the last completed day, relative to what holders actually paid?
The PA Engine — the weekly reading
The Engine condenses five on-chain measures into one number between 0 and 1, computed weekly. A low reading means the market sits in territory that, across Bitcoin’s history, has usually meant undervaluation; a high reading means the opposite. The reading places the market in one of the five zones — always relative to Bitcoin’s own history, never an absolute price level.
The Engine’s current Monday reading goes to members. The free weekly email carries the Lens read. The five measures behind the Engine:
MVRV Z-Score. Compares what all coins are worth today with the price at which each last moved — and asks how unusual today’s gap is compared with Bitcoin’s entire history. That last step matters: “expensive” and “cheap” are measured against the full record, not against a feeling. Extreme highs have historically come late in cycles; readings at or below breakeven, early.
Trend distance. Price measured against its 200-week moving average — the slowest, most stubborn trend line in Bitcoin. Every deep bear market so far has found its floor near this line, and every mania has stretched far above it. This measure captures how far the market has traveled from its own long-term base.
NUPL — net unrealized profit/loss. The share of the market’s value that exists as paper profit. When most holders sit on large unrealized gains, the temptation to sell grows with every uptick; when most sit on losses, sellers exhaust and coins gather in stronger hands. NUPL reads that pressure directly from the chain.
Reserve Risk — our own construction. A measure of long-term holder conviction. Every day a coin does not move, its holder is choosing not to sell — conviction, visible on-chain. When long-held coins finally do move, that stored conviction is spent, and it can be counted. Our version compares the market’s value with the running total of all conviction ever spent: a low reading means price is cheap relative to the resolve of the people who have held longest. We tried to reproduce the widely used third-party version exactly and could not — its precise definition is not public — so we built ours from first principles and documented every term. Where our readings differ from other providers’, that is why.
Puell Multiple. Miner economics: the value of newly created coins each day, measured against its own one-year average. Miners must sell to pay for power and hardware, which makes them the market’s structural sellers — and periods of miner stress and miner windfall have both historically clustered near cycle turning points.
Each measure is computed from the blockchain and our price index, by our own documented code, with its exact conventions recorded. Readings may differ modestly from other providers who compute the same concepts from different data or different conventions; where we know the reason, we say so.
The Record — how the Engine is graded
The Record is the Engine’s audit trail. Every time the Engine entered a decisive zone, the episode is dated and shown — and judged only after the cycle resolved, including the reads that fell short of the instrument’s own claim. Nothing is recalculated to look better afterward. You can read it in full, free, and judge the Engine’s consistency yourself.
What the testing showed
Every figure in this section counts Bitcoin, not its value in any currency. PRIORAXIS measures accumulation, and a coin-count result says nothing about value in either direction.
Two tests are reported below, and they measure different things on different bases. The first tested the pacing rule against a flat weekly benchmark, week by week. The second tested the shape that actually ships — a budget deployed over a fixed horizon. The frames are not interchangeable, so the figures are never mixed; each subsection names its own.
The benchmark itself is not a person and not a habit. It is arithmetic: the same amount deployed every single week, no missed week, no hesitation, no cost of attention — perfect frictionless execution, computed rather than lived. Nobody attempts it, in the way nobody attempts a 1:45 marathon; it is not a standard anyone falls short of, it is the ceiling the arithmetic allows. That matters for reading what follows: a result standing at 0.99 of flawless arithmetic is the strong reading of that number, not the disappointing one.
The benchmark test — contribution frame
Measured across 430 windows against that mathematical benchmark, the median result was 0.9904 — 0.96% below parity. Coverage matters more than the median, because it depends on when you happened to start: 55.3% of start dates finished below the benchmark and 44.7% finished at or above it. Of those, 179 windows finished above and 13 finished exactly at the benchmark; where the record reports 41.6%, that is wins alone. In the worst window the result was 31.2% of the benchmark’s coins.
We show this because it is what the test returned. It is the frame in which the site’s parity statement is made, and it is the less flattering of the two shown here.
The shipping rule — deployer frame
Before the figures, the finding that cuts against us, because a favourable number without it would be worth less than nothing: a rule that consults no signal at all — one that simply deploys faster — matched these figures on this history. The case for PRIORAXIS does not rest on the instrument beating anyone.
With that stated: measured over a one-year deployment horizon against the same mathematical benchmark, the median was 1.0435, with 36.7% of start dates finishing below the benchmark and a worst window of 24.8% of its coins. Over two years the median was 1.0961, with 30.1% of start dates below the benchmark and a worst window of 44.1% of its coins.
Those coverage shares are a ceiling rather than a dial we failed to tune. No price and no deployment size reaches the remainder, because below parity the benefit is negative.
Horizons are measured as 52 and 104 weeks of deployment, across 430 and 326 start windows respectively.
What the case rests on instead
One published rule, applied the same way every week, written down before it is needed and gradeable afterwards. Whether it gets followed is the part that decides the outcome, and that part is measured on your own record — not claimed here.
What we publish — and what we don’t
The Lens is public — whole, daily, forever. The Engine’s existence, its five measures, and its graded Record are public too. The Engine’s current Monday reading goes to members. And the construction — how the five measures are weighted, how the composite is built, where the zone boundaries sit — stays private.
Two reasons, stated plainly. First: the construction is what we sell; giving it away and selling it are not compatible, and we would rather say that honestly than pretend the boundary doesn’t exist. Second, and more important: publishing the recipe would feed the illusion that the number is the edge. It isn’t. The five measures are public knowledge; anyone with time and discipline could build a variation of this composite. The product is the discipline — acting on the reading, on a schedule, in public, on the record, through every kind of week. That part can’t be copied from a formula, and it’s the part this site exists to demonstrate.
Changes are on the record
From the day the Engine publishes, its methodology is on the record. Any change to how it is computed — a data source, a convention, a construction — appears in the changelog below, dated, with the before and after stated plainly. There are no silent recalibrations. A published score that can be quietly re-tuned is worthless, and we know it.
No changes yet. The changelog begins the day the Engine publishes.
What the Engine is not
It is not a prediction. It is not investment advice. And it does not tell you when to act — the Record shows the Engine’s readings against every extreme it has met, graded after the fact, flags included; what any reading means for your money is decided by your rules, not by the reading. It describes where the market stands relative to its own history — nothing more, and honestly, nothing less.
PRIORAXIS publishes a measurement, not advice. Nothing on this site is a recommendation to buy, sell, or hold any asset. The founder trades his own book; you trade yours.