46 exhibits · every chart in the framework
The Framework in Pictures
All of the framework’s exhibits, grouped by the page that teaches them and shown in the order the book presents them. Each chart makes one claim and shows the mechanism behind it. Hover over or tap any part of a chart for its explanation (on a desktop, a click pins it in place), and open Details at the foot of the chart for its sources and methodology.
The marker beside each title tells you what you are looking at: a square is plotted from a public data series, a diamond is a representative exhibit (an illustrative shape) or a simulation (computed from stated inputs), and a circle is a conceptual diagram of framework logic. One of the 46 is plotted from public data: 02 · The Bill Came Due.
L marks the framework’s three lenses and S its two signature shapes; they appear on the home page and in Part 1. Numbered exhibits belong to their Part.
Opening Foundation Lineage & Macro The Backbone Tax Architecture Construction Scoring
Opening
The shape of the whole system.
L1 · Capital Has Weather: Same assets, different regime, different behavior
The same portfolio behaves differently in different macro regimes.
A conceptual four-quadrant map of growth versus inflation, with quadrants labeled stagflation, reflation, deflation and goldilocks. An illustrative path traces capital from a disinflation boom through a deflation scare, reflation and an inflation shock, and ends at an illustrative “Now” marker near the center.
Illustrative regime path; every waypoint, including “Now”, is placed by hand. Leadership in each quadrant is the framework's reading of common market experience, not a measured result. Quadrant names follow common market usage, not any one published model. Conceptual diagram. Illustrative framework exhibit, not historical data. Leadership labels are the framework's judgment.
Sources: Rule stated in Part 2 · Identification, evaluation, and governance; Rule stated in Part 1 · The traditional playbook was built for a regime that has ended; ACF dashboard · Regime reading from the direction of rates, inflation and unemployment (software behavior as of September 2026).
L2 · The Window Opens: A new position waits out the range, adds on confirmation, then lets the move run
A new position earns its full weight only after the market confirms the thesis.
A conceptual price path for one new position. It stays range-bound inside a soft compression band, clears the range and is confirmed, then accelerates upward into an asymmetric release.
Conceptual path for one new position, with no price levels implied; the Bitcoin backbone and Ballast run on different clocks. Conceptual diagram. Illustrative framework exhibit, not historical data.
Sources: Rule stated in Part 5 · Momentum: conviction requires confirmation.
S2 · Bend the Tail: Hold the average fixed and see what the framework changes around it
With the average held fixed, ACF is designed to thin the deep-loss tail and lengthen the gain tail, paying for it with more small shortfalls.
Two outcome distributions with the same area and the same average. The symmetric reference is a normal bell centered on the base case. The ACF curve is thinner in the deep-loss tail, piles up just below the base case, is thinner through moderate gains, and runs longer into the far right tail.
Conceptual outcome distributions with the same area and average; illustrative, not measured returns. Conceptual diagram. Illustrative framework exhibit, not historical data.
Sources: Rule stated in Part 1 · The traditional playbook was built for a regime that has ended; Rule stated in Part 5 · Three behaviors, one portfolio; Rule stated in Part 2 · The intellectual foundations, thesis-agnostic.
Foundation
Why the traditional playbook is failing.
Part 1 also runs as a single page of its eight exhibits, in reading order: Part 1 in Pictures.
01 · Assets Outran CPI: Consumer prices, US home prices and an equal mix of stocks, homes and gold, end of 1999 = 100
From the end of 1999 to the end of 2024, US homes and an equal mix of stocks, homes and gold outran consumer prices by a wide margin.
Three lines indexed to 100 at the end of 1999. Consumer prices rise steadily to about 188 by the end of 2024. US home prices climb to about 185 at the July 2006 peak, fall to about 135 at the February 2012 low, and reach about 325. An equal mix of stocks with dividends, homes and gold reaches about 617. The shaded area between that mix and CPI shows how far these assets outran CPI.
Drawn through the published figures at the marked points. The path between them is illustrative, not plotted data. Representative framework exhibit. Sources support the underlying concept.
Sources: FRED (BLS) · CPIAUCSL · Consumer Price Index for All Urban Consumers: All Items in U.S. City Average; FRED (S&P Dow Jones Indices) · CSUSHPINSA · S&P Cotality Case-Shiller U.S. National Home Price Index; Aswath Damodaran (NYU Stern) · Historical Returns on Stocks, Bonds and Bills: 1928-2024 (updated January 5, 2026): S&P 500 with dividends and gold.
02 · The Bill Came Due: Federal debt held by the public and federal interest outlays, as a percent of GDP, fiscal years 1980 to 2025
From 1991 to 2021 federal debt held by the public more than doubled as a share of GDP while falling rates cut the interest bill by more than half. By 2025 the bill was back at its 1991 peak, on more than twice the debt.
Two stacked line charts on a shared timeline, fiscal years 1980 to 2025, both in percent of GDP. Top: federal debt held by the public, about 25 percent in 1980 and 44 percent in 1991, down to about 31 percent in 2001, then up to 98 percent in 2020, 94 percent in 2021 and 98 percent in 2025. Bottom: federal interest outlays, 1.8 percent in 1980, rising to a 3.2 percent peak in 1991, falling to a 1.2 percent low in 2015 and 1.5 percent in 2021, then climbing to 3.0 percent in 2024 and 3.2 percent in 2025, level with the 1991 peak marked by a reference line.
Fiscal years 1980 to 2025, plotted as published by OMB via FRED (retrieved September 30, 2026). FRED divides each fiscal-year figure by calendar-year GDP. Historical data: plotted from a named public series.
Sources: FRED (OMB) · FYPUGDA188S · Gross Federal Debt Held by the Public as Percent of Gross Domestic Product; FRED (OMB) · FYOIGDA188S · Federal Outlays: Interest as Percent of Gross Domestic Product.
03 · Correlation Turns: How US stocks and 10-year Treasuries moved together, 2014 to 2024. Below zero, bonds tended to cushion stocks; above it, they tended to fall together.
The stock-bond correlation turned positive when inflation became the main stress.
Line chart of the 24-month rolling correlation of monthly returns on US stocks and 10-year Treasuries, 2014 to 2024. The line stays below zero from 2014 through 2021, between about minus 0.2 and minus 0.55, deepest in 2020 and near minus 0.35 at the end of 2021. It crosses zero in early 2022, inside a shaded inflation-shock stretch, and stays between about plus 0.5 and plus 0.75 through 2023 and 2024.
Drawn through the published figures at the marked points. The path between them is illustrative, not plotted data. Shorter windows turned positive in mid-2021; the 24-month window drawn here crosses in early 2022. The 2023 and 2024 levels are drawn, bracketed by published 12-month and 36-month readings for 2024. Representative framework exhibit. Sources support the underlying concept.
Sources: Brixton, Brooks, Hecht, Ilmanen, Maloney and McQuinn (AQR) · A Changing Stock–Bond Correlation: Drivers and Implications. The Journal of Portfolio Management 49(4), 64–80 (2023); Marco Lombardi and Vladyslav Sushko (BIS) · The correlation of equity and bond returns. BIS Quarterly Review, December 2023, Box A; State Street Investment Management · Mind on the Market, Chart of the Week (October 2, 2025). S&P 500 and Bloomberg US Treasury Index, monthly: the 12-month stock-bond correlation peaked at 0.80 in July 2024, the 36-month at 0.66 in December 2024.
04 · The Hedge Broke: Total return of US stocks, US bonds and a 60/40 mix through 2022, indexed to 100 at December 2021
In 2022 the bonds in a 60/40 portfolio fell with the stocks instead of cushioning them.
Line chart of total return through 2022, indexed to 100 at December 2021. Stocks fall to about 76 at the September low and end the year near 82. Bonds, the supposed hedge, fall alongside them to about 85 in September and 84 in October, and end near 87. The 60/40 mix falls to about 80 and ends near 84, marked by a line for its year-end loss of about 16 percent.
Drawn through the published figures at the marked points. The path between them is illustrative, not plotted data. Representative framework exhibit. Sources support the underlying concept.
Sources: Janus Henderson Portfolio Construction and Strategy Team · Reports of the death of 60/40 have been greatly exaggerated (2023). 2022, S&P 500 TR and Bloomberg US Agg TR: 60/40 −16.1%; Aswath Damodaran (NYU Stern) · Historical Returns on Stocks, Bonds and Bills: 1928-2024 (updated January 5, 2026). 2022, S&P 500 with dividends: −18.04%; Bloomberg · Bloomberg US Agg Total Return Value Unhedged USD (LBUSTRUU).
05 · Path Changes Everything: The same 12 annual returns and the same withdrawals, in two orders: gains first or losses first
Same returns, same withdrawals, different order. At a 4 percent withdrawal, losses first ends with less than half what gains first ends with; at 6 percent it runs out in year 12.
A 12-year withdrawal simulation. A row of 12 annual returns is shown twice: gains first, and the same returns reversed, losses first. Below, two portfolio paths are computed from those returns, each starting at $1,000,000 and withdrawing $40,000 at the end of every year. Gains first rises to about $2.39 million in year 7 and ends near $1.16 million. Losses first falls to a low near $321,000 in year 8 and ends near $475,000, less than half as much. A line at zero marks where the money would run out.
A 12-year simulation. The twelve returns are illustrative, not a market record. Representative simulation. Computed from stated inputs, not a forecast or a historical backtest.
Sources: Author simulation · Twelve annual returns in two orders; both paths computed from them.
S1 · Shape the Payoff: The payoff shape the whole framework is built to produce, and what it costs
Keep the left tail survivable and let the right tail run, within the concentration limits.
A conceptual payoff curve. A straight symmetric reference line loses on adverse outcomes as much as it gains on favorable ones. The ACF curve runs slightly below that line near the base case, which is the cost of the structure. On the left it flattens toward a level bounded by sizing; on the right it bends upward and accelerates.
Conceptual; the framework limits loss through sizing and exits, not a guaranteed floor. Conceptual diagram. Illustrative framework exhibit, not historical data.
Sources: Rule stated in Part 1 · The traditional playbook was built for a regime that has ended; Rule stated in Part 2 · The intellectual foundations, thesis-agnostic.
L3 · Govern the Thesis: A thesis earns exposure, and tripwires decide when to stop and check
ACF is a closed-loop operating system, not a fixed set of weights.
A simple left-to-right path: thesis, exposure, risk, a tripwire checkpoint and a governed response, with a return arc showing that evidence updates the thesis. The tripwire triggers a check before any response is chosen.
Conceptual governance loop; the operating tripwires are in Part 5, and the signals the dashboard watches are on the Math page (as of September 2026). Conceptual diagram. Illustrative framework exhibit, not historical data.
Sources: Rule stated in Part 1 · The system at a glance; Rule stated in Part 5 · Governance through the cycle; ACF dashboard · Tripwire signals the dashboard watches (software behavior as of September 2026).
06 · Survive the Path: A fixed monthly purchase over seven years through two deep falls, against the total put in (100 = everything invested)
Buying a fixed amount of Bitcoin every month from January 2018 to December 2024, through two falls of more than 70 percent, still ended at about seven times the money put in.
Two lines over seven years. A straight line climbs to 100, the total put in by a fixed monthly purchase. The value of those purchases, drawn along a price path with two falls of more than 70 percent, starts at the first contribution, rises, drops about 70 percent from its peak in the second fall, and ends near 700, about seven times the money put in. Falls from the running peak are shaded.
A drawn price path with two falls of more than 70 percent, like those of 2018 and 2021 to 2022; it is not Bitcoin's price history. The dates and the seven-times multiple quoted above come from Coinbase prices and are not plotted. Representative framework exhibit. Sources support the underlying concept.
Sources: Rule stated in Part 3 · Optimizing for multi-cycle survivability; FRED (Coinbase) · CBBTCUSD · Coinbase Bitcoin. Source of the dates and the seven-times multiple quoted with this chart; not plotted; Author calculation · Fixed monthly purchase along the drawn price path.
Lineage & Macro
Regime, thesis, and the policy backdrop.
P2-01 · Method Before Macro: One fixed axis for the method and one moving line for the macro thesis, across three regimes
The method stays fixed while the macro thesis is revisited as regimes change.
A single conceptual field. A flat horizontal axis for the method runs through the center. The macro-thesis line runs above it in regime A, crosses it at the first regime transition, runs below it in regime B, and crosses back above it in regime C. The axis never moves; only the thesis does.
Regimes A, B and C are placeholders, not dated periods, and the line has no scale. Conceptual diagram. Illustrative framework exhibit, not historical data.
Sources: Rule stated in Part 2 · The intellectual foundations, thesis-agnostic; Rule stated in Part 1 · The traditional playbook was built for a regime that has ended.
P2-02 · Ruin Comes First: Two portfolios with similar volatility and very different worst cases
Survival comes before optimization, because ruin cannot be undone.
Two value paths with similar early volatility. One draws down and recovers. The other falls through a point-of-no-return line, is forced out, and flatlines well below where it started, never recovering.
The paths have no scale; the line marks the idea of a forced exit, not a particular percentage loss. Conceptual diagram. Illustrative framework exhibit, not historical data.
Sources: Rule stated in Part 1 · The traditional playbook was built for a regime that has ended; Rule stated in Part 2 · Nassim Nicholas Taleb: Antifragility & ruin avoidance.
P2-03 · Conviction Needs an Exit: Size may rise with conviction, but only inside the concentration cap and with an exit written in advance
Conviction earns a larger position only inside the concentration limits, and only with an exit set in advance.
Position size against conviction. The framework line stays at zero below a CIS of 50, then rises through the sizing bands and reaches the concentration cap only at the highest conviction. The conviction-alone line starts above zero, crosses the cap near the middle and keeps rising into unbounded risk.
Sizes follow Part 5’s Torque bands with no scale shown. The cap and the exit are written rules you carry out; the dashboard flags a position above the cap, records every trade and never blocks one (as of September 2026). Conceptual diagram. Illustrative framework exhibit, not historical data.
Sources: Rule stated in Part 2 · Stanley Druckenmiller: Conviction & asymmetric positioning; Rule stated in Part 5 · Torque: leverage on regime forces.
P2-04 · Markets Feed Back: Price changes capital behavior; capital changes fundamentals; fundamentals feed back into price.
Price can change the fundamentals it is supposed to reflect.
Reflexivity as one mechanism in a single row of five stages: price, capital, buildout, fundamentals and validation. Each stage holds two states. A reinforcing path runs left to right through the top states (price rises, capital flows in, buildout expands, fundamentals improve, validation confirms) and arcs back above the row into price. A reversing path runs through the bottom states (price falls, capital tightens, buildout slows, fundamentals weaken, validation breaks) and arcs back below the row into the same price. Both loops close back into price.
A diagram of Soros’s reflexivity loop; it says nothing about how fast or how far any loop runs. Conceptual diagram. Illustrative framework exhibit, not historical data.
Sources: Rule stated in Part 2 · George Soros: Reflexivity & regime feedback loops.
P2-05 · Time Changes Prudence: The same 10 percent a year for 70 years, compounded tax-free or taxed every year
A long horizon magnifies tax drag.
Two growth multiples over seventy years from the same ten percent annual return. The tax-free line reaches about 17 times its start at year thirty and about 790 times at year seventy. The line taxed every year compounds at seven and a half percent and reaches about 9 times its start at year thirty and about 158 times at year seventy. On this linear scale the two look close for decades, then split wide.
Tax-free means qualified Roth growth under current law. The taxed line assumes every year’s gain is realized, the costly case; a gain realized once costs less (Part 4). Representative simulation. The same 10 percent a year, tax-free versus taxed every year at a 25 percent blended rate; no volatility, fees or withdrawals; not a forecast.
Sources: Author calculation · 1.10^t tax-free vs 1.075^t (10% a year less a 25% blended tax on gains realized every year) · year 30: 17.4× vs 8.8× · year 70: 790× vs 158×; Rule stated in Part 2 · Tax as a structural return multiplier.
P2-06 · Liquidity Sets the Tide: One drawn liquidity cycle, and an asset like Bitcoin moving with it on a wider swing
Bitcoin has tended to move in the same direction as global liquidity over multi-month windows.
Two lines over one drawn cycle. A global-liquidity line rises and falls gently around the middle. An asset line moves the same way on a much wider swing, with short wiggles that sometimes run against the liquidity line. A dot marks where liquidity peaks and turns.
Drawn shapes, not plotted liquidity or Bitcoin prices. Swing sizes and wiggles are illustrative; the cited study measures direction, not a fixed multiple. Representative framework exhibit. Sources support the underlying concept.
Sources: Lyn Alden Investment Strategy · Bitcoin: A Global Liquidity Barometer (written by Sam Callahan, commissioned by Lyn Alden), September 2024; Rule stated in Part 3 · Why Bitcoin can be modeled; Author illustration · Method the shape illustrates: one liquidity cycle drawn as a smooth wave; the asset moves in the same direction on a wider swing, plus a faster wiggle that sometimes runs against it. No lag is drawn, and heights and widths are drawn, not measured.
P2-07 · Narrative Is Not Thesis: Four tests stand between a good story and a thesis: persistence, a capital-flow path, falsifiability and a multi-year runway
A narrative earns thesis status only by surviving four gates.
A validation gauntlet. A narrative enters and must pass four gates (structural persistence, a capital-flow implication, falsifiable indicators and a multi-year runway), and only what survives all four emerges as a thesis. A survivor band thins at each gate.
The gates are Part 2’s four tests for a valid thesis; how many stories fail at each is drawn, not counted. Conceptual diagram. Illustrative framework exhibit, not historical data.
Sources: Rule stated in Part 2 · Identification, evaluation, and governance.
P2-08 · Capital Finds the Bottleneck: A structural force becomes investable where it runs into a constraint, because that is where capital has to flow
Capital concentrates where a structural force hits its binding constraint.
A five-stage descending cascade: structural force, required buildout, bottleneck, capital pathway and investable exposure. Each stage transforms the one before it until the thesis becomes something you can own.
A reasoning sequence from Part 2’s thesis-construction steps; it names no assets and measures no flows. Conceptual diagram. Illustrative framework exhibit, not historical data.
Sources: Rule stated in Part 2 · Identification, evaluation, and governance.
P2-09 · Phase Changes Sizing: One valid thesis through three phases: early structural, mid-cycle buildout, late expression
A thesis can stay valid while the right size for it changes with its phase.
Across a thesis lifecycle (early structural, mid-cycle buildout, late expression), validity stays high and flat while sizing starts small, rises to a peak through the buildout, then eases into the late phase.
Phase is a judgment recorded with the thesis. The dashboard stores it and does not size positions by it (as of September 2026); the curve shows the idea, not a sizing rule. Conceptual diagram. Illustrative framework exhibit, not historical data.
Sources: Rule stated in Part 2 · Identification, evaluation, and governance; Rule stated in Part 5 · Torque: leverage on regime forces.
The Backbone
Bitcoin as the convexity backbone.
P3-01 · Exposure Is Not Control: Three allocations meet the same shock. The one that counts is the one you can still follow afterward.
Maximum exposure breaks under a messy path; the framework allocation survives it and still participates.
An interactive stress test, not a time series. Three allocations run as horizontal lanes through a shock zone in the middle: maximum exposure at 100 percent Bitcoin, the framework reserve at 15 percent Bitcoin with dry powder and an income sleeve, and a stress-tested reserve at 10 percent Bitcoin with half the portfolio in dry powder. The same shock hits all three. Maximum exposure climbs steepest before the shock, then plunges; under a job-loss shock its lane breaks at a forced sale. The framework lane bends, stays whole and keeps capacity to act. The stress-tested lane dips least but climbs slowly. A readout below grades each allocation under the current shock: whether it still participates, whether it keeps a reserve, how likely it is to force a mistake such as a sale at the bottom, and whether you could stick with it. Terminal value is a small secondary figure. In the opportunity window the reserves buy during the drawdown, though not at the exact bottom.
Three whole-portfolio allocations on one made-up market path; the control and participation scores are illustrative. Representative simulation. Computed from stated inputs, not a forecast or a historical backtest.
Sources: Author simulation · Three allocations on one shared market path: 100% Bitcoin; 15% Bitcoin with 15% dry powder and an income sleeve; 10% Bitcoin with 50% dry powder and an income sleeve; Rule stated in Part 3 · Optimizing for multi-cycle survivability.
P3-02 · Accumulate, Don’t Trade: Fixed-dollar DCA buys more units when Bitcoin is lower
Fixed-dollar DCA buys more units when price is lower, and the framework leans in further only when Bitcoin is deeply undervalued.
A representative Bitcoin price index over one cycle, drawn above a row of dollar-cost-averaging unit bars. Price starts below fair value, dips into a deep undervalued trough, recovers toward fair value, then runs extended. Because units received equal fixed dollars divided by price, the bars are taller when price is lower. While price is deeply undervalued the framework buys about 50 percent more, so its accent bars rise above the baseline; nearer fair value it buys the baseline; once price is extended it halves discretionary buying and banks the difference as dry powder. The readout, +26%, is the framework's units per dollar against plain DCA (28 percent more units for 2 percent more dollars). The reserve is never sold.
A representative price index, not historical prices, so the bars show relative units only. The readout compares units per dollar, and the reserve is never sold. Conceptual diagram. Illustrative framework exhibit, not historical data.
P3-03 · Volatility Is the Toll: What a deep Bitcoin fall costs a portfolio that holds Bitcoin as a sized reserve
At a 15 percent reserve, a 77 percent Bitcoin fall costs the portfolio about 11.5 percent: painful, and survivable.
Two representative lines, both starting at 100. Bitcoin rises, falls about 77 percent, climbs to a new high, falls about 54 percent and partly recovers. The portfolio line, computed from it with Bitcoin at 15 percent, dips about 11.5 percent in the first fall and about 8 percent in the second.
Representative Bitcoin path, not historical prices. The portfolio line is computed from it, with Bitcoin reset to 15 percent at each new high and the rest held flat. The reset is a simplification: the framework never trims the reserve, so a reserve left to grow goes into the next fall as a larger share and loses more. Representative framework exhibit. Sources support the underlying concept.
Sources: CNBC · March 5, 2024 · Galaxy Digital’s Alex Thorn: “bitcoin has seen four 75% [plus] drawdowns”; its record before 2024 was $68,982.20 on November 10, 2021; Author calculation · Portfolio line = Bitcoin at 15 percent (reset at each new Bitcoin high, untouched through each fall) + 85 percent held flat; the Bitcoin path is illustrative, drawn to falls of about 77 and 54 percent; Rule stated in Part 1 · The traditional playbook was built for a regime that has ended.
P3-04 · Power-Law Corridor: Price against a log power-law corridor: a trend line with euphoria and capitulation bands
On log-log axes, Bitcoin’s price has clustered around a rising power-law trend, with swings above and below it.
A log-log schematic with no price levels or dates. A rising power-law trend line runs through a shaded corridor. A representative price path swings through it, spiking above the upper euphoria band twice and closing below the lower capitulation band once, and ends back inside the corridor at a later reading.
Representative schematic, not plotted prices: the axes carry no price levels or dates, and the path is drawn. The fitted constants move with the data window, and no band is a hard floor: price has spent months below power-law floor lines, most recently in 2026. Representative framework exhibit. Sources support the underlying concept.
Sources: Giovanni Santostasi · Medium, March 2024 · The Bitcoin Power Law Theory: price as a power law of days since the genesis block (first posted on Reddit, 2018); Harold Christopher Burger · Medium, September 2019 · Bitcoin’s natural long-term power-law corridor of growth: the corridor with support and resistance bands; Author illustration · Method the shape illustrates: a straight power-law trend on log-log axes, with bands a fixed distance above and below it; the price path is drawn, not fitted.
P3-05 · Models Must Converge: The valuation models are most useful where they agree; when they fan apart, caution rises
The valuation models earn trust where they converge, and the discount to that converged estimate sets the pace of buying.
A conceptual valuation index with a shaded range between the highest and lowest model estimates. The range starts wide, narrows to a tight convergence mid-chart, then widens again into divergence. The price line moves inside the range throughout and sits near its top where the models converge.
The envelope spans four valuation models. The dashboard computes no model convergence; you read it yourself (as of September 2026). Conceptual diagram. Illustrative framework exhibit, not historical data.
Sources: Rule stated in Part 3 · Why Bitcoin can be modeled.
P3-06 · Cold Storage to Borrow: The reserve lifecycle: accumulate, self-custody, mature, and optionally borrow instead of selling
A mature reserve can supply cash by borrowing against it instead of selling it, if you choose to borrow at all.
A five-stage reserve lifecycle, left to right: accumulate, self-custody, mature reserve, an optional collateralized loan, and liquidity with no sale while the loan holds.
Borrowing is optional. A loan supplies cash without a sale, but it adds interest cost and the risk of a forced sale: if the collateral falls far enough, or a payment is missed, the lender sells it, and that sale is a taxable event. Tax treatment can change. Conceptual diagram. Illustrative framework exhibit, not historical data.
Sources: Rule stated in Part 3 · TAM, custody, and the borrow phase.
P3-07 · Reserve Share Evolves: A 15–20% reserve can grow into a 30–50% share over 15 to 20 years, and borrowing against it then becomes an option
A starting reserve can grow into a share large enough that governing it matters more than building it.
Reserve share as a percent of net worth, rising over twenty years from about 17 percent, past the 20 percent top of the target range near year 9, across the 30 percent mature guide just after year 15, to about 46 percent at year 20. A mark at year 17 shows where borrowing against the reserve becomes an option.
Illustrative trajectory, not a forecast for any portfolio. These percentages describe outcomes under given conditions, not rebalanced targets. Representative simulation. Computed from stated inputs, not a forecast or a historical backtest.
Sources: Author calculation · Illustrative trajectory: share = 17 + 29 × (year ÷ 20)^2.9 percent, with small drawn noise; Rule stated in Part 3 · TAM, custody, and the borrow phase.
Tax Architecture
Wrappers, basis, and what you keep.
P4-01 · Routing the Dollar: Where each kind of new capital goes, and why that wrapper
New money is routed by the kind of position it buys, in a set order, with no fixed split between accounts.
A routing diagram. Six nodes on the left name kinds of new money: contributions up to the employer match, Torque that may rotate, never-sold Bitcoin, return-of-capital Ballast with harvesting satellites, ordinary-income payers, and a high-income year. Three nodes on the right name wrappers: Roth, taxable and pre-tax. Torque connects to Roth. Never-sold Bitcoin and return-of-capital Ballast connect to taxable. Ordinary-income payers and the high-income year connect to pre-tax. The match connects to both Roth and pre-tax: your own contribution can go into a Roth 401(k) bucket, and the match itself lands pre-tax unless the plan offers Roth matching.
A conceptual map under current US federal rules. The order (match first, Roth IRA next, then by position type) is Part 4’s routing sequence; the diagram shows only where each kind of dollar lands. Conceptual diagram. Illustrative framework exhibit, not historical data.
Sources: Rule stated in Part 4 · Three wrappers, three roles; IRS · 2026 IRA contribution limit ($7,500; $1,100 catch-up at 50+), shared across traditional and Roth IRAs; IRS · Notice 2025-67 · 2026 traditional IRA deduction phase-outs for workplace-plan participants ($81,000–$91,000 single; $129,000–$149,000 joint when the contributor is covered); IRS · Notice 2024-2 · SECURE 2.0 Roth matching contributions: the plan may permit them, only for fully vested matches, taxable in the year allocated.
P4-02 · The Tax Wedge: What each account lets you keep of every dollar inside it, as a winning position grows
The after-tax gap between a Roth and a taxable account that sells widens as the win grows.
Three lines show what each account keeps of every dollar inside it as a winning position grows from one times to thirty times. The Roth line keeps the full outcome and reaches thirty times. The taxable line starts with it at one times and rises less steeply, keeping the principal plus about 76 percent of each unit of gain, to about 23 times at the right edge; the widening gap between the Roth and taxable lines is shaded as the tax wedge. The pre-tax line starts lower, at 0.7 times, because its whole withdrawal, principal included, is taxed as ordinary income; it keeps about 70 percent all the way and reaches about 21 times. The chart does not show the deduction a pre-tax contribution earned going in.
Per dollar inside each account, at representative federal rates: 23.8 percent on a large one-time sale and about 30 percent blended on pre-tax withdrawals spread over years. The deduction a pre-tax contribution earns going in is not drawn; at the same tax rate going in and coming out, pre-tax and Roth leave the same after-tax money. Illustrative; the numbers depend on income, filing status, state tax and the rules in force. Representative framework exhibit. Sources support the underlying concept.
Sources: IRS · Publication 590-B · Traditional IRA distributions are taxed as ordinary income; qualified Roth IRA distributions are not taxed; IRC · 26 U.S.C. §1(h) · Long-term capital gains rates (top rate 20 percent); IRC · 26 U.S.C. §1411 · The 3.8 percent net investment income tax; IRS · 2026 IRA contribution limit ($7,500; $1,100 catch-up at 50+), shared across traditional and Roth IRAs; Author calculation · Roth 100% · taxable keeps the principal plus ~76% of the gain (23.8% on a large one-time sale) · pre-tax ~70% of the whole withdrawal at a representative 30% blended ordinary rate (withdrawals spread over years) · illustrative; depends on income, filing status and state.
P4-03 · ROC Changes the Yield: One distribution, two tax treatments, and how much of it goes back to work
Return of capital defers the tax, so more of each distribution goes back to work now.
Two aligned bars split the same 100-unit distribution. The return-of-capital bar sends the full 100 back to work now, with a hatched strip below it marking about 24 units of tax deferred into a lower cost basis, due at a later sale. The taxed-dividend bar, taxed at the top federal rate, redeploys about 76 units now and gives up about 24 to tax. A dashed line marks the 76 point; the return-of-capital bar reaches past it to 100, and that surplus of about 24 units is the capital still working now under return of capital.
Conceptual comparison at the top federal rate on qualified dividends (23.8 percent). Return of capital is not tax-free: it lowers your basis, the deferred gain is taxed when you sell unless a step-up at death resets the basis first, and once basis reaches zero further distributions are taxed as capital gain when received. Under current law; characterization is set each tax year and can change. Conceptual diagram. Illustrative framework exhibit, not historical data.
Sources: Rule stated in Part 4 · The work taxable accounts do; IRS · Publication 550 · A nondividend distribution (return of capital) reduces basis and is not taxed until basis is fully recovered; after that it is capital gain; IRC · 26 U.S.C. §301(c) · The part of a distribution that is not a dividend reduces the stock’s basis; any excess over basis is treated as gain; IRC · 26 U.S.C. §1(h) · Qualified dividends are taxed at the long-term capital gains rates (top rate 20 percent); IRC · 26 U.S.C. §1411 · The 3.8 percent net investment income tax.
P4-04 · Gross Is Not Net: One gross balance, split into the part you keep and the part you owe
A pre-tax statement shows 100 percent, and part of it is tax you have not paid yet.
A donut divides one gross pre-tax balance into two shares: about 70 percent is the net you keep, drawn in the framework accent, and about 30 percent is the deferred tax claim, drawn muted. A horizon control (today, plus 12 years, plus 25 years) grows the balance from one times to about 8.6 times, at about 9 percent a year, while the 70/30 split holds, so the claim grows from 0.3 of a unit to about 2.6 units. The split assumes a steady 30 percent rate; the claim’s final size is set by the rate when the money comes out.
Conceptual composition at a representative 30 percent blended rate on withdrawals, with the balance growing about 9 percent a year. The claim’s final size is set by your rate when the money comes out. Tax is deferred, not eliminated, under current law. Conceptual diagram. Illustrative framework exhibit, not historical data.
Sources: Rule stated in Part 4 · Pre-tax accounts: inferior, occasionally useful; IRC · 26 U.S.C. §72 · Ordinary-income tax on qualified-plan and annuity withdrawals.
Construction
Postures, sizing, and the rules of change.
P5-01 · Three Jobs. One Cycle. How each posture behaves when the market advances, breaks, and recovers
Each posture behaves differently through the same market cycle, and that behavior is what the framework classifies.
Four indexed value paths cross one stylized market cycle divided into advance, stress and recovery phases. Torque climbs steepest, falls roughly sixty percent through the stress phase and recovers to finish highest. Ballast stays close to flat, dipping about twelve percent in stress; a ring at the trough marks the rotation moment where reserves deploy into Torque. Hype spikes fastest, collapses early in the stress phase and ends at a dot marked stopped at breakeven, with no recovery path. A dashed Bitcoin line on its own lower register falls about seventy-five percent in the stress phase, deeper than Torque, and recovers; a note marks it as never rotation capital.
Indexed shapes on one stylized cycle, not returns or forecasts. Bitcoin is drawn on a lower register only to keep it visually separate, and its fall of about 75 percent sits at the shallow end of the falls of 75 percent or more that Part 3 says to plan for. Conceptual diagram. Illustrative framework exhibit, not historical data.
Sources: Rule stated in Part 5 · Three behaviors, one portfolio; Rule stated in Part 3 · Bitcoin as the convexity backbone.
P5-02 · Position Size Must Be Earned: Six stages of evidence, from an idea worth investigating to a proven platform
Position size is earned through evidence: the most a thesis could justify is not what the evidence supports today.
An ascending ladder of six evidence stages with no numeric axis. Stage one, thesis exposure: the opportunity is real enough to investigate. Stage two, commercial validation: customers, contracts, backlog or adoption begin confirming demand. Stage three, initial execution: management turns opportunity into measurable delivery. Stage four, scale execution: success repeats without breaking the model. Stage five, economic proof: margins, cash generation and operating leverage validate the business. Stage six, exceptional platform quality: durability, scarcity, execution and optionality together, the strongest evidence a position can show. The exhibit carries no percentages and assigns no score bands.
Percentages come from the score, through Part 5’s posture sizing tables and under its concentration caps. Conceptual exhibit. The six evidence stages are the framework’s written sizing ladder and carry no sizing numbers.
Sources: Rule stated in Part 5 · Torque: leverage on regime forces; Rule stated in Part 5 · Adds and freezes: discipline on the way in.
P5-03 · Ballast Preserves the Right to Buy: The reserve that does the buying when Torque goes on sale
Ballast is rotation capital: the reserve that turns a drawdown into an allocation decision.
A governed cycle of five stations: harvest strength by trimming Torque that has outgrown its limits, rebuild reserves toward their working range, wait without urgency, deploy into validated weakness at a governed checkpoint, and participate in recovery. A return arc closes the cycle back to harvesting the next run.
Bitcoin sits outside this loop by rule and never funds it. Conceptual diagram. Illustrative framework exhibit, not historical data.
Sources: Rule stated in Part 5 · Ballast: the reserve that buys the drawdown; Rule stated in Part 3 · Bitcoin as the convexity backbone.
P5-04 · The Same Score Does Not Create the Same Position: The score measures quality; the posture decides what that quality is worth in capital
The same CIS score earns a different position size in each posture.
Three columns on one shared percent scale, each with its own ceiling mark. Torque stacks three ranges, two to four percent for scores in the fifties, four to eight for the sixties and eight to fifteen at seventy and above, under a posture ceiling at fifteen percent that only a documented override extends to eighteen. Ballast stacks one to three, three to five and five to eight percent for the same score bands, under an exceptional single-position maximum of ten percent, with aggregate Ballast working between twenty and thirty-five percent, never below a fifteen percent floor or above a thirty-five percent ceiling. Hype shows a single two to five percent range under a hard five percent cap, eligible from a CIS of fifty, with a ten percent aggregate cap. A single dashed rule above all three columns marks the eighteen percent portfolio outer bound; each posture ceiling binds before it.
Hype has no score bands: it is eligible from CIS 50 and sized within Part 5’s 2–5% range under hard caps of 5% a position and 10% in aggregate. Each posture’s own ceiling binds before the 18% maximum, and the 15% Ballast floor is Part 5 doctrine. Conceptual comparison. Ranges and caps are the framework’s sizing parameters (Part 5 tables).
Sources: Rule stated in Part 5 · Torque: leverage on regime forces; Rule stated in Part 5 · Sizing and placement; Rule stated in Part 5 · Hype: quarantined, capped, and pre-committed to exit; ACF dashboard · The portfolio builder sizes Hype from CIS 50 at no more than 5 percent a position and 10 percent in aggregate; on the book you hold it flags a Hype position above 5 percent or a sleeve above 10 percent (software behavior as of September 2026).
P5-05 · Conviction Does Not Eliminate Binary Risk: A 10% position walks the earnings protocol: trim · observe · assess · re-earn
The framework cuts exposure before a binary event, and the position earns its size back afterward.
A timeline from twenty-one trading days before earnings to fifteen days after. The position holds ten percent through the initiation-blackout window, is trimmed to the three percent cap in one step at T-5, and holds the cap through the announcement and the five-day assessment window. From day six it splits into three branches: thesis confirmed rebuilds toward ten percent, adding no more than a quarter of the target a week and reaching eight percent by day fifteen; thesis uncertain stays at three percent; thesis damaged exits to zero. A dashed rule marks the three percent cap.
The dashboard flags a position above the 3% cap in the days before its report. It counts calendar days, not trading days, so the flag can arrive a trading day or two after T-5; start the trim from the earnings calendar instead of waiting for the flag. It trims nothing: the trim is yours, and so is the T-21 to T-6 initiation blackout (as of September 2026). Conceptual protocol path. The 3% cap, the window boundaries and the branch rules are Part 5 doctrine; the 10% starting position is illustrative.
Sources: Rule stated in Part 5 · Earnings proximity: cap the binary event; ACF dashboard · Earnings flag on a position above the 3 percent cap as its report nears, counted in calendar days; it trims nothing (software behavior as of September 2026).
P5-06 · Conviction Requires Market Confirmation: Absolute trend, relative performance, and breadth decide how much size a thesis can carry
Full size needs three momentum confirmations, and losing all three is an exit even at a high score.
A validation gauntlet. A high-conviction position enters from the left and passes three gates: absolute momentum, its own trend measured from the fifty-two-week high; relative momentum, its performance against the sector and the market; and breadth, participation across related names. The surviving band thins at each gate, and only a position confirmed on all three reaches full sizing. Gate order is illustrative; the action depends on how many dimensions break: one broken reduces sizing twenty-five to thirty percent, two put the position on watch with minimal new exposure, and all three are an exit even at a high score.
Doctrine gauntlet; the action ladder is the Part 5 protocol. The dashboard does not yet measure the three dimensions; it watches stress across your holdings through a confluence tripwire: flag, hedge, then a displayed 25% Torque-trim playbook it does not execute (as of September 2026). Conceptual diagram. Illustrative framework exhibit, not historical data.
Sources: Rule stated in Part 5 · Momentum: conviction requires confirmation; ACF dashboard · Momentum Death tripwire reports unavailable; the cohort confluence tripwire escalates from a flag to hedging only to a displayed 25 percent Torque-trim playbook it does not execute (software behavior as of September 2026).
P5-07 · When Correlation Rises, Diversification Shrinks: Positions that looked independent can become one trade when liquidity drains
When correlation rises, the framework reduces exposure to the shared failure mode and makes no call on direction.
A governed cycle of four stations. Normal: correlations contained, liquidity functioning, standard sizing rules. Stressed: the VIX sustained above twenty-five, credit spreads widening, Torque correlation above zero point seven. Throttle, the governed checkpoint: pause new adds, hold posture limits strictly, and consider cutting gross exposure ten to twenty percent. Repair: indicators normalize. A return arc labeled stable two-plus weeks closes the cycle back to normal rules.
Doctrine protocol, not market timing. The dashboard watches the VIX and credit spreads through its macro tripwires and flags average pairwise correlation above 0.7 across all non-cash holdings; the Torque-only reading and the throttle actions are yours (as of September 2026). Conceptual diagram. Illustrative framework exhibit, not historical data.
Sources: Rule stated in Part 5 · Liquidity and correlation: throttle the regime; ACF dashboard · Macro tripwires on the VIX (watch 20, caution 25, critical 35) and high-yield credit spreads; the Correlation Spike tripwire flags average pairwise correlation above 0.70 across non-cash holdings and reports unavailable until every holding has 60 daily returns (software behavior as of September 2026).
P5-08 · One Regime Force. Multiple Economic Expressions. One thesis, seven businesses that fail in different ways
High conviction in one regime force can still be diversified, across the force’s distinct economic channels.
A fan diagram. One node on the left, an illustrative regime force labeled AI infrastructure, connects to seven channel nodes on the right: compute, memory and networking, power generation, grid equipment, cooling, datacenter owners and builders, and physical security and defense. Each channel is a distinct bottleneck within the same regime force.
The seven channels are examples of how one force can be expressed; the list is illustrative, not exhaustive. Conceptual methodology illustration. AI infrastructure is the example force, not a recommendation.
Sources: Rule stated in Part 5 · Allocate by force, not by sector; Rule stated in Part 2 · Identification, evaluation, and governance.
P5-09 · Tax Drag Compounds Too: Identical asset, identical return, different wrapper: $100,000 over 30 years
The same asset earning the same return ends about 2× apart after 30 years, depending only on the wrapper.
Two curves start together at one hundred thousand dollars in year zero. The tax-free curve compounds at ten percent and reaches about one point seven four five million dollars by year thirty. The after-tax curve compounds at seven and a half percent and reaches about eight hundred seventy-five thousand dollars. Checkpoints mark both values at years ten and twenty; the shaded gap between the curves is the wrapper drag, roughly eight hundred seventy thousand dollars at year thirty.
A constant-return illustration of tax drag at federal, illustrative rates: the taxable account realizes every gain every year, and Roth withdrawals are assumed qualified, which makes them tax-free under current law. Representative simulation. Computed from stated inputs, not a forecast or a historical backtest.
Sources: Author calculation · $100,000 × 1.10^t (tax-free) vs × 1.075^t (10% pre-tax return, all gains realized every year at a 25% blended rate); 10y $259k vs $206k · 20y $673k vs $425k · 30y ≈$1,745k vs ≈$875k; Rule stated in Part 4 · Reserving Roth for Torque.
P5-10 · Know What You Are Changing: Some changes calibrate the system; some suspend a rule; some abandon the framework
Calibrating a parameter, documenting an override and touching doctrine are three different acts with three different meanings.
A routing diagram. One node on the left, a proposed change, connects to three levels on the right. Doctrine: defines the framework; changing it means you are no longer implementing ACF. Parameters: calibrate implementation; adjustable from their documented defaults, with a stated reason, without changing what the framework is. Overrides: temporary, explicit, reasoned, time-bounded, with reversion conditions.
The three levels and their examples are stated in Part 5’s change-governance section. Conceptual diagram. Illustrative framework exhibit, not historical data.
Sources: Rule stated in Part 5 · Doctrine, parameters, overrides.
Scoring
Two scores, and the weekly workflow that reads them.
P6-01 · CIS Measures the Position, Not the Portfolio: One asset, four components, no portfolio context
CIS weighs four components, and at the reference weights convexity carries the most because it is what the framework is built to capture.
A composition donut of the Convexity Integrity Score at its reference weights. Convexity and optionality fill 40 percent of the ring; risk and fragility 25 percent, where a higher score means greater survivability; macro alignment 25 percent; execution and sentiment 10 percent. The center is labeled CIS, a position-level score from 0 to 100.
The weights move with the thesis you select: under the Capital Preservation profile, Risk (35 percent) outweighs Convexity (30), and under Conflict Economy the first three tie at 30 (as of September 2026). Conceptual composition. Drawn at the reference weights. The dashboard’s scores use the active thesis profile’s weights (each within 0.10 of these, renormalized) and move weight from Macro to Convexity and Risk when macro evidence is thin. The four-component structure is doctrine.
Sources: ACF dashboard · Reference weights C40/R25/M25/E10; thesis-profile bounds ±0.10, floor 0.05, ceiling 0.50, renormalized; macro weight cut to 60 or 80 percent at low or medium macro confidence (software behavior as of September 2026); Rule stated in Part 6 · CIS: scoring the position, not the portfolio.
P6-02 · FIS Starts at 100: Five capped penalty buckets, and every deduction names the rule it breaks
FIS is subtractive: every point below 100 is a named, repairable violation.
A waterfall chart. The score starts at 100 and steps down through five deductions for one example book: allocation minus 4, governance minus 1, dead capital minus 1, concentration minus 10, complexity minus 1. It lands at 83, inside the Strong band. Horizontal guides mark the band boundaries at 70, 60 and 50.
The deductions belong to one example book scored under the framework’s FIS rules. The dashboard’s score does not yet see all of its inputs: nothing sets the momentum-breakdown flag, and the thesis, distribution-type, hold-rationale, remediation-plan and override records you keep on a holding are not passed to it, which in the current build bills every position as undocumented and unclassified (as of September 2026). Conceptual diagram. Bucket caps, thresholds and bands follow the FIS rules; the deduction sizes belong to this example only.
Sources: ACF dashboard · Five FIS buckets capped at 25/15/15/15/10; governance and dead capital weighted by position share between 0.2 and 12 percent; concentration charges 8/6/4 (software behavior as of September 2026); ACF · worked example · Author calculation under the FIS rules: 11 positions, 80 percent taxable and 20 percent Roth; top three 41.5 percent and top five 61.5, none above 15; four documented names scoring in the 60s at 6.25 percent each; a fifth of value without a thesis; one unclassified distribution. Allocation 3.85, governance 1.00, dead capital 1.00, concentration 10, complexity 1, FIS 83.15, drawn rounded; Rule stated in Part 6 · Subtractive scoring: every lost point has an owner.
P6-03 · Good Assets Can Still Form a Bad Portfolio: CIS grades what you own; FIS grades what it becomes when assembled
CIS never reads FIS, and FIS reads CIS only through its four bands; neither score can stand in for the other.
A two-by-two matrix. The horizontal axis is construction integrity, FIS, split at 70, the action line. The vertical axis is position quality, CIS, split at the Strong band’s floor of 70, a band boundary, not an action trigger. Top right: maintain, strong positions well assembled. Top left: repair construction and keep the positions. Bottom right: sub-core conviction; honor the posture-specific sizing bands and strengthen, resize or replace only as warranted. Bottom left: construction first, then reassess and size each position by its CIS band. A path runs from the bottom left through the bottom right to the top right, showing the repair order.
Conceptual diagnosis matrix. Both axes split at 70, the Strong band’s floor on each score and, on FIS only, the action line. In the dashboard, FIS below 70 produces a fix-the-top-penalty recommendation when you run the Weekly Review (as of September 2026). Conceptual diagram. Illustrative framework exhibit, not historical data.
Sources: Rule stated in Part 6 · Neither score compensates for the other; ACF dashboard · The shared four-band register (70 is the Strong floor on both scores); FIS below 70 raises a fix-the-top-penalty recommendation in the Weekly Review (software behavior as of September 2026).
P6-04 · The Weekly Evidence Loop: Regular measurement, explicit attribution, action only on a governing threshold
The framework runs on a weekly evidence loop: measure, check, act or hold, record, repeat.
A weekly operating loop of five stations: update CIS for every position, calculate FIS for the portfolio, run the governance checks (earnings, momentum, tripwires, posture drift), then act or deliberately hold, and log what changed and why. A return arc labeled weekly closes the loop back to the first station.
The dashboard’s Weekly Review runs the same order when you start it: a budgeted refresh of stale scores, then ingest, read CIS and FIS, governance checks, action determination, decision log. Its governance checks do not yet receive earnings dates, price trends or correlation data, so the earnings flag comes from the Framework Rule Register and the positions table instead (as of September 2026). Conceptual diagram. Illustrative framework exhibit, not historical data.
Sources: Rule stated in Part 6 · The weekly workflow; ACF dashboard · Weekly Review order and trigger precedence: three-level check, earnings, FIS below 70, CIS drift (software behavior as of September 2026).
P6-05 · Failure Rarely Arrives All at Once: What happens to a healthy portfolio in the twelve months after reviews stop
Framework failure is usually slow: small tolerable deviations compound while they go unmeasured.
A conceptual chart of twelve months without reviews. A dashed reference line holds flat at 100, labeled still assumed healthy. Beneath it five normalized indicators erode, each steepening as the months pass: stale conviction slides first and falls furthest, thesis evidence follows, correlation stacking holds up for months and then drops fastest of all, posture drift steepens gently, and wrapper leakage loses least. An early marker notes that a point-in-time audit two to three months in still looks acceptable.
The erosion tempos are illustrative; the diagnostic thresholds cited for the lines are the dashboard’s (as of September 2026). Conceptual diagnostic. Normalized health indicators, no market data.
Sources: Rule stated in Part 6 · How implementations decay unnoticed; ACF dashboard · Stale-score warning at 60 days and failure past 90; refresh flag at 50 percent above cost on a score older than 60 days; posture drift warning at 10 points and failure at 20; Correlation Spike tripwire above 0.70 average pairwise correlation (software behavior as of September 2026).