In 1952, a young economist named Harry Markowitz published the paper that taught the world how to build a portfolio. It would eventually win him a Nobel Prize; to this day, nearly every institutional portfolio on Earth — every pension, every endowment, every target-date fund holding your retirement — descends from its math. Around the same time, Markowitz faced a smaller problem: how to invest his own retirement account. The inventor of portfolio optimization did not optimize. He split his contributions fifty-fifty between stocks and bonds, and later explained why with disarming honesty: he imagined how much he’d regret missing a rally, imagined how much he’d regret riding a crash, and chose the split that let him live with either.
The man who invented the optimal portfolio sized his own for sleep.
That story is this entire chapter in miniature. Position sizing looks like a math problem, and there is math in it — we’ll do some in a moment. But the math is the easy part. Underneath, sizing is a psychology problem, and the goal is not the maximum return. The goal is a position you can actually hold through everything the last chapter described.
First, the promise from the introduction, kept and restated: this is education, not advice. No percentage in this chapter is a recommendation, because a book that has never met you — your debts, your dependents, your taxes, your tolerance — cannot allocate for you, and any book that tries is lying about what it knows. What follows is how sophisticated allocators actually reason about assets like these, laid out so you can reason alongside them.
Start with the arithmetic that makes this entire category ownable at all.
Imagine a portfolio that places two percent into bitcoin, or into the pair on this book’s cover — the monetary asset and the infrastructure, one small sleeve; the sizing logic treats them as a single decision. Now run the two endpoints.
The bad end first, at full strength: the asset goes to zero. Not down eighty percent — zero, the total loss the bear case’s exhibits gestured at. The portfolio loses two percent. That is a bad week for an ordinary balanced portfolio; markets routinely hand diversified investors a two percent move without anyone canceling dinner. Painful, survivable, bounded. You already know the worst case on the day you enter, which is more than a holder of “safe” dollars could say in 1971.
Now the other end — carefully, because this book promised no forecasts and will keep the promise. If the argument of Parts I and II plays out — if scarce, un-printable assets continue to reprice as the unit measuring them melts — the history of such repricings is that they are measured in multiples, not percentage points. No one can promise that history repeats. But the structure is knowable in advance: a two percent position that goes to zero subtracts two percent, and the same position, if the thesis is even substantially right, can contribute more to a portfolio’s decade than asset classes ten times its size.
Bounded downside. Unbounded upside. That shape — not conviction — is why serious allocators can own an asset they only half believe in. You don’t need to be persuaded by this book to justify a small position. You need only be unable to rule it out. An allocator who gives the debasement thesis one chance in five still finds the math compelling at some size, because the losing branch costs little and the winning branch pays for every year of doubt. This is also why the range you actually hear from institutions and advisors who’ve done this work clusters between one and five percent — and why the last chapter’s takeaway was meant literally. More doubt, smaller number. Less doubt, larger. Zero doubt in either direction, reread Chapter 11.
Here is the number that disciplines all the others — the one the last chapter entered into evidence: four drawdowns of roughly eighty percent, and the honest assumption that the next one begins tomorrow.
Volatility is not a flaw in these assets that patience will cure; it is the toll the trade charges, and everyone pays it in one currency or another. The only choice you get is which currency. Sized correctly, an eighty percent drawdown in the position costs your portfolio a percent or two — annoying, ignorable, boring. Sized wrong, the same drawdown costs a third of everything, and at a third of everything nobody stays rational: the position gets sold at the bottom, and temporary volatility is converted — by you, at the worst possible moment — into permanent loss.
Read that mechanism again, because it is the entire game. The size determines the behavior. The behavior determines the outcome. The asset doesn’t have to change for you to fail at owning it; the graveyard from the last chapter was filled by sizing errors long before it was filled by anything Satoshi built. The right size is not the one that maximizes what you make if you’re right. It’s the largest one that leaves you bored when it’s cut in half.
Sizing has a maintenance schedule, and it answers two questions that otherwise invite disaster.
The discipline is mechanical: pick the target, and on a calendar — quarterly, annually, it matters less than doing it — trim back to target when the position has grown and refill it when it has shrunk. That’s all. But look at what the rule quietly does. It answers when do I take profits? without a crystal ball: when the asset runs, you skim some of the run into everything else, by rule rather than by nerve. It answers do I buy the dip? the same way: when the asset bleeds, the rule adds at prices your courage never would. And it prevents the silent failure nobody plans for — the two percent that quietly became fifteen because it worked, leaving you with a concentration you never actually chose. Un-rebalanced winners rewrite your risk without asking.
Chapter 11 treated volatility as the enemy. On a rebalancing schedule, the enemy takes a job: every violent swing becomes a small, forced act of buying low or selling high. Volatility stops being the reason you can’t own the asset and becomes one of the returns to owning it correctly.
Match the horizon to the argument, or the argument will feel wrong even while it’s being right.
Everything in Part I runs on a slow clock — interest compounding across budget cycles, demographics, the deliberate one-percent-at-a-time melt. The scheduled events in Chapter 10 are dated in years. This is a regime trade, and a regime does not resolve by Christmas. Money you will need within roughly five years — the down payment, the tuition bill, next year’s living expenses — has no business anywhere near this trade, because Chapter 11’s exhibits can cut the position in half at any moment for reasons that have nothing to do with whether the thesis is true. The dollar’s decade is the argument. Your position has to be built to see the decade.
Owning these assets means choosing a door, and each door prices the same two ingredients differently: trust and effort.
Door one: the ETF. A ticker in the brokerage account you already have. Simplest by far — the taxes, the beneficiaries, the password-recovery, the estate all work the way everything else in your account works. What you own is a claim on a fund that owns the coins, which means you’ve reintroduced the intermediaries the asset was designed to remove. That is not a contradiction; it’s a trade — convenience for trust — and it’s a perfectly reasonable one if made knowingly.
Door two: the exchange account. Coins held in your name on a company’s internal ledger. Convenient, direct, and the door with the worst history in the asset class: Chapter 6 showed that essentially every “bitcoin hack” headline of the last seventeen years was one of these companies failing, not the ledger breaking. The money is just as gone either way. If you use this door, the old trading-floor rule applies: an exchange is a checking account, not a vault.
Door three: self-custody. The asset as designed — your keys, your coins, no one to trust. Which means, symmetrically, no one to call: no fraud department, no reset link, no undo. A short list of words on paper controls everything, and the technology for managing that responsibility, while improving, is not yet casual. This door offers the most sovereignty and demands the most competence, and it punishes overconfidence in both directions.
There is no universally right door. There is only the honest question of which ingredient you’d rather pay — trust or effort — and the requirement that you choose on purpose instead of by default.
One paragraph, because omitting it would be malpractice. In the United States these assets are treated as property, not currency: selling is a taxable event, swapping one coin for another is a taxable event, and yes, the rebalancing this chapter just praised generates them too. Accounts matter — the same strategy can have very different after-tax results inside and outside retirement wrappers, which is one reason Chapter 10’s 401(k) door is a bigger deal than its paperwork suggests. The rule here is the same as the disclaimer up top: this book has never met you. A professional who has is worth their fee.
Now bring back the safe-deposit box from the introduction.
Your grandfather didn’t lose eighty-seven percent of his caution because he gambled. He lost it because he made a bet without knowing he was making one — long the dollar, at one hundred percent of everything, for fifty-five years, unhedged. Nobody ever asked him to size that position, so it sized itself, and it was all-in by default.
Everything in this chapter compresses into the refusal to repeat that: know which bets you hold, and choose their sizes on purpose. The introduction made you a promise about where this was heading, and now it has the machinery behind it — the asymmetry, the boring-sized position, the rebalancing rule, the decade clock, the deliberately chosen door.
One chapter of positioning remains, because the standard menu — “own stocks, own real assets, collect your dividends” — smuggles in one last exposure to the printer, wearing an income costume. The next chapter names it.
The goal isn’t to bet the farm. It’s to make sure the farm isn’t priced in melting ice.
[Charts for this chapter: (1) the asymmetry — outcome of a 2% allocation across scenarios from total loss to historical-repricing analogs, vs. the same portfolio at 0%; (2) size determines survival — portfolio impact of an 80% asset drawdown at 1%, 5%, 20%, and 50% allocations; (3) the rebalancing effect — a fixed-target sleeve vs. buy-and-hold through two boom-bust cycles, schematic. Endnotes: Markowitz anecdote (Zweig); 60/40 volatility data; institutional allocation frameworks (e.g., BlackRock paper from SOURCES.md #7, published advisor studies on 1–5% sizing); exchange-failure history; IRS guidance on digital assets. Author flags: (1) the “one chance in five” probability framing is illustrative reasoning, not a stated forecast — confirm comfort; (2) ETH is folded into the sleeve in one paragraph per the plan’s one-decision framing — say if you want a standalone ETH-sizing beat; (3) no specific rebalancing frequency is endorsed, deliberately. All inline [VERIFY] tags resolved in the 2026-07-26 fact-check pass (see VERIFICATION-partIIIb.md).]