Once you have decided how much to risk per trade, a second question remains: on what basis do you set the size of each position? There are two broad answers. A fixed method sets size from a number that never looks at your account — a set dollar amount, a set quantity, a set dollar risk. A percentage method sets size as a fraction of your current equity, so the stake recomputes every time the account changes. The choice sounds administrative. It is not. It quietly determines whether your account compounds, whether it de-risks in a drawdown, and whether a long losing streak can end it.
01 — Two answers to one question
Every sizing method is a rule for turning a decision into a quantity — how many coins, shares, or contracts to hold. What separates the two families is a single question: does the rule reference your current account balance? A fixed method does not. Its input is external and unchanging until you choose to change it, so the position stays the same whether the account has doubled or halved. A percentage method does the opposite: it takes a fraction of present equity as its input, so the position is recalculated from the latest balance on every trade.
That one difference — whether the rule looks at the account — is the source of almost everything that follows. It decides how the equity curve grows, how it behaves after losses, and how much arithmetic you must do before each entry. Everything else is detail on top of it.
02 — The fixed family, and its three faces
"Fixed" is not one method but three, and they are routinely confused. Fixed-dollar sizing commits the same capital outlay each trade — say a set stake of exposure per position — so the number of units falls as price rises. Fixed-units sizing holds the same quantity every trade: the same number of contracts or coins, whatever they cost, so the outlay swings with price. Fixed-dollar-risk sizing fixes neither outlay nor quantity but the amount lost if the stop is hit — risk equals size times stop distance — so it is the fixed-method cousin of the percentage-risk rule.
The distinction is easy to lose but real. To illustrate — with hypothetical numbers, not a claim about any market — a trader who always commits a set outlay buys fewer units as price climbs; one who always buys the same quantity spends more as price climbs; and one who fixes dollar-risk buys whatever quantity makes the loss-if-stopped constant, which depends entirely on how far the stop sits. Three rules, three different orders, from the same chart.
What unites them is the defining trait of the whole family: the input never references equity, so none of them adjusts as the account grows or shrinks. That is their virtue and their flaw at once — simplicity bought at the price of never adapting. It also exposes a confusion worth naming early: position size is how many units you hold, while risk is how much you lose if the stop is hit — size multiplied by stop distance. The two only move together when the stop distance is fixed; the moment stops vary, a larger position can carry less risk than a smaller one.
03 — The percentage method
Percentage sizing — more formally, fixed-fractional sizing — risks a constant fraction of current equity on each trade. The position quantity follows a compact formula: N = (f × equity) ÷ per-unit-risk, where f is the chosen fraction and per-unit-risk is the dollar risk per unit, usually the distance from entry to stop. Because f is applied to current equity, the absolute stake is rebuilt from the latest balance every time.
The familiar "risk one or two percent per trade" guidance is simply this method with a small f. Ralph Vince gave the fixed-fractional approach its book-length treatment in The Mathematics of Money Management (1992), and is credited specifically with optimal f — the fraction that maximises the geometric growth rate of equity, derived from a system's worst historical loss. The broader idea of betting a fixed fraction is older and more general than any one author; Vince's contribution is the formal optimisation, which, like the Kelly criterion it refines, tends to prescribe fractions far larger than most traders can psychologically or financially tolerate.
04 — The mathematical divide
Here is where the two families genuinely part. A fixed stake makes your equity an arithmetic series: each trade adds or subtracts an amount unrelated to the current balance, and wealth is a running sum. A percentage stake makes equity a geometric series: each trade multiplies the balance by a factor, and wealth is a running product. Sums grow in straight lines; products compound.
Percentage sizing is therefore an anti-martingale structure — a generic term for betting more after wins and less after losses, the mirror image of the doubling-down martingale. As equity rises, the same fraction commands a larger stake, so gains compound on a winning run. As equity falls, the stake automatically shrinks, so the account de-risks itself in a drawdown without any decision on your part. Fixed sizing does neither: the stake is constant, so it cannot compound on the way up and does not step down on the way in. The account climbs in straight lines and takes every drawdown at full size.
Fixed sizing asks the same of your account at its strongest and its weakest. Percentage sizing asks less exactly when the account is weakest and more exactly when it is strongest — which is the whole point, and also the whole cost.
05 — Drawdown, ruin, and the recovery tax
The de-risking property has a clean mathematical edge. If every loss removes a fixed fraction of equity, the balance is multiplied by a number less than one each time — it can fall toward zero but, in the idealised model, never quite reaches it. Fixed-dollar sizing carries no such floor: because the stake does not shrink, a long enough losing streak can drive the account to zero. On paper, percentage sizing cannot be ruined and fixed sizing can.
That "cannot be ruined" claim is a model, not a promise, and it is worth saying plainly: gaps that jump past a stop, slippage on the fill, leverage, and minimum tradeable sizes all let a real loss exceed the intended fraction, and any of them can breach the theoretical floor. The same de-risking also carries a tax. After a drawdown the position is smaller, so each recovering trade earns less, and it takes proportionally longer to climb back to the old high. The mechanism that protects capital on the way down is the very mechanism that slows the return — you cannot keep one without the other.
06 — Which equity — and the cousins
Percentage sizing hides a question fixed sizing never asks: a fraction of what? Van Tharp's equity models name the choices — total equity (cash plus open positions plus open profit), core equity (a set amount committed per position as you ladder in), and reduced total equity (total less the risk currently locked in open trades). Each yields a different stake from the same account, so the definition is not a footnote; it is part of the method.
Two relatives are worth naming. Fixed-ratio sizing, from Ryan Jones's The Trading Game (1999), keys size to accumulated profit rather than balance: a delta parameter sets the profit needed to add each further contract, and position grows with the square root of profit — slower and more conservative than fixed-fractional as the account builds. The Kelly criterion is the percentage family taken to its optimising limit, prescribing the growth-maximising fraction. And volatility- or ATR-based sizing scales the stake to how much the instrument is moving, so risk stays steadier as conditions change. All are variations on the same axis: how much, if at all, should the bet track the account and the market.
07 — Choosing between them
The honest summary is that neither method wins in the abstract; each buys something with something. Fixed sizing buys simplicity, predictability, and a constant profit-and-loss per point, with no arithmetic before a trade — at the cost of never compounding, never de-risking, growing too small as the account swells and too large, in relative terms, as it shrinks. Percentage sizing buys compounding, automatic de-risking, and risk that stays proportional to the account — at the cost of slower recovery, a recalculation before every entry, and a dependence on how you define equity.
The fit follows from that. Fixed methods suit beginners, small accounts, coarse or indivisible instruments, and quick strategy testing, where their bluntness is a feature. Percentage methods are the default of systematic and money-management frameworks, where scaling with the account is the entire aim. There is a behavioural dividend too: percentage sizing mechanically cuts the stake after losses, removing the temptation to bet bigger to win it back, while a fixed stake after a losing streak is an unchanged bet that has quietly become a larger share of a weaker account.
08 — Why crypto tilts the answer
Crypto changes the calculus in specific ways. Spot coins are finely divisible — bitcoin to eight decimals, ether to eighteen — so a percentage target fills far more cleanly than in markets that force you to round to a whole share or contract. That divisibility makes the clean fixed-fractional formula genuinely implementable, though perpetual and futures contracts, with their lot sizes and minimum notionals, reintroduce the rounding that spot avoids.
Volatility pushes the same way. Because crypto's volatility is large and shifts regime, a fixed-size position can carry wildly different risk from week to week, strengthening the case for sizing that rescales with the account or with volatility. But percentage sizing meets a problem crypto sharpens: which equity? When the account itself is denominated in a volatile asset, its measured value — and every stake computed from it — moves even when you place no trades. Leverage adds another gap: a percentage risk cap limits modelled loss per trade but not notional or leverage, and a liquidation can close a position before your stop is ever reached. Continuous markets mean equity drifts around the clock, so strict percentage sizing implies constant resizing and the fees that come with it; and sizing each trade at a fixed fraction does nothing to cap total exposure when holdings are correlated — and crypto assets often move together, most of all under stress.
"A just weight and balance are the LORD's: all the weights of the bag are his work." — Proverbs 16:11
Methodology & Sources
This report describes properties of sizing methods, not their relative performance; no method is presented as superior. Definitions and formulas — the fixed-fractional equation, the anti-martingale and geometric-growth distinction, and Van Tharp's total/core/reduced-total equity models — are standardised money-management conventions. Attributions are given as documented: optimal f to Ralph Vince (The Mathematics of Money Management, 1992), fixed-ratio and its delta parameter to Ryan Jones (The Trading Game, 1999), and the Kelly criterion to John L. Kelly Jr. (1956). The idealised "cannot be ruined" property holds only in a frictionless, infinitely divisible model and is stated with its caveats. We omit volatility, correlation, funding, fee, and liquidation figures, all of which are venue- and time-specific. Related reading: position sizing and entry tiers.