Why Position Sizing Is the Most Important Decision You Make
Your entry tells you where to play. Your position size determines whether you survive long enough to find out if you were right.
There is an asymmetry at the heart of every market that most traders never fully reckon with.
Every participant contributes to the market’s structure. Your trades, like everyone else’s, become part of the price action that follows. This is the nature of a complex adaptive system. There is no outside from which to act on the market, only the inside in which all participants shape and are shaped by what emerges.
But contribution is not parity. The market as a whole absorbs the action of any single participant without depending on it. The collective outlasts any individual contributor to the collective. The market continues. Your ability to participate in it does not.
In the early days of what became the modern casino, operators discovered a structurally similar problem the hard way. A single wealthy player on a lucky run could threaten solvency. Exposure was concentrated. The house edge was real but the path to realising it was fragile.
Over time, through failure and near-disaster, they learned that survival required spreading risk across so many independent outcomes that no individual sequence could place the institution in danger. Only then did the long-run edge become reachable.
Traders face an analogous structural problem from the opposite side of the table, and solve it with the same tool. Not the signal. Not the market selection. Not the entry timing. The position size. How much of your capital is exposed to any single outcome is the decision that determines whether your programme survives long enough for its edge to matter.
Everything in Foundation 1 was preamble to this.
The Arithmetic of Survival
The relationship between losses and recoveries is not linear. This is the single most important mathematical fact in systematic trading, and it is the one most consistently misunderstood at the level of intuition.
A 10% loss requires an 11% gain to recover. A 25% loss requires a 33% gain. A 50% loss requires a 100% gain. A 75% loss requires a 300% gain. The numbers are exact and they are asymmetric in a way that compounds with severity.
Losses shrink the base from which the recovery must be made. Every percentage point of loss is more expensive than the equivalent percentage point of gain is valuable.
The practical implication is this. A drawdown that appears survivable in percentage terms may be effectively unsurvivable in compounding terms.
A programme that enters a 50% drawdown needs to double the remaining capital before it is back to where it started. At typical compounding rates, that recovery takes years. Years in which the capital is not compounding forward. Years in which the psychological strain of the drawdown is testing the commitment to the process. Years in which the most likely outcome is abandonment of the system before the recovery arrives.
Position sizing is the lever that controls how deep those drawdowns can go. Get it right and even long sequences of losses remain within the region of practical recovery. Get it wrong and a normal losing streak, the kind every programme encounters and every backtest contains, becomes an event from which the programme does not return.
Why the Path Matters More Than the Average
Standard finance treats risk and return as if they can be summarised by their averages. Expected return is what you make, on average, across many trials. Standard deviation is how variable those returns are. If the expected return is positive, the system has an edge.
This framework lives in an ergodic world. One where the average across many parallel universes equals the average across a single long timeline. In an ergodic world, running a positive-expectancy system for long enough guarantees convergence toward the expected outcome.
Markets are not ergodic. Your trading career is not ergodic. There is only one path through time, and the sequence of outcomes along that path matters enormously.
A severe loss early in the sequence shrinks the capital available to benefit from subsequent gains. Two traders with identical average returns but different sequences of those returns end up at different places. The one who suffered the deep early loss ends up poorer, sometimes dramatically so, even if every individual trade result was drawn from the same distribution.
This is path dependence, and it is the reason why maximising expected return is the wrong objective. The correct objective is maximising the geometric return. The rate at which capital actually compounds along the single path you live.
The geometric return is always lower than the arithmetic average return, and the gap between them grows with the volatility of the return stream. More volatility means a lower geometric return, even when the arithmetic average is unchanged.
A precise distinction matters here, one this essay returns to in detail later. The volatility that damages geometric return is volatility in the realised compounding base, the closed equity that actually represents the programme’s progress through time. Volatility in unrealised, open-position equity is a different phenomenon entirely, and is treated differently.
For now, when this section refers to volatility of the return stream, it means the volatility of realised results. That is what position sizing primarily controls, and that is what compounds.
Position sizing is the primary control over the volatility of the realised return stream. Reducing position size from the over-sized region toward the optimum lowers volatility per closed trade, which raises the geometric return by reducing compounding drag.
The relationship is not infinite. Past a certain point, positions become too small to deploy the available capital usefully, and the geometric return falls again from the opposite direction. But for almost every trader who has not deliberately sized down, the available improvement lies in the direction of smaller, not larger.
In a non-ergodic system where path matters, that improvement often translates to more accumulated capital over time, even though the per-trade gains feel smaller.
“You do not get to choose whether losing streaks will arrive. You only get to choose how much damage they do when they come.”
The Ruin Boundary
There is a region of capital from which recovery is possible and a region from which it is not. The boundary between them is not a matter of preference. It is geometric.
At a 90% drawdown, recovery to the starting point requires a 900% gain. At that depth, the programme has not merely suffered a setback. It has crossed into territory from which return within any realistic timeframe is essentially impossible.
The capital technically exists, but the ability to participate meaningfully in markets does not. Ruin, in the practical sense, has occurred.
The ruin boundary defines the outer limit of how much risk any position can carry. A position sizing rule that exposes the programme to a realistic path toward that boundary is not a rule at all. It is a delayed guarantee of elimination.
Operating far inside the ruin boundary is not timidity. It is the architecture of survival. The programme that remains solvent through ten years of adverse regimes, difficult drawdowns, and long periods without outliers is the programme that is present when the fat-tail event arrives. Presence is the precondition of everything else.
ATR-Based Sizing: Turning Survival Into a Formula
Once survival is understood as a geometric problem, position sizing needs a measuring stick. In systematic trend following, that measuring stick is volatility.
The standard approach uses the Average True Range as the denominator. The ATR measures how much a market moves on an average day, capturing the current volatility of that specific instrument at that specific moment.
The logic is precise. If you commit to risking a fixed dollar amount per unit of ATR, your position size adjusts automatically to the volatility of the market being traded. In a volatile market, the ATR is large, and the position is smaller. In a quiet market, the ATR is small, and the position is larger.
The result is that each trade carries approximately the same dollar risk regardless of which market is being traded or what volatility regime it is currently in.
This solves several problems at once. It prevents a single volatile market from dominating the portfolio’s risk simply because its daily moves are larger. It normalises exposure across markets that have different nominal prices, contract sizes, and historical volatility profiles. And it automatically reduces position size during volatile periods, precisely when the risk of a large adverse move is greatest.
The specific multiplier applied to the ATR varies between programmes and practitioners. What does not vary is the underlying principle. Position size is a function of current market volatility, not a fixed number of contracts, not a fixed percentage of the notional position value, not a judgment about the quality of the signal. The formula runs. The trader executes.
The fixed dollar amount risked per unit of ATR is itself derived from closed balance equity, the realised compounding base. The next section addresses why this specific equity figure governs sizing rather than total equity, and how the same realised base then determines when the programme as a whole reduces exposure under stress.
For now, note only that the per-trade calculation and the programme-level response to drawdown share the same foundation. Both run off the closed equity figure. Neither runs off the open position value or the high-water mark.
The formula does one thing with precision and one thing only. It normalises risk across markets so that each position carries approximately the same expected dollar movement per day, regardless of which instrument is being traded or what volatility regime it currently occupies. What it cannot do is assess the compression state of the market being sized into.
A quiet market with a small ATR produces a large position. A volatile market with a large ATR produces a small one. The formula is blind to whether the quiet market’s low volatility represents genuine calm or accumulated tension approaching a release.
This is not a flaw. It is a boundary condition, and one that is managed at a different level of the programme entirely. The compression risk that Foundation 5 addresses, the specific danger of markets that have been quiet long enough to warehouse significant unrealised energy, is handled by the breadth of diversification established in Foundation 3.
A single quiet market producing a larger-than-average position is a position-level risk. A programme positioned across many markets, each at a different point in its own volatility cycle, absorbs the release of any single market’s compression without threatening survival.
The position sizing formula governs individual trade risk. The portfolio construction governs what happens when the formula gets the quiet market wrong.
The Cut Back Rule: When the Programme Defends Itself
A complete position sizing framework does not only govern individual trade risk. It responds to the state of the programme as a whole. But what it responds to, and what it ignores, depends on which measure of equity is treated as the line that cannot be breached.
This is where most traders confuse two very different forms of equity.
There are two equity figures in any live programme, and they behave very differently.
Closed balance equity is the realised compounding base. Calculated precisely, it is the cash balance in the account plus the cost basis of any open positions. It is not the cash balance alone, and it is not the cash balance plus the unrealised mark-to-market value of open positions. The cost basis of an open position is the price at which the position was entered, preserved in the closed balance figure precisely so that mark-to-market fluctuations do not move it. A trader who starts with one hundred thousand dollars and deploys fifty thousand dollars of margin into open positions still has one hundred thousand dollars of closed balance equity. The figure changes only when a position is closed, at which point the realised gain or loss is added to the cash balance, the position is removed from the open register, and the closed balance figure updates by the realised amount. This is the compounding base, the figure from which position sizes are calculated, and the figure that determines whether the programme is genuinely making money or merely watching open profits fluctuate.
Total equity, by contrast, includes unrealised profit and loss on every open position. It moves every day. It can be substantially higher than closed equity during a strong trending period, and can give back significant portions of that excess as trends retrace before the trailing stops are struck.
The Outlier Hunter sizes positions from closed balance equity, not total equity. This is not a technicality. It is the structural choice that makes letting winners run mathematically possible.
The reason is asymmetric. There is no path to capturing fat-tail returns without accepting fat-tail risk on the positions that are producing them. The trades that define the year’s performance run for months, accumulate large unrealised profits, and give back significant portions of those profits at the end as the trend exhausts and the trailing stop is finally struck.
A programme that responds to every retracement of unrealised equity by cutting position size is a programme that cannot hold an outlier to its conclusion. The mechanism designed to protect capital becomes the mechanism that prevents the capital from ever being made.
The resolution is to be precise about what is being risked.
Position size is set off closed equity, which means the initial bet is small relative to the capital actually available for compounding. Once the position is open and moves into profit, the unrealised equity sitting between the high-water mark and the trailing stop is what the trade swings with.
That unrealised equity is risk capital in the strict sense. It can be given back without threatening the compounding base, because it was never part of the compounding base in the first place. What is risked, in the open trade, is future profit. What is protected, by the sizing rule, is realised capital.
This distinction governs when the Cut Back Rule fires and when it does not.
Consider what happens when unrealised equity retraces. A winning trade gives back open profit on its way to the trailing stop. The Cut Back Rule does not engage. The programme is not in a drawdown in the relevant sense. It is in the structural mid-life of a trend following position, doing exactly what a trend following position is supposed to do.
The equity curve will be volatile through this period. That volatility is the cost of being present for the outlier, and it has been priced in to the design.
The opposite case is when closed equity itself is threatened. The sequence of realised results is genuinely adverse. Losing trades are accumulating into the realised account rather than open trades retracing toward their stops. Here the Cut Back Rule engages.
At specific drawdown thresholds in closed equity, defined in advance, the entire programme scales position size down proportionally. Not panic-driven reduction. Not discretionary reduction based on a judgment about whether the drawdown feels like a regime change or bad luck. A formulaic reduction that runs on the realised compounding base, the only figure that represents the programme’s actual trajectory through time.
A simple example illustrates the mechanism. If closed equity falls by a defined amount, say ten percent from the prior peak, the programme reduces all new position sizes by a defined proportion. If the realised drawdown deepens to a deeper threshold, the reduction increases again. When closed equity recovers, exposure scales back up according to the same rule, in the same steps, in reverse.
The thresholds and proportions vary between programmes. The mechanism does not. The programme defends itself by stepping back from the ruin boundary at predefined points, then stepping back toward full exposure as the realised account is rebuilt.
This is the Cut Back Rule in its essential form. It is the mechanism by which a programme under genuine stress automatically moves itself further from the ruin boundary. It does not require the trader to feel the stress and respond wisely. It runs mechanically, as every rule in the system must.
When closed equity recovers, sizing scales back up. The programme returns to full exposure as the realised account is restored, not before.
The Cut Back Rule is position sizing extended across time, applied to the only equity figure that matters for survival. The individual trade rule governs how much realised capital is exposed to any single position. The drawdown response rule governs how the programme protects realised capital when the sequence of closed results is adverse.
Unrealised equity is left free to do its work, swinging for the outliers within the boundary defined by the trailing stop. Both rules are necessary. Neither requires judgment at the moment of application.
“Sizing is not a detail to optimise after the entry logic is established. It is the architecture the entry logic depends on to mean anything at all.”
The Psychological Trap of Oversizing
Oversizing is the most common and most costly error in systematic trading. It is easy to understand why. Larger positions mean larger gains on winning trades. The psychological reward of a large win is immediate and vivid.
The cost of oversizing is delayed, statistical, and arrives all at once in the form of a drawdown that exceeds what the programme was designed to handle.
The trader who oversizes in a good period does not feel the error as it accumulates. The equity curve rises. The risk looks retrospectively justified. Confidence in the programme, and in the sizing specifically, increases.
Then the adverse regime arrives, the losing streak that is a structural feature of every trend following programme, and the oversized positions translate the normal drawdown into an abnormal one. The programme crosses into territory from which recovery is not practically available.
The correct response to this trap is not discipline in the moment of temptation. It is a sizing rule that does not offer the temptation in the first place.
A fixed, formula-driven sizing methodology removes the decision. There is no moment at which the trader chooses how large to go based on how confident they feel about the signal or how well the programme has been performing. The formula produces a number. The trader executes the number.
This is what it means for sizing to be part of the system rather than a judgment attached to it. The Outlier Hunter’s programme does not grow more aggressive when conditions feel favourable. It applies the same formula in every condition, because the formula was designed for all conditions, not for the ones that are currently visible.
What Comes Next
Position sizing answers the question of how much to risk on any single position. The next question is how many positions to hold, across how many markets, running on how many independent systems simultaneously.
The answer is not a matter of comfort or operational convenience. It is a structural property of Outlier Hunting. The Outlier Hunter’s edge is concentrated in rare, fat-tail events. The programme that is positioned in ten markets has ten chances to be present when the next outlier arrives. The programme positioned across a hundred has a hundred. Diversification, in this frame, is not risk reduction. It is opportunity amplification.
That argument is Foundation 3.
READ DEEPER
→ Position Sizing: The Geometry of Survival
→ Who Holds the Risk: Why Markets Survive and Traders Do Not
→ The Cut Back Rule: Engineering Survival in a Fractal World
→ Dynamic Position Sizing: Insights, Implications, and the Outlier Hunting Perspective
Previous: Foundation 1: What a System Actually Is | Next: Foundation 3: What Diversification Actually Does (and What It Cannot Do)
Want the theoretical foundation for why trend following works?
The Fractals of Finance: Determinism, Adaptation and the Geometry of Markets
The book explores the full architecture of feedback, fat tails, and fractal structure in financial markets, and what it means for how we trade, invest, and understand risk.
Available now on Amazon in paperback, hardcover, and Kindle.
Want a practical field manual for trading trends and capturing outliers?
The Aussie Turtles Trend Following Guide: A Field Manual for Hunting Outliers adapts the timeless principles of the original Turtle traders into a systematic, rules-based approach for modern markets. Co-authored with Adam Havryliv.
Available now on Amazon in paperback, hardcover, and Kindle.
Lorem ipsum dolor sit amet, consectetur adipiscing elit. Ut elit tellus, luctus nec ullamcorper mattis, pulvinar dapibus leo.Lorem ipsum dolor sit amet consectetur adipiscing elit dolor