What Is Position Sizing?

Same entry, opposite outcomes

Two traders buy the same stock at $80 on the same morning with the same $75 stop. The stock gaps down through the stop a week later. Trader A had put 100 shares against a $50,000 account and loses $500 — one percent, an ordinary Tuesday. Trader B had put $30,000 of the same account into 375 shares and loses about $1,900, nearly 4%, and after three more trades like it is down enough that the strategy never gets a fair test.

Identical entry, identical stop, identical market. The only variable was size, and size decided everything: who survived the losing stretch, who could think clearly during it, and whose account was still intact when the strategy's winners finally showed up. Position sizing is the discipline of choosing that variable on purpose, before entry, with arithmetic instead of conviction.

Fixed-fractional, worked start to finish

The fixed-fractional method fixes the fraction of your account you are willing to lose on any single trade, then lets the trade's own geometry set the share count. Three inputs: account size, risk percent, and the distance from entry to stop. The formula: shares = (account x risk %) / (entry price - stop price).

Worked all the way through: account $50,000, risk 1% per trade, so the dollar risk budget is $50,000 x 0.01 = $500. The setup is an entry at $80 with a stop at $75, so the per-share risk is $80 - $75 = $5. Shares = $500 / $5 = 100 shares. Position cost: 100 x $80 = $8,000, which is 16% of the account. If the stop is hit, the loss is 100 shares x $5 = $500. Check it against the budget: $500 / $50,000 = 1.0%. The circle closes exactly.

Notice what the method did not ask: how confident you feel, how good the chart looks, how much you like the company. It asked one question — where is this trade proven wrong? — and converted the answer into a share count. The stop placement is the analysis; the size is just the analysis expressed in dollars you can afford to be wrong with.

The stop sets the size, and that cuts both ways

Tighten the stop and the formula buys more shares for the same dollar risk. Same account, same $500 budget, same $80 entry, but a stop at $78: per-share risk is $2, so shares = $500 / $2 = 250, and the position costs 250 x $80 = $20,000 — 40% of the account. The dollar risk at the stop is still $500. The formula is internally consistent, and it has just told you to put 40% of your money in one name.

That is where a second constraint has to enter from outside the formula: a notional cap. Many professionals cap any single position at 20-25% of the account no matter what the stop math allows, for two reasons. First, tight stops are noisy — a $2 stop on an $80 stock is 2.5%, inside the ordinary daily wiggle of many names, so the 250-share version gets stopped out by randomness far more often than the 100-share version. Second, stops are promises the market never signed.

The unsigned-promise problem is gap risk. A stop at $75 assumes someone will buy your shares at $75 on the way down. Hold through an earnings report and the stock can open at $68 without ever trading at your stop: the 100-share trader loses (80 - 68) x 100 = $1,200 — 2.4% instead of the budgeted 1%. The 250-share trader with the $78 stop loses (80 - 68) x 250 = $3,000, six times the budget. Fixed-fractional sizing bounds the ordinary loss; only smaller positions and avoiding binary events bound the extraordinary one.

Why size, not entries, is the professional edge

Losing streaks are not a possibility; they are a schedule. A strategy that wins 50% of the time will produce a streak of five straight losses about once every 60 or so trades, and most active traders take hundreds of trades a year. The design question is never whether the streak comes — it is what the streak does to the account when it arrives.

Run the compounding: ten consecutive 1% losses leaves 0.99^10 = 90.4% of the account, a 9.6% drawdown, and getting back to even requires a 10.6% gain — annoying, recoverable. Ten consecutive 5% losses leaves 0.95^10 = 59.9% of the account, a 40.1% drawdown, and recovery now requires a 67% gain. Same strategy, same streak, same entries. The 1% trader is bruised; the 5% trader needs a career year just to reach zero.

This asymmetry — losses require disproportionately large gains to reverse — is why experienced traders talk about sizing more than setups. Entry signals are probabilistic and partly luck; the mapping from a losing streak to account damage is pure arithmetic, chosen in advance, and it is the one part of trading where you control the outcome completely. The Kelly criterion formalizes the same idea from the other direction, deriving the growth-optimal fraction from win rate and payoff ratio; in practice most who use it size at half-Kelly or less, because the inputs are estimates and the penalty for oversizing is the 40% drawdown above.

Correlation stacking: three bets that are one bet

The per-trade budget has a blind spot: it counts positions, not exposures. Suppose the $50,000 account holds three semiconductor names, each sized to risk exactly 1% — $500 at its stop. On paper that is three independent, well-behaved bets. Then a sector-wide selloff hits: chip stocks trade together, all three positions fall through their stops in the same week, and the account takes 3 x $500 = $1,500, a 3% hit from what was effectively a single decision — being long semiconductors.

The stops did not fail; the independence assumption did. Correlated positions share a risk factor, and when the factor moves, it moves all of them. Book five correlated 1% positions and your true single-bet exposure is 5%, which puts you on the ugly side of the streak math from the previous section without ever violating the per-trade rule.

The fix is a second budget layered on the first: a cap on total risk per theme. A common structure is 1% per trade, but no more than 2-3% of combined stop-distance risk in any one sector or driver (rates, oil, one country, one earnings story). Counting exposure by factor rather than by ticker is most of what separates a sized portfolio from a sized trade.

The deadly habit: sizing up after losses

The most dangerous sizing behavior has a familiar shape. Three straight $500 losses take the account to roughly $48,500, and the next idea looks clean, and a thought arrives: risk $1,500 on this one and the ledger is even again. It feels like decisiveness. It is the martingale — doubling into losses — wearing a trader's clothes, and it inverts the entire logic of fixed-fractional sizing at the exact moment the logic matters most.

Two things are reliably true right after a losing streak. Your capital is smaller, so any fixed dollar loss is a larger percentage of what remains. And your judgment is at its worst — loss-chasing is one of the best-documented behavioral failures in trading, and the urge to make it back is precisely the signal that you are inside it. Sizing up at that moment increases risk on degraded capital with degraded judgment, three compounding errors in one order ticket.

Fixed-fractional sizing already contains the correct response, and it happens automatically: because the risk budget is a percent of current equity, a $48,500 account risking 1% puts $485 on the next trade, not $500 and not $1,500. Size shrinks as the account shrinks and rebuilds as it recovers. The discipline is not adding a rule — it is declining to override the one already running.

What sizing cannot do

Position sizing does not create an edge. Size a coin-flip strategy with flawless 1% discipline and you still bleed to zero, just slowly and politely — the expectancy of the underlying strategy is untouched by how you portion it. Sizing determines whether a genuine edge survives long enough to pay, and how much damage a bad strategy does before you notice. It is a survival technology, not a profit one.

The formula also inherits every weakness of its inputs. The stop is a judgment call dressed as a number; place it inside the stock's normal noise and the formula confidently sizes a position that gets stopped out on randomness. Gap risk means the budgeted loss is a floor estimate, not a ceiling — position sizing bounds the loss you planned for, not the overnight surprise. And the risk percent itself is a choice with no formula behind it: 1% is a convention that lets a bad month stay boring, not a law. What the method actually guarantees is narrower and still worth having — no single ordinary trade, and no ordinary streak of them, gets to end the account.