Why position sizing outranks stock picking
Here is a counterintuitive truth: a trader with a 40% win rate and excellent position sizing will outperform a trader with a 60% win rate and poor position sizing over 100 trades. The reason is asymmetry. Losses compound devastatingly; gains cannot recover from them fast enough when sizes are wrong.
A trader who puts 20% of their account into a single trade and hits a stop loss has lost 20% of their capital in one moment. That requires a 25% gain to break even — on the next trade. Repeat this twice and they need to triple their account to get back to the starting point. Position sizing determines whether you survive long enough to let your edge play out.
The 3-variable formula
Every position size calculation flows from a single formula:
Stop Distance = Entry Price − Stop Loss Price (in dollars per share)
With a $10,000 account, 1% risk rule, and a $1 stop distance, you can buy 100 shares ($10,000 × 0.01 = $100 risk; $100 ÷ $1 = 100 shares). If the stop is $2 away, you buy 50 shares. If it's $5 away, you buy 20 shares. The dollar risk stays constant — the share count changes with the stop.
Position Size Formula — Two Examples
Doubling the stop distance halves your position size — your dollar risk stays fixed at $200 (1% of account) in both cases.
Worked examples across different stops
| Account | Risk % | Max $ Risk | Stop Distance | Shares to Buy |
|---|---|---|---|---|
| $10,000 | 1% | $100 | $1.00 | 100 |
| $10,000 | 1% | $100 | $2.00 | 50 |
| $10,000 | 1% | $100 | $5.00 | 20 |
| $10,000 | 2% | $200 | $2.00 | 100 |
| $25,000 | 1% | $250 | $2.50 | 100 |
Why 1% vs. 2% matters over 100 trades
The difference between risking 1% and 2% per trade seems trivial — but compound loss math makes it significant over any extended losing streak.
| Consecutive Losses | Account left (1% risk) | Account left (2% risk) | Account left (5% risk) |
|---|---|---|---|
| 10 losses | 90.4% | 81.7% | 59.9% |
| 20 losses | 81.8% | 66.8% | 35.8% |
| 30 losses | 74.0% | 54.5% | 21.5% |
After 30 consecutive losses at 5% risk, a trader has lost 78.5% of their account. They now need a 460% gain to return to breakeven — mathematically near-impossible without taking the same reckless risks that caused the drawdown. At 1%, 30 straight losses leave 74% of capital intact and recovery is a normal winning streak away.
Thinking in R — normalising trade outcomes
Once you define your risk per trade as 1R (one unit of risk), every trade outcome can be expressed as an R-multiple. A trade that made $300 when you risked $100 = +3R. A trade that lost $50 = -0.5R. A stopped-out full loss = -1R.
This system lets you compare trades across different account sizes and evaluate the quality of your trading decisions independent of dollar amounts. A system with average wins of +2R and average losses of -1R has positive expectancy as long as your win rate exceeds 33%.
A trader has a $20,000 account and uses the 1% rule. Their stop loss is $2 away from entry. How many shares can they buy?
Kelly Criterion basics
The Kelly Criterion is a mathematical formula for optimal bet sizing developed by John Kelly at Bell Labs in 1956. It calculates the percentage of capital to risk on each trade to maximise long-run geometric growth:
Example: If your strategy wins 55% of the time and your average win is 2× your average loss (2:1 R:R), then Kelly = (2 × 0.55 − 0.45) ÷ 2 = (1.10 − 0.45) ÷ 2 = 0.325, or about 32.5% of your bankroll per trade.
In practice, full Kelly (32.5%) is far too aggressive for trading. Variance in live markets is significantly higher than in controlled probability environments. Most professional traders use Half Kelly (16.25% in this example) or Quarter Kelly (~8%) — and even that is high. The 1-2% fixed risk rule is effectively very conservative Kelly, which is the right approach for most traders.
Volatility-adjusted sizing using ATR
A fixed stop distance (e.g. always $2/share) works only if all stocks have similar volatility — which they don't. A stock with an Average True Range (ATR) of $15 will frequently breach a $2 stop just from normal daily noise. The solution is to scale your stop to the stock's volatility, then calculate shares accordingly.
| Stock | ATR | Stop (ATR × 1.5) | Risk ($200) | Shares |
|---|---|---|---|---|
| Low-vol stock | $1.50 | $2.25 | $200 | 88 shares |
| Mid-vol stock | $5.00 | $7.50 | $200 | 26 shares |
| High-vol stock | $15.00 | $22.50 | $200 | 8 shares |
Notice how the high-volatility stock receives only 8 shares vs. 88 for the low-volatility stock — but the dollar risk is identical at $200. This is the core principle: you normalise risk, not share count.
The correlation warning
Holding five positions at 1% risk each sounds like proper diversification — but only if those positions are genuinely uncorrelated. Five technology stocks (AAPL, MSFT, NVDA, META, GOOGL) all falling on a Fed rate announcement will all hit their stop losses simultaneously. Five 1% positions in the same sector can behave like a single 5% position during market stress.
Four position sizing red flags
Case study: Nick Leeson and Barings Bank (1995)
Barings Bank had been in operation for 230 years when it collapsed in February 1995. The cause was a single trader — Nick Leeson — operating in the Singapore futures market with no position sizing system and no risk limits enforced on his activities.
Leeson began accumulating a massive long position in Japanese Nikkei 225 futures, betting that the Japanese market would recover. When the Kobe earthquake hit on January 17, 1995, the Nikkei fell sharply — and Leeson's losses exploded. Rather than cutting the position, he doubled down. He ultimately controlled positions worth approximately $27 billion in notional value, on behalf of a bank with capital of roughly $600 million.
The final loss was £827 million ($1.4 billion at the time) — more than twice the bank's available capital. Barings was sold to ING for £1 and Leeson was sentenced to six and a half years in prison.
Expectancy — the formula that ties it all together
Position sizing and risk management ultimately serve a single goal: letting your system's expectancy play out. Expectancy is the average profit or loss per dollar risked over many trades:
Example: 45% win rate, avg win = 2R, avg loss = 1R:
E = (0.45 × 2R) − (0.55 × 1R) = 0.90R − 0.55R = +0.35R per trade
A positive expectancy system is only valuable if you apply it consistently across enough trades. Inconsistent position sizing destroys expectancy even in a winning system.
Key terms — full table
| Term | Definition |
|---|---|
| Position Size | Number of shares/contracts — calculated, not guessed |
| Risk Per Trade | Max dollar loss if stop is hit. Keep to 1-2% of account |
| R-Multiple | Trade result in units of risk. 3R = profit of 3× initial risk |
| Kelly Criterion | Formula for optimal bet size. Use 1/4 Kelly max in trading |
| ATR | Average True Range — measures daily price volatility in dollars |
| Volatility Sizing | Stop = ATR × 1.5–2; Shares = Risk $ ÷ Stop Distance |
| Expectancy | Avg profit per trade: (WR × Avg Win) − (LR × Avg Loss) |
| Drawdown | Peak-to-trough account decline. 50% drawdown needs 100% gain to recover |
Kelly Criterion suggests betting 25% of your bankroll on each trade. Should a trader follow this literally?