Defining risk before entering is the foundation of survival. Position size — how many shares you put on — is the most controllable variable a trader has, and it dictates how long an account survives a losing streak. Van Tharp argues position sizing is the dominant driver of long-run performance variability (more than entry or win rate). The hard truth that frames everything below: sizing cannot create an edge from a negative-expectancy system — it can only preserve a real edge or destroy it.

Stops

A stop converts open-ended market risk into a defined, known loss. Two approaches:

• Structural stops — placed just beyond price structure: below the recent swing low (long) or above the swing high (short), or below a support level. The logic: if price breaks that level, the trade thesis is wrong.
• ATR-based / volatility stops — distance scaled to volatility via ATR (defined in Technical Analysis → Volatility Indicators). The Chandelier Exit (Chuck LeBeau) is the standard trailing form:

- Long: highest high (22) − ATR(22) × 3 - Short: lowest low (22) + ATR(22) × 3 The 3× multiplier is the common default; tighten toward 1.5–2.5× in high-volatility names/regimes, loosen toward 3–3.5× for stable large-caps. The stop ratchets up with new highs, trailing the move while leaving room for normal noise.

A stop too tight gets shaken out by noise; too wide bleeds more than necessary. ATR sizing is the objective middle ground.

Position sizing

The 1–2% rule: risk no more than 1–2% of account equity on any single trade (2% is the common swing default). This guarantees a losing streak can't be fatal.

The core formula:

Shares = (Account Equity × Risk %) / (Entry − Stop)   ← per-share risk

Portfolio heat = the sum of risk across all open positions if every stop hits at once. Cap it (commonly ~5–6%). Crucially, correlated positions count together — two 2% tech names aren't 4% diversified risk, they're closer to a single ~4% tech bet. Capital preservation outranks any one opportunity (a 20% drawdown needs a 25% gain to recover; the math gets punishing fast).

R-multiples & expectancy

R = the initial risk per trade (entry − stop). All outcomes are measured in R: exit at +2× your risk = +2R. This normalizes results across position sizes and price levels.

Expectancy = the average P&L per trade over many trades:

Expectancy (R) = (Win% × Avg-win R) − (Loss% × Avg-loss R)

Positive expectancy = a real edge; negative = the system loses long-term no matter how disciplined the execution.

The 2:1 reward:risk filter: reject setups whose (target − entry) / (entry − stop) is below ~2:1. At 2:1 you break even winning just 1 in 3; at 3:1, 1 in 4. R:R is often more decisive than win rate — 40% wins at 3:1 (+0.6R) beats 60% wins at 1:1 (+0.2R).

Worked example

$25,000 account, 2% rule → max risk $500/trade. Entry $100, stop $98 → per-share risk $2. Shares = $500 / $2 = 250. If stopped, the loss is exactly 250 × $2 = $500 (2% of account). As equity grows to $50,000, the same rule auto-scales risk to $1,000 / 500 shares.

Expectancy check — 100 trades, 52% win rate, avg win +$150, avg loss −$100: (0.52 × 150) − (0.48 × 100) = 78 − 48 = +$30/trade → +$3,000 over 100 trades. In R (avg win ≈ 1.5R, avg loss = 1R): (0.52 × 1.5) − (0.48 × 1) = +0.30R/trade.

Strengths, limitations & evidence

Position sizing and risk definition are the best-grounded part of trading — the math (drawdown recovery, expectancy, ruin probability) is deterministic and survives every market regime, which is not true of setup "edges." This is why it's the highest-confidence layer in this whole corpus.

But the limitation is absolute: sizing is necessary, not sufficient. A −0.1R system fails slowly at 1% risk and fast at 5% — it still fails. The edge must exist first.

Common misuses: moving a stop to dodge a loss (destroys the defined risk); oversizing on a hot streak (emotional, not formula-based); ignoring portfolio heat / correlation; and confusing win rate with profitability (65% wins at 0.8:1 is −0.17R — a losing system).

Sources

• Van Tharp, Trade Your Way to Financial Freedom — position sizing, R-multiples, expectancy (the "% of performance variability" claim is Tharp's framing, widely cited and debated)
• StockCharts ChartSchool — Chandelier Exit (LeBeau); Wilder — ATR
• Investopedia / professional risk-management literature — 1–2% rule, position-size formula, reward:risk math, portfolio heat
• Note: specific percentages in popular write-ups vary; the formulas above are standard and the worked-example arithmetic is verified.