position-sizing

Position Sizing

Purpose

Determine how much to risk on a single position, so that a losing streak is survivable and no single idea can end the account. Position sizing determines survival; entry selection determines returns. Survival comes first.

When to Use

• Deciding how many shares or contracts to buy.
• Setting a risk-per-trade policy.
• Checking whether a position breaches concentration limits.
• Reviewing why a drawdown was deeper than expected.

Capabilities

• Fixed-fractional sizing from a stop distance.
• Volatility-adjusted sizing using ATR.
• Kelly criterion, and fractional Kelly in practice.
• Concentration limits: per position, per sector, per correlated group.
• Aggregate portfolio risk.

Inputs

• Account equity.
• Entry price and stop price (or an ATR and a multiplier).
• Risk per trade, as a percentage of equity.
• Existing exposure, including correlated positions.

Outputs

• A share or contract count.
• The dollar risk if the stop is hit, stated explicitly.
• Whether the position breaches any limit.

Workflow

1. Fix the risk per trade first — Typically 0.5% to 2% of equity. This is a policy decision made when calm, not a per-trade judgment made when excited.
2. Determine the stop — At a level where the thesis is wrong, not at a level that produces a comfortable size. This is the discipline that everything else depends on.
3. Size from the stop distance — Shares = (equity × risk%) / (entry − stop). A wider stop means a smaller position, not more risk.
4. Check the concentration limits — Position size as a percentage of equity, sector exposure, and correlated exposure. A limit breached is a size reduced.
5. Check the aggregate — Ten positions each risking 1% is 10% at risk if they are correlated and all stop out together. In a market-wide selloff, they will be.

Best Practices

• Never widen a stop to justify a larger position. That inverts the entire logic: the stop defines the risk, and the size follows from it.
• Correlated positions are one position. Five semiconductor names is one bet on semiconductors, and it should be sized as one.
• Full Kelly is theoretically optimal and practically unusable: it produces enormous drawdowns, and it assumes you know your edge precisely. You do not. Half-Kelly or less is the practical range.
• A 50% drawdown requires a 100% gain to recover. Sizing that permits deep drawdowns is not aggressive; it is arithmetically self-defeating.
• Volatility-based sizing normalizes risk across instruments: a fixed dollar risk on a volatile stock and a quiet one produces very different position values, and that is correct.
• Recalculate the size on the current equity, not on the starting equity. Sizing off a high-water mark after a drawdown compounds the damage.