Volatility-Based Position Sizing
Learn how volatility-based position sizing uses ATR or return volatility to scale exposure, normalise risk across assets and avoid using equal notionals for unequal price behaviour.
Reading progress — saved on this device
Volatility-based sizing reduces exposure when an asset is moving more violently and increases it when volatility is lower, aiming to make risk contributions more comparable across positions.
Learning objectives
- Translate a volatility estimate into a position-size rule.
- Explain why equal notional positions can contribute very different risk.
- Identify estimation, regime and leverage risks in volatility targeting.
What it is
A £10,000 position in an asset that typically moves 1% per day is not equivalent to a £10,000 position in an asset that moves 8% per day. Volatility-based sizing adjusts notional exposure so expected price variability contributes a chosen amount of portfolio risk.
Common inputs include Average True Range (ATR), standard deviation of returns, exponentially weighted volatility or realised volatility over a specified horizon. The estimator must match the intended holding period and strategy.
Volatility sizing can be used alone for portfolio allocation or combined with a structural stop. In the latter case, volatility helps decide whether the planned stop is unusually tight or wide relative to current market noise.
How it works
A simple volatility-target rule sets position notional approximately equal to risk budget divided by volatility. If two assets have volatilities of 2% and 8%, the higher-volatility asset receives roughly one quarter of the notional for an equal volatility contribution.
ATR sizing works in price units. If 14-day ATR is £2,000 and the risk rule allows a 2×ATR stop, loss per BTC is £4,000. A £400 risk budget therefore supports 0.10 BTC before slippage adjustments.
Volatility estimates can be procyclical: low measured volatility encourages larger positions just before volatility expands, while a spike forces deleveraging after losses. Using minimum volatility assumptions and gradual resizing can reduce this instability.
Crypto volatility differs across time zones, weekends, assets and venues. The chosen measurement window should be robust to 24/7 trading and should not mix incompatible return frequencies.
How to analyse and apply it
| Check | Why it matters | What to verify |
|---|---|---|
| Volatility estimator | Determines the scaling signal. | Specify horizon, sampling frequency and whether weighting is used. |
| Volatility floor | Prevents oversized positions in unusually calm periods. | Set a minimum stress volatility rather than trusting near-zero readings. |
| Target risk | Controls total exposure. | Express in portfolio or trade-risk terms, not just leverage. |
| Rebalance rule | Avoids constant turnover. | Define thresholds or scheduled resizing rather than reacting to every small change. |
Risk rules should be written before a live position is opened and evaluated across many trades or scenarios. A control that is changed only after losses appear is discretionary damage control, not a repeatable risk system.
Worked example and thought exercise
A strategy allocates £1,000 of volatility budget between two assets. Asset A has 20% annualised volatility; Asset B has 80%. Ignoring correlations and scaling constants, equal-risk sizing gives B about one quarter of A's notional.
For an ATR example, ETH is £2,500, ATR is £125 and the stop is 2 ATR away. Loss per ETH is £250. With £500 risk, the raw size is 2 ETH. If liquidity stress suggests £40 additional adverse execution per ETH, the risk-aware size should be lower.
Thought exercise: why might mechanically increasing leverage when realised volatility falls from 50% to 15% be dangerous?
Common mistakes and practical workflow
- Using equal notionals and assuming that means equal risk.
- Treating a short historical-volatility window as a stable forecast.
- Increasing leverage without a volatility floor or maximum notional cap.
- Ignoring correlation when several volatile assets respond to the same crypto beta factor.
Practical workflow
- Choose a volatility estimator appropriate to the holding period.
- Apply a volatility floor and stress estimate.
- Translate the target risk into notional or units.
- Check correlation, liquidity and leverage caps.
- Resize only under a predefined rebalance rule.
✅ Knowledge checkpoint
- Why are equal-notional positions not necessarily equal-risk?
- How does ATR translate into loss per unit?
- Why can volatility targeting become procyclical?
- What purpose does a volatility floor serve?
FAQs
❓ Is ATR the same as standard deviation?
No. ATR measures trading range including gaps, while standard deviation measures dispersion of returns around their mean.
❓ Does lower volatility always justify more leverage?
No. Low realised volatility can precede abrupt regime changes; leverage still needs independent caps and stress tests.
❓ Can volatility sizing remove correlation risk?
No. It normalises standalone variability, but correlated positions can still move together.
❓ How often should volatility positions be resized?
That depends on strategy horizon and transaction costs. Threshold or scheduled rebalancing is often more stable than continuous resizing.
📋 Summary
Volatility-based sizing recognises that a pound of notional is not a pound of risk. It improves comparability across assets, but only when volatility estimates are paired with floors, leverage caps, liquidity checks and portfolio-level correlation controls.
Want this in a personalised order?
Take the crypto assessment and get a custom path of 10 modules matched to what you already know. Free, no card required.
Build my path →