The calculation, without a black box.
Divide the annual volatility target by estimated annual volatility. Apply the exposure cap, then multiply by capital. For an ordinary cash asset, dividing that notional value by its price gives fractional units.
Position value = capital × exposure
Modeled volatility = exposure × estimated volatility
Worked example: 100,000 capital, a 10% target and a 20% estimate imply 0.5× exposure: a position worth 50,000, or 500 units at a price of 100. With the same target and a 40% estimate, the position falls to 25,000. If the estimate falls to 5%, the uncapped rule asks for 2× exposure; a 1× cap limits the position to 100,000 and modeled volatility to 5%.
The simplified relationship assumes one linear exposure and constant-value cash. It omits financing, cash interest, transaction costs, taxes, market impact, rounding and changing prices. A fractional result may not be executable.
Which kind of position sizing do you mean?
| Method | Main input | What it answers |
|---|---|---|
| Volatility sizing | Variability of returns | How much exposure corresponds to a chosen volatility scale? |
| Stop-loss sizing | Entry-to-stop distance and a loss budget | How many units fit the intended loss at the stop? Execution can slip or gap. |
| Kelly sizing | Probabilities and payoffs | Which fraction maximizes expected logarithmic growth in the assumed model? |
This tool implements the first model. It does not calculate forex lots, stop orders or an optimal bet. Explore the separate Kelly game and calculator for the third.
A lower estimate can produce a larger position.
That is the mechanism, and also a failure mode. A calm recent period can make a historical estimate too low for what happens next. The cap prevents this illustration from demanding unlimited exposure, but it does not cap losses or guarantee that a target is achieved. The model refuses a zero or negative volatility estimate because division would be undefined or meaningless.
For several holdings, portfolio risk also depends on weights and correlations. Applying this rule independently to each holding does not establish the total portfolio’s volatility. See the volatility-targeting guide for a fuller explanation of estimation lag, leverage and rebalancing.
Method notes and sources
Educational model by Positive Skew · Method version 1 · 17 October 2026. Formulas and examples are reproducible from the inputs shown.
For a published implementation of capped target-volatility exposure, see FTSE Russell’s Custom Volatility Target Index ground rules, section 5. Its full methodology differs from this illustration. For financing and leveraged-loss risks, see the SEC investor bulletin on leveraged investing.