Risk
Position Sizing and the Kelly Criterion: How Much to Risk Per Trade
Most investors obsess over entries. The ones who compound wealth obsess over sizing. Kelly explains why - and what most traders get wrong about it.
SIGMASEVEN ResearchIn short
Position sizing determines how much capital to allocate per signal. The Kelly Criterion is the mathematical framework that maximises long-run growth given a known edge and win rate. In practice, fractional Kelly - using 25 to 50 percent of the full Kelly fraction - is the standard for risk-adjusted growth without catastrophic drawdowns.
Why sizing matters more than entry
Two traders can use the same strategy and produce radically different outcomes if they size positions differently. Oversizing converts a positive expected value process into ruin. Undersizing converts it into mediocrity. Correct sizing is the difference between compounding and surviving.
This is not a secondary consideration. It is the primary lever private investors have over their long-run outcome.
What the Kelly Criterion calculates
Kelly gives the fraction of capital to bet on each opportunity to maximise the geometric growth rate of a portfolio. The formula is: Kelly fraction equals edge divided by odds. Edge is expected profit per unit risked. Odds is the ratio of win size to loss size.
A strategy with a 55 percent win rate and 1:1 payoff has a Kelly fraction of 10 percent. That is the maximum allocation per signal for optimal long-run compounding.
Why professionals use fractional Kelly
Full Kelly maximises long-run growth but produces drawdowns that almost no human will sustain behaviourally. At full Kelly, the strategy will regularly halve the portfolio before recovering. Half Kelly produces roughly 75 percent of the growth with dramatically smaller drawdowns.
Quarter Kelly is common in institutional systematic strategies - it captures meaningful compounding while keeping worst-case drawdowns within the range investors will actually hold through.
Applying Kelly to a systematic Bitcoin strategy
A signal-based Bitcoin strategy with a documented win rate and average win/loss ratio can calculate its theoretical Kelly fraction from historical data. The key discipline is using out-of-sample win rates, not the best period of a backtest.
In practice, systematic strategies often run at 10 to 20 percent of the Kelly-optimal allocation per signal, especially when the underlying asset has high volatility. This is not timidity - it is the mathematically correct response to uncertainty about the true edge. Our quantitative systems use fractional Kelly sizing calibrated from out-of-sample signal history.
Frequently asked questions
- What is the Kelly Criterion?
- The Kelly Criterion is a mathematical formula that calculates the optimal fraction of capital to allocate per bet to maximise the long-run geometric growth rate of a portfolio, given a known edge and win rate.
- How do I calculate my Kelly fraction?
- Kelly fraction equals (win probability × average win) minus (loss probability × average loss), divided by the average win. You need a documented set of historical trades to calculate this accurately.
- What is fractional Kelly?
- Fractional Kelly means using a percentage - typically 25 to 50 percent - of the full Kelly-optimal allocation. It sacrifices some long-run growth in exchange for significantly smaller drawdowns.
- Does Kelly Criterion work for cryptocurrency trading?
- Yes, but with smaller fractions than traditional assets due to higher volatility. Crypto strategies typically run at 10 to 25 percent of full Kelly to keep drawdowns within manageable ranges.