Forex glossary
Kelly Criterion
The Kelly criterion selects the capital fraction maximizing expected log wealth in a specified repeated-bet model. For binary outcomes, f = p − (1 − p)/b, where b is profit per unit staked. Estimation error can cause excessive risk; the formula does not guarantee account growth.
Illustrative example
How does it work in practice?
For hypothetical independent bets with p = 0.6, a win gains the stake and a loss loses the stake. Kelly gives 0.6 − 0.4 = 20% of capital; this mathematical model is not a position-size recommendation.
A case to work through
One changed assumption
Binary model: win and loss equal the stake, b = 1; compare assumed p = 0.55 with p = 0.50.
- At p = 0.55, the formula gives 0.55 − 0.45 = 0.10.
- At p = 0.50, it gives 0.50 − 0.50 = 0.
What follows? Assess input uncertainty before interpreting the fraction.
Example boundary: These are independent hypothetical bets without costs, not a risk recommendation.
A useful distinction
Kelly Criterion / 1% Rule
The 1% rule sets a budget; Kelly optimizes a fraction in a specified distribution model.
Check your understanding
Calculate a binary formula’s sensitivity to probability.
Show the answer and explanation
0%
The binary model has no edge at equal chances.
Net results and risk measures
Step 10 / 10 · 0 / 10 reviewed
View the full sequence →Progress is saved only in this browser.
Read the fuller explanation
Related terms
Sources and documentation
- Paper on Kelly models and volatility
Mathematical capital-fraction assumptions rather than a risk recommendation or growth guarantee.
The example illustrates the mechanism. Check current account conditions and instrument specifications with the provider.