← Back to GugoPro AcademyRisk management · Personal input · Not financial advice
Quant lab · Risk budget

Kelly risk budget

Using the statistical summary of your completed trades, convert 'edge' into an upper bound and a conservative fractional‑Kelly. If win rate, payoff ratio or sample are unstable, the output should not be used as an automatic sizing rule.

Enter a summary based on your closed trades; this page does not save records.
Formula:b＝平均獲利 R ÷ 平均虧損 R；Full Kelly＝p−(1−p)÷b；分數 Kelly＝Full Kelly×fraction。Position size is then estimated by account × risk fraction and stop distance; gaps, slippage, correlations and margin rules are not included。
Risk budget

Theoretical upper bound and conservative ratio

—
Awaiting calculation
Full Kelly—
Profit/loss ratio b—
Risk budget—
Stop position cap—

Full Kelly and Fractional Kelly

FullFractional
Important constraints:Kelly is highly sensitive to estimation errors in win rate and win/loss ratio; if Full Kelly is less than 0, it means the historical edge is insufficient to support a positive Kelly — do not invert a negative value to obtain leverage. This page is a research calculator, not personalized investment or trading advice.

How to use this finance tool

This page turns a finance concept into checkable inputs, formulas, and scenarios. Read the variable definitions first, then compare conservative, base, and stress cases; one result is not a promise of return.

Suggested workflow: confirm units and time periods, enter your own assumptions, then review sensitivity, costs, and downside cases.

Limitations: the model does not forecast markets and may not include every tax, slippage, liquidity, credit, regulatory, or contract term. Verify important decisions with current primary information and a qualified professional.

Financial risk disclaimer

This page, its calculators, and examples are for education, research, and scenario estimation only. They are not personalized investment, trading, betting, tax, legal, or financial advice. Markets and local rules can change quickly; verify current primary information and take responsibility for your decisions. Past performance, model outputs, and simulations do not guarantee future results.