Kelly Criterion Calculator

Result definition

This page calculates Full Kelly stake from the visible Bankroll and Estimated win probability entries. The output applies only to the scenario and market boundaries stated on the form.

Keep dutching outside this formula and use Dutching, and retain the current inputs for reproducibility.

Before replacing sample values

• Label Bankroll as observed, quoted, or projected. Its role is current betting bankroll.
• Under the entered assumptions, check the timestamp and unit for Estimated win probability because it supplies your probability estimate.
• Keep American odds on the event basis defined here: enter positive or negative American odds.

Formula mechanics

Growth and ruin describe repeated wagers rather than a guaranteed sequence.

Actual drawdowns depend on result order and dependence.

A separate arithmetic check

In the current scenario, this worked case verifies the method without describing a typical market.

• Bankroll: $910
• Estimated win probability: 57.75%
• American odds: -125

Applying the Kelly Criterion rule: Kelly fraction = (decimal odds × probability − 1) ÷ (decimal odds − 1).

Event information that still matters

On this page, use an executable price for the exact selection and stake, not an earlier screenshot or an unavailable best quote.

Within this calculation, cash, restricted credit, gross return, and net profit may require separate accounting.

Testing result sensitivity

Lower estimated win probability before accepting the stake or growth path.

Compare the baseline with a capped or fractional policy.

A longer losing-sequence case provides a cautious boundary.

How to read the result

Treat the stake or growth figure as a policy output, not a command.

A worse probability case shows how quickly it contracts.

Use Fractional Kelly for the narrower fractional kelly result, with assumptions saved separately from Full Kelly stake.

Settlement and data limitations

Kelly staking can be highly volatile when the probability estimate is wrong.

Within this calculation, an event update can stale the input set before the arithmetic changes.

What deserves a second calculation

Look for dependency between the visible entries. If Estimated win probability was already used to construct Bankroll, applying both independently may repeat the same information. Calculate the simpler version first, add the disputed adjustment in a second case, and compare the two values of Full Kelly stake.

A boundary check should cover realistic low and high values for Estimated win probability, not arbitrary extremes. Hold American odds steady during that test. If Full Kelly stake crosses the relevant line throughout a defensible range, the conclusion is sturdier than one supported by a single point estimate.

Questions about the inputs