Skip to main content
Calnex LogoCalnex
Finance

Kelly Criterion Calculator

Work out the Kelly criterion bet size as a percentage of a bankroll from a win probability and a profit to loss ratio, with an optional fractional multiplier to scale the result down.

Optimal Bet Size10.00 %View results
%

Real-time Results

Optimal Bet Size
10.00%
Expected Value (EV)
10.00$

Risk Analysis

ItemValue
Full Kelly %

Mathematically optimal, but extremely volatile

10.00%
Half-Kelly (0.5) %

Highly recommended for most investors (less variance)

5.00%
Recommendation

Favorable bet. Follow the calculated percentage of your bankroll.

Bet

Disclaimer

The calculation results of this financial tool are for estimation and research reference only and do not constitute any investment advice or commitment. Actual rates and yields are subject to the final contract with the financial institution.

How it works

This calculator applies the Kelly criterion, a mathematical formula for sizing a repeated bet as a fraction of a bankroll, using only two assumptions you supply: the probability of winning and the profit to loss ratio. It reports the resulting bet size as a percentage, an expected value figure, and an analysis list with the full Kelly percentage, a half-Kelly reference figure, and a simple Bet or Skip status. The optional fractional multiplier scales the headline bet size down (or up) by the factor you choose. Every output is a direct consequence of the numbers you type in: if the win probability is overstated, the size is overstated. The model assumes the same bet can be repeated many times with a known, fixed probability, which real bets and trades do not offer. This is an explanation of a formula and its assumptions, not investment or gambling advice, and no output here promises or predicts a profit.

How the Kelly size and expected value are calculated

The formula balances two forces. Winning more often, or winning more per unit risked, increases the suggested fraction; losing more often decreases it. Concretely, with win probability as a fraction p, loss probability q = 1 - p, and profit to loss ratio b, the expected value per unit risked is p times b minus q. Only when that expected value is positive does the formula produce a size: the full Kelly fraction is p minus q divided by b, shown as a percentage.

The headline bet size multiplies that full Kelly percentage by the fractional multiplier you enter, which defaults to 1. The expected value card rescales the per-unit expected value to 100 units wagered, so an expected value of 0.10 per unit appears as 10.00. The status in the analysis list is Bet when the expected value is positive and Skip when it is zero or negative, and in the zero or negative case every percentage is set to 0.

Reading the analysis list and the multiplier correctly

The Risk Analysis list needs careful reading, because its lines do not all respond to your multiplier. The full Kelly line always shows the unscaled Kelly percentage, and the half-Kelly line always shows exactly half of that, whatever multiplier you typed in. Only the main bet size card applies your multiplier. In the fourth example the multiplier is 0.25, so the bet size is 12.50 percent, a quarter of full Kelly, while the list still displays a half-Kelly figure of 25.00 percent as a reference.

The multiplier field is also unchecked: validation only requires the win probability to be strictly between 0 and 100 and the ratio to be above 0. If the multiplier is missing, zero, negative, or not a number, the calculation silently falls back to 1, the full size. There is no upper limit on the multiplier or the ratio in validation, so a multiplier above 1 will produce a bet size larger than full Kelly, and the formula can suggest very large fractions when the assumed edge is large, as the fifth example shows with 33.33 percent.

Model assumptions and why the result is not advice

The Kelly criterion is a long-run growth model, not a safety rule. It assumes the win probability is known exactly, that identical bets repeat independently many times, that the payout ratio never changes, and that fractional stakes of any size are possible. If the true probability is lower than the one entered, the suggested size is too large, and full Kelly staking is known within the model itself to produce severe swings in the bankroll even when the edge is real.

For those reasons the outputs must not be read as advice. A positive expected value is an arithmetic consequence of your inputs, not evidence that a real bet or trade has an edge, and no percentage here is a recommended stake. Nothing in this calculator verifies your probability, accounts for costs, taxes, or changing conditions, or promises any return. Use it to understand how the formula behaves and how sensitive it is to its assumptions, and treat any real decision as your own.

Examples and interpretation(5)

How do I size a bet with a 55 percent win probability and a profit to loss ratio of 1?

Fractional Kelly Multiplier
1
Profit/Loss Ratio (Odds)
1
Win Probability (%)
55 %
  1. Set Win Probability to 55, Profit/Loss Ratio to 1, and Fractional Kelly Multiplier to 1.
  2. Expected value per unit = 0.55 x 1 - 0.45 = 0.10, shown as 10.00 per 100 units.
  3. Full Kelly = 0.55 - 0.45 / 1 = 10.00 percent, and the multiplier of 1 leaves the bet size at 10.00.
Optimal Bet Size
10.00
Expected Value (EV)
10.00
Risk Analysis: Recommendation
Bet

The model suggests 10.00 percent of the bankroll. This rests entirely on the 55 percent win probability being accurate; if the true probability were 50 percent at the same ratio, the expected value would be zero and the model size would be 0.

How do I size a bet with a 60 percent win probability, a ratio of 2, and a half multiplier?

Fractional Kelly Multiplier
0.5
Profit/Loss Ratio (Odds)
2
Win Probability (%)
60 %
  1. Set Win Probability to 60, Profit/Loss Ratio to 2, and Fractional Kelly Multiplier to 0.5.
  2. Expected value per unit = 0.60 x 2 - 0.40 = 0.80, shown as 80.00 per 100 units.
  3. Full Kelly = 0.60 - 0.40 / 2 = 40.00 percent. Bet size = 40.00 x 0.5 = 20.00 percent.
Optimal Bet Size
20.00
Risk Analysis: Full Kelly %
40.00%

Halving the full Kelly through the multiplier gives a bet size of 20.00 percent, which in this case happens to match the separate half-Kelly line in the analysis, because the multiplier chosen was exactly 0.5.

What size does the model give for a 40 percent win probability and a ratio of 1?

Fractional Kelly Multiplier
1
Profit/Loss Ratio (Odds)
1
Win Probability (%)
40 %
  1. Set Win Probability to 40, Profit/Loss Ratio to 1, and Fractional Kelly Multiplier to 1.
  2. Expected value per unit = 0.40 x 1 - 0.60 = -0.20, shown as -20.00 per 100 units.
  3. Because the expected value is not positive, the Kelly percentage is set to 0 and the bet size is 0.00.
Optimal Bet Size
0.00
Expected Value (EV)
-20.00
Risk Analysis: Recommendation
Skip

Within this model the answer is to stake nothing: a negative expected value means the assumed odds work against you over many repetitions. The calculator does not suggest a smaller stake instead; it sets the size to zero.

How do I size a bet with a 70 percent win probability, a ratio of 1.5, and a quarter multiplier?

Fractional Kelly Multiplier
0.25
Profit/Loss Ratio (Odds)
1.5
Win Probability (%)
70 %
  1. Set Win Probability to 70, Profit/Loss Ratio to 1.5, and Fractional Kelly Multiplier to 0.25.
  2. Expected value per unit = 0.70 x 1.5 - 0.30 = 0.75, shown as 75.00 per 100 units.
  3. Full Kelly = 0.70 - 0.30 / 1.5 = 50.00 percent. Bet size = 50.00 x 0.25 = 12.50 percent.
Optimal Bet Size
12.50
Risk Analysis: Half-Kelly (0.5) %
25.00%

The final size of 12.50 percent is a quarter of the full Kelly, not the half-Kelly figure: the analysis list always shows half of full Kelly as a reference line, whatever multiplier you actually entered. Keep the two ideas separate when you read the results.

How do I size a bet with a 50 percent win probability and a ratio of 3?

Fractional Kelly Multiplier
1
Profit/Loss Ratio (Odds)
3
Win Probability (%)
50 %
  1. Set Win Probability to 50, Profit/Loss Ratio to 3, and Fractional Kelly Multiplier to 1.
  2. Expected value per unit = 0.50 x 3 - 0.50 = 1.00, shown as 100.00 per 100 units.
  3. Full Kelly = 0.50 - 0.50 / 3 = 33.33 percent after rounding, and the multiplier of 1 leaves the bet size at 33.33.
Optimal Bet Size
33.33
Expected Value (EV)
100.00

A model size of one third of the bankroll is very aggressive and would swing the bankroll sharply from one outcome to the next. It follows mechanically from the favourable assumed numbers; it is not a recommendation to stake that much on anything real.

Formula

Let p be the win probability as a fraction, q = 1 - p the loss probability, and b the profit to loss ratio (profit per unit risked when you win, for each 1 unit lost when you lose). Expected value per unit = p x b - q. When that expected value is positive, the full Kelly fraction = p - q / b; when it is zero or negative, the Kelly fraction is 0. The displayed bet size = full Kelly fraction x the fractional multiplier, expressed as a percentage.

How to use

  1. Enter the Win Probability as a percentage strictly between 0 and 100, for example 55 for a 55 percent chance of winning.
  2. Enter the Profit/Loss Ratio as the profit per unit risked: 1 means you profit 1 unit when you win for each 1 unit you lose.
  3. Optionally enter a Fractional Kelly Multiplier such as 0.5 to halve the suggested size or 0.25 to quarter it; leaving it blank uses the full size.
  4. Read the Optimal Bet Size percentage and the Expected Value figure.
  5. Check the Risk Analysis list for the full Kelly percentage, the half-Kelly reference figure, and the Bet or Skip status, remembering that only the main bet size applies your multiplier.

Frequently asked questions

What happens when the expected value of my assumed bet is zero or negative?

The engine sets the Kelly percentage to 0, the bet size to 0.00, and the status to Skip. It does not scale the stake down to a smaller positive amount; within this model, a bet with no edge is sized at nothing.

What does the expected value figure in dollars actually mean?

It is the model's expected profit for each 100 units wagered, in the long run, if the probability and ratio you entered were exactly right and the same bet could be repeated many times. It is a property of your assumptions, not a prediction or a promise of profit on any single bet or in real markets.

Why does the half-Kelly line in the analysis not match the multiplier I entered?

The full Kelly line always shows the unscaled Kelly percentage, and the half-Kelly line always shows exactly half of it, regardless of the multiplier you entered. Only the main bet size card applies your multiplier. With a multiplier of 0.25, as in the fourth example, your bet size is a quarter of full Kelly even though the list still displays the half-Kelly reference figure.

What happens if I leave the fractional multiplier blank or set it to 0?

If the multiplier is empty, zero, negative, or not a number, the calculation quietly treats it as 1, meaning full Kelly. Validation does not check this field at all, so check the value yourself before relying on the bet size; a blank field does not mean a smaller, safer stake.

How should I express the payout as a profit to loss ratio when my odds are quoted differently?

Enter the net profit you would make per unit risked. If a winning outcome pays you 1.5 units of profit for each 1 unit you would lose, the ratio is 1.5. This field expects that profit to loss ratio, not decimal odds that include your stake back; entering total payout odds instead would overstate the edge and inflate the size.

Is the optimal bet size a safe amount to stake on a real bet or trade?

No. The formula assumes your probability estimate is exact, that the same bet repeats independently many times, and that you can stake any fraction of a bankroll. Real probabilities are estimates that can be badly wrong, full Kelly sizes are highly volatile, and a wrong edge can shrink a bankroll quickly. This calculator explains a mathematical model; it is not investment or gambling advice and it does not promise any return.

Complementary Tools

Back to FinanceAll Calculators