You have learned how to find value bets using Expected Value. You know which bets have a positive edge. But that answers only half the question. The other half: how much should you stake?

Bet too little, and you leave profit on the table. Bet too much, and a bad streak can wipe out your bankroll. The Kelly Criterion solves this problem mathematically -- it tells you the exact percentage of your bankroll to stake on each bet to maximize long-term growth.

What is the Kelly Criterion?

The Kelly Criterion was developed by John Kelly Jr. at Bell Labs in 1956. Originally designed to optimize signal noise in telephone lines, it was quickly adopted by gamblers and investors who realized it solved a universal problem: how to size bets when you have an edge.

The core idea is simple. If you have a positive expected value, the Kelly formula tells you the fraction of your bankroll that maximizes the rate of growth over time. Stake this exact fraction, and your bankroll grows faster than with any other staking strategy -- in the long run.

Too aggressive, and volatility destroys you. Too conservative, and you grow too slowly. Kelly finds the optimal point between the two.

The Kelly Formula

The Kelly Criterion formula for decimal odds is:

f* = (b × p − q) / b

Where:

• f* = the fraction of your bankroll to stake
• b = the decimal odds minus 1 (the net odds, i.e. your potential profit per unit staked)
• p = the probability of winning (as a decimal)
• q = the probability of losing (1 − p)

Worked Example

You find a bet at decimal odds of 2.50. You estimate the true probability of winning at 45%.

b = 2.50 − 1 = 1.50
p = 0.45
q = 0.55

f* = (1.50 × 0.45 − 0.55) / 1.50
f* = (0.675 − 0.55) / 1.50
f* = 0.125 / 1.50
f* = 0.0833 (8.33%)

With a 1,000 EUR bankroll, Full Kelly says stake 83.30 EUR on this bet.

Try It Yourself

Full Kelly vs Fractional Kelly

In theory, Full Kelly maximizes long-term growth. In practice, it produces extreme volatility. Your bankroll can swing 30-50% in a single losing streak. Most bettors cannot tolerate this psychologically, and small errors in probability estimation make Full Kelly dangerously aggressive.

That is why experienced bettors use fractional Kelly:

• *Half Kelly (f/2)**: The most popular choice. It captures roughly 75% of the growth rate of Full Kelly while cutting variance in half. In the example above, you would stake 4.17% instead of 8.33%.
• *Quarter Kelly (f/4)**: Even more conservative. Ideal when your probability estimates have high uncertainty.

Why Over-Betting Is Catastrophic

Staking more than Full Kelly does not just increase risk -- it actually decreases your expected growth rate. At 2x Kelly, your expected growth rate drops to the same level as not betting at all. Beyond that, ruin becomes almost certain.

The chart below simulates 200 bets with four different staking strategies, all with the same edge (odds 2.50, true probability 45%):

Notice how 2x Kelly (over-betting) crashes. The Full Kelly line grows fastest but with high volatility. Half Kelly delivers smooth, steady growth. Flat staking (2% per bet) is safe but leaves returns on the table.

Practical Kelly Workflow

Here is how to apply Kelly in your daily betting routine:

Step 1: Estimate the true probability

Use prediction aggregators, your own model, or consensus data from sites like MetaPredictions. The better your probability estimate, the more reliable your Kelly stake.

Step 2: Calculate the Kelly fraction

Use the formula or the calculator above. If the result is negative or zero, do not bet -- there is no edge.

Step 3: Apply a fractional Kelly

Multiply the Kelly fraction by 0.5 (Half Kelly) or 0.25 (Quarter Kelly). This protects you against estimation errors.

Step 4: Calculate your stake

Multiply the adjusted fraction by your current bankroll. Not your starting bankroll -- your current bankroll. This is key: Kelly automatically reduces your stakes after losses and increases them after wins.

Step 5: Track and adjust

Keep a spreadsheet of every bet with the estimated probability, odds, Kelly fraction, and result. Over time, review your probability calibration. If your 60% estimates are winning only 50% of the time, adjust your model.

Kelly Limitations

The Kelly Criterion is powerful but not perfect. Be aware of these limitations:

Estimation errors amplify risk. Kelly assumes you know the true probability. In reality, you are estimating. A 5% overestimation of your edge can turn an optimal bet into over-betting. This is the primary reason to use fractional Kelly.

Simultaneous bets are tricky. The standard Kelly formula is for a single bet. If you have 5 bets on the same day, you cannot simply apply Kelly to each independently -- the total stake could exceed your bankroll. A simple fix: divide each Kelly stake by the number of simultaneous bets.

It assumes you can bet fractional amounts. Kelly might suggest staking 2.37% of your bankroll. If your bankroll is small, rounding to the nearest unit can introduce meaningful error.

It does not account for correlations. If two of your bets are on teams in the same match, the outcomes are correlated. Standard Kelly does not handle this.