Bankroll ManagementPublished 1/25/2026· Updated 1/25/2026· 14 min read

Kelly Criterion Calculator: Optimize Your Betting Stake Size (Free Tool)

Betting the right amount matters more than picking winners. You can hit 60% winners and still lose money if you're betting too much during cold streaks or too little when you have an edge.

Kelly Criterion Calculator: Optimize Your Betting Stake Size (Free Tool)

Betting the right amount matters more than picking winners. You can hit 60% winners and still lose money if you're betting too much during cold streaks or too little when you have an edge. The Kelly Criterion is a mathematical formula that tells you exactly how much to stake based on your edge and bankroll—maximizing long-term growth while managing risk.

This guide explains how the Kelly Criterion works, shows you how to use our free calculator, and reveals why most bettors use it wrong (hint: overconfidence in win probability destroys bankrolls)

What Is the Kelly Criterion?

The Mathematical Formula Explained

The Kelly Criterion is a formula for optimal bet sizing developed by J.L. Kelly Jr. in 1956 at Bell Labs. Kelly's original paper, "A New Interpretation of Information Rate," was actually about signal transmission—but gamblers quickly realized its application to betting.

The core principle: Kelly maximizes long-term wealth growth while minimizing risk of ruin. According to Wikipedia's analysis of the Kelly criterion, this approach "leads almost surely to higher wealth compared to any other strategy in the long run."

Here's the formula: Kelly % = W - [(1-W)/R]

Where:

  • W = Your win probability (as a decimal)

  • R = Win/loss ratio (how much you win vs. how much you risk)

Why Kelly Criterion Matters for Bettors

Most bettors either bet the same amount every time (flat betting) or use arbitrary percentages. Kelly tells you the mathematically optimal stake based on your actual edge. If you're betting Lakers -110 with a 55% win probability, Kelly calculates exactly what percentage of your bankroll maximizes growth without risking ruin.

The catch? Kelly assumes you know your exact win probability—and that's where most bettors get destroyed. Overestimate by 5%, and Kelly recommends stakes that will drain your bankroll during inevitable variance.

How to Use Our Kelly Criterion Calculator

Our Kelly Criterion calculator does the math for you. Here's how to use it:

Step 1: Enter Your Win Probability

This is the most critical input—and the hardest to estimate accurately. Your win probability should come from:

  • Historical tracking: If you've logged 500+ bets and won 54%, use 0.54

  • Model outputs: If your betting model gives 58% probability, use 0.58

  • Market analysis: Compare your assessment to implied probability from odds

Don't guess. If you think Lakers "should probably win," that's not enough data for Kelly. You need a quantifiable edge based on actual performance or rigorous analysis.

Step 2: Input Your Odds

Enter the odds in your preferred format (American, Decimal, or Fractional). The calculator converts automatically. If you're betting Lakers -110, enter "-110" in American odds format. Need help converting between formats? Try our American to Decimal odds converter.

Step 3: Add Your Bankroll

Enter your current total bankroll—not the amount in your sportsbook account, but your dedicated betting fund. If you have $1,000 set aside for betting, enter 1000.

Step 4: Get Your Optimal Stake

The calculator shows your recommended stake as both a percentage and dollar amount. For example, with a $1,000 bankroll, 60% win probability, and +150 odds:

  • Kelly %: 10%

  • Recommended stake: $100

The calculator also shows Half Kelly (50% of the recommended stake) and Quarter Kelly (25%)—more on why these matter below.

Pro tip: Track your bets with mybets.gg to refine your win probability over time. The more data you have, the more accurate your Kelly calculations become.

The Kelly Criterion Formula: Breaking Down the Math

Understanding Each Variable

W (Win Probability): Calculate this from historical data. If you've won 265 of 500 bets, W = 265/500 = 0.53 or 53%. This should be based on a significant sample size—at least 200 bets for a reliable estimate.

R (Win/Loss Ratio): This is the net odds received on a win divided by the amount lost on a loss. For a $100 bet at +150 odds, you win $150 and risk $100, so R = 150/100 = 1.5.

Converting Odds to the Win/Loss Ratio

Different odds formats require different conversions:

  • American odds (positive): R = (Odds/100). For +200, R = 200/100 = 2.0

  • American odds (negative): R = (100/|Odds|). For -110, R = 100/110 = 0.909

  • Decimal odds: R = (Decimal odds - 1). For 2.50, R = 2.50 - 1 = 1.5

Manual Calculation Example

Let's work through an NFL bet at -110 odds with a 53% estimated win probability:

  1. W = 0.53 (your estimated win probability)

  2. R = 0.909 (100/110 for -110 odds)

  3. Kelly = 0.53 - [(1-0.53)/0.909]

  4. Kelly = 0.53 - [0.47/0.909]

  5. Kelly = 0.53 - 0.517

  6. Kelly = 0.013 or 1.3%

With a $5,000 bankroll, Kelly recommends betting $65 (1.3% of $5,000). That's it—simple math that tells you exactly how much to stake.

The calculator saves you from doing this manually for every bet. Just plug in your numbers and get instant results.

Why Most Bettors Use Kelly Wrong (And How to Fix It)

The Overconfidence Problem

Here's the contrarian truth most betting sites won't tell you: Kelly Criterion is presented as a simple "plug and play" solution, but it has a critical flaw. The formula assumes you know your exact win probability—and you almost certainly don't.

Think you have a 58% edge on that Celtics spread? What if it's actually 55%? That 3% difference means Kelly recommends a stake that's too large for your actual edge. Over hundreds of bets, this overconfidence compounds into significant losses.

According to Investopedia's analysis of Kelly limitations, "The biggest limitation is that it requires an accurate assessment of probability—something that's extremely difficult in practice."

Fractional Kelly: The Safer Approach

Full Kelly can recommend 10-20% of your bankroll on a single bet if you have a strong edge. That creates massive variance. Even if your edge is real, variance can produce brutal losing streaks that test your discipline.

Enter fractional Kelly: betting a fraction of the Kelly-recommended stake. The most common approaches:

  • Half Kelly: 50% of the calculated stake

  • Quarter Kelly: 25% of the calculated stake

Here's a comparison over 100 bets with a 2% edge:

Method Average Growth Worst Drawdown Volatility Full Kelly +18.2% -32% High Half Kelly +16.1% -18% Medium Quarter Kelly +12.4% -9% Low

When Full Kelly Is Too Aggressive

Even professional bettors rarely use full Kelly. The volatility is too extreme for most bankrolls and psychological comfort levels. Half Kelly gives you 75% of the growth with 50% of the volatility—a much better risk/reward tradeoff.

Recommendation: Start with Half Kelly (50% of the calculated stake) until you have 500+ tracked bets validating your win probability estimates. Then consider moving to 60-70% Kelly if your edge proves consistent.

Kelly Criterion vs. Other Staking Methods

How does Kelly stack up against other bankroll management approaches?

Method Risk Level Bankroll Growth Complexity Best For Kelly Criterion Medium-High Optimal (if edge accurate) Medium Bettors with proven edge Flat Betting Low Slow but steady Simple Beginners, conservative bettors Fixed Percentage Low-Medium Steady growth Simple Regular bettors Martingale Extreme Negative expectation Simple Nobody (avoid completely)

Flat Betting: Bet the same amount every time ($50 per bet regardless of odds or edge). Simple and safe, but doesn't maximize growth or adjust for varying edges.

Fixed Percentage: Bet a fixed percentage of your current bankroll (2% per bet). Automatically adjusts to bankroll changes but doesn't account for varying edge across different bets.

Martingale: Double your bet after each loss to recover losses. This strategy guarantees eventual ruin—don't use it. Ever.

According to the research on Kelly's long-term performance, Kelly provides superior wealth accumulation compared to any other strategy when you have an edge. The key phrase: "when you have an edge."

Real-World Kelly Criterion Examples

Example 1: NFL Point Spread Bet

Scenario: You're betting on Ravens -3.5 at -110 odds. Your model shows 52% win probability. Your bankroll is $5,000.

Calculation:

  • W = 0.52

  • R = 0.909 (100/110)

  • Kelly = 0.52 - [(1-0.52)/0.909] = 0.52 - 0.528 = -0.008

Result: Kelly recommends 0% stake. Why? Because 52% isn't enough edge to overcome -110 juice. You need at least 52.4% to break even at -110 odds. This bet has negative expected value—skip it.

Example 2: Underdog Moneyline

Scenario: Hawks +250 underdog. You estimate 35% win probability (implied probability is 28.6%). Bankroll: $5,000.

Calculation:

  • W = 0.35

  • R = 2.5 (250/100)

  • Kelly = 0.35 - [(1-0.35)/2.5] = 0.35 - 0.26 = 0.09 or 9%

Result: Full Kelly recommends $450 (9% of $5,000). Half Kelly recommends $225. This is a high-variance bet—even with an edge, you'll lose 65% of the time. Half Kelly is the smarter choice here.

Example 3: Multiple Bets with Different Edges

Scenario: You have three simultaneous bets, each with different Kelly recommendations:

  • Bet A: Kelly recommends 4% ($200)

  • Bet B: Kelly recommends 3% ($150)

  • Bet C: Kelly recommends 2% ($100)

Total Kelly allocation: 9% ($450)

When you have multiple simultaneous bets, Kelly percentages add up. According to Bloomberg's analysis of Kelly for spreading bets, you need to consider total exposure. If 9% feels too aggressive, reduce each bet proportionally (Half Kelly on all three = 4.5% total exposure).

Tools like mybets.gg help you track which bet types consistently give you the best edge, allowing you to refine your Kelly approach over time.

Common Kelly Criterion Mistakes to Avoid

  1. Overestimating your win probability: This is the #1 killer. Most bettors think they're better than they are. Use historical data, not gut feelings. If you don't have 200+ tracked bets, you don't have reliable win probability data.

  2. Using Kelly on bets without a proven edge: Kelly optimizes bet sizing for +EV bets. It doesn't create an edge. If you're betting -110 lines with 50% win probability, Kelly correctly tells you to bet 0%—because you have no edge.

  3. Not adjusting for correlation in parlays: Kelly assumes independent bets. Parlays and same-game parlays have correlated outcomes. Standard Kelly calculations don't apply—you need more complex models.

  4. Ignoring bankroll fluctuations: Recalculate Kelly stakes when your bankroll changes by 20% or more. Using an outdated bankroll figure leads to oversized or undersized bets.

  5. Applying full Kelly without understanding variance: Full Kelly can recommend 15-20% stakes on strong edges. That creates wild swings. Start with Half Kelly until you're comfortable with the volatility.

  6. Failing to track results to validate assumptions: You can't improve what you don't measure. Track every bet to see if your estimated win probabilities match reality. According to Investopedia's Kelly limitations analysis, "Without accurate probability estimates, Kelly can lead to overbetting and significant losses."

Mybets.gg's automatic tracking prevents mistake #6 by logging all your bets and calculating actual win rates across different bet types, odds ranges, and sports.

Kelly Criterion for Different Betting Markets

Sports Betting Applications

Sports betting is ideal for Kelly because you can build historical data. After 500+ NFL spread bets, you know your actual win rate. You can segment by favorite/underdog, home/away, division games, etc., to refine your edge estimates.

According to academic research on Kelly in sports betting contexts, the criterion performs best in markets where you can develop information advantages through analysis, modeling, or faster information access.

Horse Racing and High-Odds Events

Horse racing and other high-odds markets require more conservative fractional Kelly. When you're betting +800 longshots, even a small overestimation of win probability leads to massive overbetting. Use Quarter Kelly (25%) in markets with high variance and less predictable outcomes.

Live Betting Considerations

Live betting probabilities change constantly. That Lakers game you estimated at 60% win probability shifts to 45% after they go down 10 points in the first quarter. Kelly requires quick recalculation—or better yet, use our calculator to adjust stakes in real-time as probabilities shift.

Kelly works best in markets where you have an information edge. If you're betting the same lines as everyone else with the same information, you probably don't have the edge Kelly requires. Track your performance across different markets with mybets.gg to identify where your edge actually exists.

Frequently Asked Questions

What happens if Kelly Criterion recommends betting more than my bankroll?

This means your estimated edge is unrealistic or you've miscalculated. Kelly never recommends betting more than 100% of your bankroll. If you're getting results like 150% Kelly, recheck your win probability—you're almost certainly overestimating your edge. According to the mathematical properties of Kelly, the formula is bounded and won't recommend over-bankroll stakes unless inputs are wrong.

Can Kelly Criterion guarantee profits?

No. Kelly optimizes bet sizing but doesn't create an edge. You need a genuine advantage for Kelly to work. If you're betting random games with no edge, Kelly will correctly recommend 0% stakes. Even with an edge, variance means you'll experience losing streaks. Kelly maximizes long-run growth—it doesn't eliminate short-term volatility.

How often should I recalculate my Kelly stake?

Recalculate when your bankroll changes significantly (±20%) or when you have new data about your win rate. If you started with $5,000 and you're now at $6,000, your Kelly stakes should increase proportionally. mybets.gg tracks this automatically, updating your available bankroll after each bet so your Kelly calculations stay current.

What's the difference between Kelly Criterion and expected value?

Expected value (EV) tells you if a bet is profitable. Kelly tells you how much to stake on profitable bets. You need positive EV before using Kelly. If a bet has negative EV, Kelly correctly recommends 0% stake. Use our expected value calculator to identify +EV opportunities, then use Kelly to size your stakes optimally.

Yes. Kelly is just a mathematical formula for bankroll management. It's a tool for sizing bets, not a betting system that violates any regulations. Sportsbooks don't care what staking method you use—they care whether you're exploiting technical vulnerabilities or colluding. Kelly is completely legal everywhere betting is legal.

Track Your Kelly Performance with mybets.gg

Kelly Criterion requires accurate win probability data—and that comes from tracking. You can't estimate your edge on NFL spreads without knowing your historical win rate on NFL spreads. You can't refine your estimates without data showing where you're over or underestimating.

mybets.gg automatically logs your bets via Chrome extension. Take a screenshot of your betslip, and the extension captures all the details—odds, stake, bet type, sport. No manual entry. No spreadsheets. Just automatic tracking that builds the historical data Kelly needs.

The dashboard shows your ROI by sport, bet type, and odds range. You can see that you're hitting 56% on NBA spreads but only 51% on NFL spreads. That's the difference between a bet Kelly recommends at 4% and one it recommends at 0.5%.

Bankroll tracking keeps your Kelly calculations current. As your bankroll grows or shrinks, mybets.gg updates automatically, so you're always betting the right percentage based on your current capital—not what you had three months ago.

Start tracking your bets free to improve your Kelly calculations and validate whether your estimated edges match reality.

Key Takeaways

  • The Kelly Criterion calculates optimal bet size using the formula: Kelly % = W - [(1-W)/R], where W is win probability and R is win/loss ratio

  • Most bettors should use Half Kelly (50% of the calculated stake) to reduce variance while still optimizing growth

  • Kelly only works if you have a genuine edge—it optimizes bet sizing but doesn't create profitable opportunities

  • Accurate win probability is critical: overestimating by even 5% can lead to significant bankroll damage over hundreds of bets

  • Track your betting results with tools like mybets.gg to refine your win probability estimates and validate your Kelly approach

Conclusion

The Kelly Criterion is the mathematically optimal way to size your bets—but only if you use it correctly. Start with our free calculator above, use fractional Kelly to manage risk, and track your results to refine your win probability estimates over time.

Remember: Kelly doesn't create an edge, it maximizes the edge you already have. If you're betting without an edge, Kelly will correctly tell you to bet nothing. If you have an edge but overestimate it by 5%, Kelly will recommend stakes that are too large for your actual advantage.

The solution? Data. Track every bet. Calculate your actual win rates. Segment by sport, bet type, and odds range. Build the historical foundation that makes Kelly work.

Want to improve your Kelly calculations? Start tracking your bets with mybets.gg to build the historical data you need for accurate win probabilities. The Chrome extension captures betslips automatically, calculates ROI by market, and keeps your bankroll current—giving you the foundation for smarter Kelly-based betting.

---

Disclaimer: Betting involves risk. Only bet what you can afford to lose. If you or someone you know has a gambling problem, seek for help.

Share this post:
MB

Written by the Mybets.gg Team

Our team of sports betting analysts, data scientists, and professional bettors working together to bring you the most accurate strategies and insights. We combine data-driven analysis with real-world experience.