Kelly Criterion Calculator Excel

Kelly Criterion Calculator

Calculate optimal bet sizing for maximum long-term growth using the Kelly Criterion formula

Use 1 for full Kelly, 0.5 for half-Kelly, etc.

Kelly Criterion Calculator: The Complete Guide to Optimal Bet Sizing

The Kelly Criterion is a mathematical formula developed by John L. Kelly Jr. in 1956 that determines the optimal size of a series of bets to maximize logarithmic utility (long-term growth) of capital. Originally applied to gambling, it has since become a cornerstone of modern portfolio theory and quantitative finance.

The Kelly Criterion Formula

The basic Kelly Criterion formula for positive expectation bets is:

f* = (bp - q) / b
where:
f* = fraction of current bankroll to wager
b = net odds received on the wager (decimal odds - 1)
p = probability of winning
q = probability of losing (1 - p)

Why Use the Kelly Criterion?

  • Maximizes long-term growth – The Kelly strategy grows capital faster than any other strategy in the long run
  • Bankroll management – Automatically adjusts bet sizes based on current bankroll
  • Risk control – Never risks more than the calculated edge suggests
  • Mathematically optimal – Derived from information theory and probability

Practical Applications

  1. Sports Betting – Determine how much to wager on a +EV (positive expected value) bet
  2. Stock Trading – Calculate position sizes for asymmetric risk/reward trades
  3. Poker – Manage bankroll when you have an edge over opponents
  4. Blackjack – Optimal betting when card counting gives you an advantage

Kelly Criterion vs. Fixed Fractional Betting

Metric Kelly Criterion Fixed Fractional (1%) Fixed Fractional (5%)
Long-term growth rate Maximized Suboptimal Suboptimal
Bankroll volatility High (but optimal) Low Moderate
Risk of ruin Minimized for given edge Higher for same edge Much higher
Adaptability Adjusts to changing edges Fixed regardless of edge Fixed regardless of edge
Psychological comfort Challenging for most Easier to maintain Moderate difficulty

Fractional Kelly: The Practical Compromise

While full Kelly (f* = 1) maximizes growth, it can lead to extreme volatility. Most professional investors use fractional Kelly (typically 0.3 to 0.7) to:

  • Reduce psychological stress
  • Limit maximum drawdowns
  • Account for estimation errors in edge calculation
  • Maintain more stable equity curves
Fractional Kelly Growth Rate Max Drawdown Psychological Ease
Full Kelly (1.0) 100% ~50-80% Very difficult
0.75 Kelly ~94% ~40-60% Difficult
0.5 Kelly (Half-Kelly) ~75% ~25-40% Manageable
0.25 Kelly ~38% ~10-20% Easy

Common Mistakes When Using Kelly Criterion

  1. Overestimating edge – The formula is extremely sensitive to input accuracy. A 1% overestimation of win probability can lead to disastrous results.
  2. Ignoring transaction costs – Commissions, spreads, or vig reduce your actual edge and should be factored into calculations.
  3. Using with negative expectation bets – Kelly only works for +EV situations. Using it on -EV bets will quickly deplete your bankroll.
  4. Not adjusting for changing conditions – Market edges and probabilities change; your Kelly fraction should be recalculated regularly.
  5. Psychological inability to follow – Many abandon Kelly during drawdowns, which is when discipline matters most.

Advanced Kelly Criterion Concepts

Multi-Asset Kelly

For portfolios with multiple independent bets, the formula extends to:

f*i = (b*i p*i - q*i) / b*i
where the sum of all f*i should not exceed 1 (full Kelly)

Continuous-Time Kelly

For continuous trading (like forex or algorithmic trading), the formula becomes:

f* = μ / σ²
where:
μ = expected return (drift)
σ² = variance of returns

Kelly Criterion in Excel

To implement Kelly Criterion in Excel:

  1. Create cells for:
    • Win probability (e.g., B2)
    • Decimal odds (e.g., B3)
    • Current bankroll (e.g., B4)
  2. In another cell, enter the formula: 
    =((B2*(B3-1))-(1-B2))/(B3-1)
  3. Multiply the result by your bankroll to get the bet size: 
    =MIN(Result_from_step_2*B4, B4)
  4. For fractional Kelly, multiply the result by your desired fraction (e.g., 0.5 for half-Kelly)

Academic Research on Kelly Criterion

The Kelly Criterion has been extensively studied in academic literature. Key papers include:

Kelly Criterion vs. Other Bet Sizing Strategies

Strategy Growth Rate Risk of Ruin Psychological Stress Best For
Full Kelly ★★★★★ ★★☆☆☆ ★★★★★ Mathematically optimal players
Fractional Kelly (0.5) ★★★★☆ ★☆☆☆☆ ★★★☆☆ Most practical applications
Fixed Fractional (1-3%) ★★☆☆☆ ★★☆☆☆ ★☆☆☆☆ Conservative investors
Martingale ★☆☆☆☆ ★★★★★ ★★★★★ Never (guaranteed ruin)
Flat Betting ★☆☆☆☆ ★☆☆☆☆ ★☆☆☆☆ Casual gamblers

Implementing Kelly Criterion in Real World Scenarios

To successfully apply Kelly Criterion:

  1. Accurately estimate probabilities – Use historical data, simulations, or expert analysis to determine true win probabilities
  2. Account for all costs – Include commissions, spreads, taxes, and any other frictional costs in your edge calculations
  3. Start with fractional Kelly – Begin with 0.2-0.3 Kelly to test your edge estimates and emotional resilience
  4. Track results meticulously – Compare actual win rates to estimated probabilities and adjust accordingly
  5. Set strict bankroll limits – Never exceed your predetermined fractional Kelly, even during losing streaks
  6. Combine with position sizing rules – Use Kelly to determine what percentage of capital to risk, then apply additional position sizing rules
  7. Regularly review performance – Reassess your edge and Kelly fraction at least quarterly

Limitations of Kelly Criterion

While powerful, Kelly has important limitations:

  • Assumes known probabilities – In reality, we only have estimates with uncertainty
  • Ignores utility preferences – Doesn’t account for individual risk tolerance
  • Sensitive to estimation errors – Small errors in edge estimation can lead to large suboptimal results
  • Assumes infinite divisibility – In practice, bet sizes are often constrained
  • No consideration for liquidity – Doesn’t account for market impact of large bets
  • Psychologically demanding – The optimal strategy can be emotionally difficult to follow

Kelly Criterion in Different Markets

Sports Betting

In sports betting, you must:

  • Convert moneyline odds to decimal format
  • Estimate true win probability (not the bookmaker’s implied probability)
  • Account for the bookmaker’s vig (overround)
  • Adjust for closing line movements

Stock Trading

For stock trading applications:

  • Use historical backtests to estimate win probability
  • Calculate expected return and volatility for position sizing
  • Consider correlation between positions in portfolio Kelly
  • Adjust for slippage and transaction costs

Poker

In poker, Kelly helps with:

  • Bankroll management for cash games
  • Tournament buy-in selection
  • Deciding when to move up/down stakes
  • Managing variance in high-stakes games

Alternative Bet Sizing Methods

When Kelly isn’t appropriate, consider:

  • Fixed Fractional – Bet a fixed percentage (1-5%) of bankroll per trade
  • Volatility-Based – Size positions based on instrument volatility (e.g., ATR)
  • Risk Parity – Allocate capital to equalize risk contribution across positions
  • Equal Position Sizing – Same dollar amount for each position
  • Anti-Martingale – Increase position size after wins, decrease after losses

Building Your Own Kelly Criterion Calculator in Excel

To create a robust Excel implementation:

  1. Create input cells for:
    • Win probability (as decimal, e.g., 0.55 for 55%)
    • Decimal odds
    • Current bankroll
    • Fractional Kelly multiplier
  2. Add calculation cells for:
    • Implied probability (1/decimal odds)
    • Edge (true probability – implied probability)
    • Full Kelly fraction
    • Adjusted Kelly fraction (with your multiplier)
    • Recommended bet size
  3. Add data validation to prevent:
    • Probabilities outside 0-1 range
    • Odds less than 1
    • Negative bankrolls
    • Fractional Kelly outside 0-1 range
  4. Create a results dashboard showing:
    • Bet size in dollars and percentage
    • Expected growth rate
    • Risk of ruin estimates
    • Sensitivity analysis (how changes in inputs affect outputs)
  5. Add visualizations:
    • Bankroll growth projections
    • Bet size vs. bankroll percentage
    • Sensitivity charts

Advanced Excel Techniques for Kelly Calculations

For sophisticated implementations:

  • Use DATA TABLES for sensitivity analysis
  • Implement GOAL SEEK to find required edge for desired growth
  • Create MONTE CARLO SIMULATIONS with RAND() functions
  • Use CONDITIONAL FORMATTING to highlight +EV opportunities
  • Build DYNAMIC CHARTS that update with input changes
  • Implement VBA MACROS for complex portfolio Kelly calculations

Kelly Criterion in Portfolio Management

For investment portfolios, the Kelly approach becomes more complex:

Portfolio Kelly = V⁻¹ μ
where:
V = variance-covariance matrix of asset returns
μ = vector of expected excess returns

This requires:

  • Estimating expected returns for each asset
  • Calculating pairwise correlations
  • Inverting the covariance matrix
  • Applying constraints (e.g., no short selling)

Psychological Aspects of Kelly Betting

The mathematical optimality of Kelly often conflicts with human psychology:

  • Loss aversion – People feel losses ~2x as strongly as equivalent gains
  • Recency bias – Overweighting recent results in edge estimation
  • Overconfidence – Overestimating edge and underestimating risk
  • Disposition effect – Holding losers too long and selling winners too soon
  • Herding behavior – Following crowd rather than mathematical edge

Solutions include:

  • Starting with very low fractional Kelly (0.1-0.2)
  • Automating bet sizing to remove emotion
  • Keeping detailed records to combat recency bias
  • Setting strict bankroll stop-loss limits
  • Regular psychological reviews of trading decisions

The Future of Kelly Criterion

Emerging applications include:

  • Cryptocurrency trading – Managing volatility in crypto markets
  • Algorithmic trading – Dynamic position sizing in quant funds
  • Esports betting – Applying to new betting markets
  • Machine learning – Using AI to estimate probabilities for Kelly inputs
  • Behavioral finance – Combining Kelly with prospect theory insights

Final Recommendations

  1. Always use fractional Kelly (0.3-0.5) when starting
  2. Conservatively estimate your true edge
  3. Never risk more than 1-2% of bankroll on any single bet
  4. Combine Kelly with other risk management techniques
  5. Regularly backtest and validate your edge estimates
  6. Be prepared for significant drawdowns even with optimal play
  7. Consider using Kelly as a maximum limit rather than exact sizing

Leave a Reply

Your email address will not be published. Required fields are marked *