Kelly works – but almost never on casino games. The formula maximizes long-term growth when you have a positive expected return. On slots, roulette, and table games without card counting, expected return is always negative. In that case Kelly says you should bet zero.
Source: Kelly, J.L. (1956), Bell System Technical Journal
What is the Kelly Criterion?
The Kelly Criterion is a formula for deciding how much of your bankroll to bet on a single wager. John L. Kelly Jr. published it in 1956 at Bell Labs. Since then, Warren Buffett mentor Ed Thorp has used it to beat blackjack in Vegas, and hedge funds like Renaissance Technologies have built billion-dollar fortunes on the same principle.
Kelly maximizes the long-term growth rate of your bankroll – given that you have an edge. In mathematical terms: the formula maximizes the logarithm of wealth over an infinite number of bets. It is the only known strategy that is neither too cautious nor too aggressive.
The Formula Explained
For betting with binary outcomes (win or lose), the Kelly formula is:
- f* = Optimal fraction of bankroll to bet
- b = Net odds (profit per unit wagered)
- p = Probability of winning
- q = Probability of losing (1 − p)
The formula requires that you have a positive expected return (edge). If bp − q is negative or zero, Kelly says you shouldn't bet at all. In casino games, the house edge is usually against you, which makes Kelly strategy not directly applicable to most games.
Practical Example
Let us walk through a concrete example. Say you have a game with the following parameters:
Step-by-step calculation
- b = 2.10 − 1 = 1.10 (net odds)
- bp = 1.10 × 0.55 = 0.605
- bp − q = 0.605 − 0.45 = 0.155
- f* = 0.155 / 1.10 = 0.141 (14.1%)
Result: Kelly says you should bet 14.1% of your bankroll, which is 1,410 of 10,000.
Fractional Kelly – what the pros actually do
Full Kelly is mathematically optimal. But in practice almost no serious gamblers run full Kelly – they use half (50%) or quarter Kelly (25%). We explain why.
- Extremely volatile – large bankroll swings
- Requires exact knowledge of probabilities
- Psychologically difficult to handle drawdowns
- Small estimation errors can be catastrophic
- Retains 75% of growth rate
- Halves volatility
- More robust against estimation errors
- Psychologically sustainable
The Kelly curve is parabolic. Betting 2x the Kelly fraction yields exactly the same expected growth as not betting at all (zero). Betting even more leads to guaranteed ruin over time.
Application to casino games – short answer: no
This is where it gets uncomfortable for casino players. Kelly requires positive expected return, and most casino games have a house edge. That means the formula mathematically says you should bet zero. We still looked at where Kelly works – and where trying to apply it is pure self-deception.
- Blackjack with card counting: can give 0.5–1.5% edge at high counts
- Poker against other players: skilled players can have an edge
- Certain promotions: rare +EV offers
- Slots: always negative EV (house edge 2–15%)
- Roulette: fixed house edge (2.7–5.26%)
- Baccarat: no strategic edge possible
Do not use Kelly. Use the 1–2 percent rule instead: set a budget you can afford to lose and split it into many small bets (1–2 % of your bankroll per game). Your money lasts longer, you reduce the risk of impulsive big losses, and gambling stays where it belongs – as entertainment.
Common mistakes we have seen
- Overestimate your edge. If you think you have 5% edge but actually have 2%, Kelly will give you bets that are too large. This is the most common mistake.
- Apply to –EV games. The Kelly Criterion is not designed for games where you don't have an edge. Trying to apply it to slots or roulette is mathematically incorrect.
- Ignore volatility. Full Kelly can result in 50%+ drawdowns. If you can't handle seeing your bankroll temporarily halve, use fractional Kelly.
- Not updating bankroll. Kelly requires recalculating bet size based on current bankroll after each bet. Using a fixed amount loses Kelly's benefits.
- The Kelly Criterion is mathematically proven to maximize long-term growth – when you have an edge.
- On most casino games you do not have an edge. In that case Kelly says: bet zero.
- Pros run half or quarter Kelly to handle volatility and estimation errors.
- Betting more than Kelly actually reduces your expected returns. Betting 2× Kelly gives the same result as not betting at all.
- For casino games without edge: use the 1–2 percent rule and treat gambling as entertainment.
The best math in the world cannot turn a –EV game into a +EV game. Treat casino gambling as entertainment, not investment. Set a budget before you start, and walk away when it is gone. If gambling starts to control you instead of the other way around, seek help early.
Helplines: BeGambleAware.org · GamCare 0808 8020 133 (free, 24/7)
- Kelly, J.L. (1956). A New Interpretation of Information Rate. Bell System Technical Journal, 35(4), 917–926.
- Thorp, E.O. (2006). The Kelly Criterion in Blackjack, Sports Betting, and the Stock Market. Handbook of Asset and Liability Management.
- MacLean, L.C., Thorp, E.O., & Ziemba, W.T. (2011). The Kelly Capital Growth Investment Criterion. World Scientific.