Find the optimal stake size based on your edge, probability, and bankroll.
Kelly Calculator
Enter the bookmaker's odds, your estimated probability of winning, your bankroll, and pick a Kelly fraction. The calculator returns the optimal stake for full, half, quarter, and custom Kelly side-by-side.
Recommended Stakes
Edge: Implied Probability: Your Probability: Probability Gap:
Strategy
Stake %
Stake Amount
Profile
Full Kelly
Aggressive
Half Kelly ★ Recommended
Recommended
Quarter Kelly
Conservative
Custom (50%)
Your choice
Why fractional? Full Kelly maximizes long-term growth in theory but produces large bankroll swings in practice. Most professional bettors stake a quarter to half of full Kelly to reduce drawdown.
⚠ DON'T BET — No value detected
Your Probability: Implied Probability: Probability Gap: Edge:
Kelly Criterion only recommends a stake when your estimated probability gives a positive edge over the bookmaker's implied probability. With these inputs, betting would have negative expected value over time — you'd lose money.
Consider:
Re-evaluating your probability estimate — is your number really above the bookmaker's implied probability?
Looking for better odds elsewhere — a different bookmaker may offer enough value.
Skipping this bet — discipline beats forced action over the long run.
Kelly answers the question every bettor faces after spotting a value bet: "How much should I stake?" Bet too little and you leave growth on the table; bet too much and a normal losing streak wipes you out.
How the Kelly Criterion works
The formula is f* = (b·p − q) / b, where b = decimal odds − 1, p = your estimated win probability, and q = 1 − p. With odds 2.50 and p = 0.45: b = 1.5, q = 0.55. f* = (1.5 × 0.45 − 0.55) / 1.5 = 0.0833, i.e., bet 8.33% of bankroll. If (b·p − q) is zero or negative, the optimal stake is zero — don't bet.
Why use fractional Kelly
Full Kelly is mathematically optimal for unlimited horizons and perfectly accurate probability estimates — neither of which exist in real betting. Even a small overestimation of your edge causes full Kelly to over-bet and crash the bankroll. Half Kelly cuts variance roughly in half while still capturing most of the long-term growth, which is why it's the practitioner default.
Common mistakes
Plugging in the bookmaker's implied probability and expecting Kelly to find an edge — by definition there is none.
Using a static stake instead of recomputing as the bankroll changes.
Confusing Kelly with progressive systems like Martingale — Kelly is edge-driven, not loss-chasing.
Ignoring correlation between simultaneous bets — Kelly assumes independent outcomes.
Worked Examples
Positive edge: a clear value bet
Odds 2.50, your estimated probability 45%, bankroll $1,000. Full Kelly recommends 8.33% of bankroll ($83.33). Half Kelly cuts that to 4.17% ($41.67). Edge is +12.5%, probability gap is +5pp — a clean value bet that justifies a measured stake.
Negative edge: don't bet
Odds 2.50, your estimated probability 35%, bankroll $1,000. Bookmaker's implied probability is 40% — your estimate is below it. Edge is −12.5%. The calculator returns "DON'T BET" rather than a stake. Skipping this is the right move; over time, you'd lose 12.5% of every stake.
Compare odds across bookmakers
Same outcome, your probability 45%. Bookmaker A offers 2.40 (edge +8%, full Kelly 5.7%), Bookmaker B offers 2.60 (edge +17%, full Kelly 10.5%). The same bet at better odds nearly doubles your optimal stake — always check multiple books before staking.
The Kelly Criterion formula is f* = (b·p − q) / b, where f* is the optimal fraction of bankroll to bet, b is the decimal odds minus 1, p is your estimated probability of winning, and q = 1 − p. The result is multiplied by your bankroll to get the absolute stake.
Most professional bettors use somewhere between quarter and half Kelly. Full Kelly is mathematically optimal only when your probability estimates are perfectly accurate, which is virtually never the case. Half Kelly cuts variance roughly in half at the cost of about 25% of long-term growth — a trade most bettors gladly accept.
The calculator returns a "DON'T BET" warning when your estimated win probability gives zero or negative expected value at the offered odds. In that case, Kelly's mathematically optimal stake is zero — betting would lose money on average. Either re-evaluate your probability, look for better odds, or skip the bet.
