Kelly · Calc

Kelly Criterion
Calculator.

Plug in bankroll, decimal odds, and your model probability — get the optimal stake, fractional Kelly (¼ default), EV %, and edge against the implied probability. Built for paper-trading discipline, not get-rich-quick math.

Fractional Kelly Risk-of-ruin aware ¼ Kelly default
01 · Inputs · Outputs

Run the math

Real-time output: change any input, see the result update instantly. Currency-agnostic — pick your unit.

Inputs
Total amount you allocate to betting. Currency-agnostic.
Bookmaker odds in decimal format. Minimum 1.01.
Your model's win probability (0.01–0.99). Implied prob from odds = 1/k.
Industry standard: ¼ Kelly. Cuts variance ~75% with ~25% less log-growth.
Outputs
Optimal stake
—bankroll units
Stake as % of bankroll
—
Expected Value (EV)
—
Edge · p − implied
—
⚠ Full Kelly maximizes growth but has ~50% drawdown variance. ¼ Kelly is industry standard for sustainable bankroll.
Watch us apply ¼ Kelly live on /track
See live track → Read methodology
02 · Theory

The math, the trade-offs, the limits

The Kelly criterion is a sizing formula, not a strategy. It tells you how much to bet given a known edge — it does not find the edge for you.

[ Formula ]

Kelly fraction f*

f* = (b · p − q) / b b = k − 1 (net odds) p = win probability q = 1 − p (loss probability)

f* is the share of bankroll that maximizes the long-run logarithmic growth rate. If f* ≤ 0, the bet has no edge and Kelly says: don't bet. Negative Kelly means the implied probability already exceeds your model — passing is the right move.

[ Fractional ]

Why ¼ Kelly, not Full

Full Kelly is mathematically optimal only if your probability estimate is exact. In practice, p comes from a model with error. A 5pp overestimate at Full Kelly can ruin you; at ¼ Kelly it's a small leak.

¼ Kelly captures ~94% of the geometric growth of Full Kelly while cutting drawdown variance by roughly 75%. The trade-off is overwhelmingly worth it.

[ When NOT to use Kelly ]

Kelly assumes things that rarely hold

  • Low sample size — your p estimate has wide confidence intervals. Kelly compounds estimation error.
  • Model uncertainty — if you can't quote a 95% CI around p, drop to ⅒ Kelly or pass.
  • Correlated bets — multiple simultaneous bets on the same outcome (e.g., parlays, same-game) violate Kelly's independence assumption.
  • Illiquid markets — when your stake moves the line, the odds at execution are not the odds Kelly used.
  • p from soft books — using the very same bookmaker's implied probability as p is circular. Always use sharp consensus (Pinnacle / Betfair / Matchbook) as the reference.

Read /methodology →

[ How Σ applies it ]

¼ Kelly · 1.5% hard cap · paper-only

Σ Value Sniper applies ¼ Kelly with a hard cap of 1.5% of bankroll per bet, on top of a CLV-positive filter and a sharp-consensus probability. This combination is intentionally conservative — the goal is to survive variance long enough for CLV to compound, not to ride the curve at the edge of ruin.

See /track →

⚠️ Calculator is a math tool, NOT financial advice. Sports betting carries financial risk. The Kelly criterion only protects you when your probability estimate is honest and well-calibrated. Garbage-in, garbage-out applies. Σ Value Sniper is paper-trading only — no real money is at stake on our end. 18+. Do not bet what you cannot afford to lose.