ServedBets

Free tool

Kelly Criterion & Value Calculator

Enter your own estimated win probability and the odds on offer. This tells you the market's implied probability, your edge, and what a full or fractional Kelly stake would be, using the same math behind every ServedBets pick.

Implied prob.
-
Edge
-
Full Kelly stake
-
Suggested stake
-
Embed this calculator on your own site

Free to use anywhere. A small "Powered by ServedBets" link stays on the embed. Paste this where you want it to appear:

How this is calculated, exactly

Implied probability:1 ÷ decimal odds. What the market is pricing the outcome at, margin included. Seethe full breakdown.

Edge: your estimate minus the implied probability. Positive means you think the true chance is higher than the price suggests.

Full Kelly stake uses the standard formula, with b as the net odds (decimal odds − 1) and p as your probability:

f = (b·p − (1 − p)) / b

This is the fraction of your bankroll that maximizes long-run growthif your probability estimate is exactly right. It's provably optimal under that condition, and provably punishing when the estimate is off, because it stakes hardest on exactly the bets where your edge (and any error in it) is largest.

Suggested stake is the full Kelly fraction multiplied by whatever fraction you pick above (a quarter by default). Real quarter-Kelly still lost to flat staking on our own tracked picks. Seewhy, and why we default to flat 1-unit stakes ourselves.

This calculator does not know anything about tennis. It only does the math on the numbers you give it. Our own probability estimates come from the model described on the Methodology page. Want to see the realistic range of outcomes over many bets instead of a single one? Try theVariance & Bankroll Calculator.

Worked example: 55 % at odds 2.00

  1. Expected value: 0.55 × 2.00 − 1 = +10 % per unit staked.
  2. Full Kelly: (b × p − q) ÷ b with b = 1, p = 0.55, q = 0.45 gives (0.55 − 0.45) ÷ 1 = 10 % of the bankroll.
  3. Quarter Kelly: 2.5 % of the bankroll, the far more common choice in practice.
  4. If your 55 % is really 52 %, full Kelly would already be overbetting, which is why fractions are used.

Common mistakes

  • Plugging in an optimistic probability. Kelly assumes your estimate is right; small errors lead to big overbets.
  • Using the bookmaker’s implied probability as "p". Then the edge is zero and Kelly says bet nothing, correctly.
  • Staking full Kelly. The swings are brutal even with a real edge; see the variance calculator.

In tennis

Our own picks need an edge of at least 3 percentage points and +2 % expected value. Stakes follow fixed tiers, and Kelly is kept only as a capped reference, because a probability estimate is never exact.

Read next: Variance & bankroll calculator · Flat staking vs. Kelly · What is value betting?

Want the probability estimate too, not just the math?

We publish our own model's win probability and edge for every tennis pick in the Discord, live.

Join Discord (opens in a new tab)