Value bet calculator for expected value and Kelly criterion stakes

What expected value means

Expected value, or EV, is the single most useful idea in betting, and it is simpler than it sounds. It is the average outcome of a bet if you could repeat it endlessly. Positive EV means the price is in your favour and you would profit over the long run; a negative figure means the opposite. Everything here flows from comparing two numbers: how likely you think an outcome is, and how likely the odds say it is.

The arithmetic is short. Multiply your estimated probability by the decimal odds, then subtract one. A 55 percent chance at odds of 2.00 gives 0.55 times 2.00 minus one, which is 0.10, a positive expected value of ten percent per unit staked. Flip the estimate to 45 percent and the same price returns minus ten percent, a bet to avoid. The calculator above runs that calculation the instant you change either input, and colours the result so a losing price is obvious at a glance.

How the calculator works

The value bet calculator asks for two things: your honest estimate of the true win probability, and the decimal odds on offer. From those it derives the implied probability behind the price, compares it with your estimate, and reports the expected value. When your number beats the market\'s, the price carries an edge; when it does not, it does not. That is the whole of expected value betting, stripped of jargon.

It then goes one step further and sizes a stake, because knowing a bet is good is only half the decision. The bankroll field and the Kelly fraction selector turn the edge into a concrete amount, so the answer is not just "yes, this has value" but "and here is a stake that respects how uncertain you are". If you want to sanity-check the market price first, the odds converter turns any reference price into an implied probability, and the parlay calculator shows what happens when several value bets are combined.

The Kelly criterion, and why to use a fraction of it

The Kelly criterion answers a question expected value alone cannot: not whether to bet, but how much. It sets the stake that grows a bankroll fastest over the long run, given your edge and the odds. The formula divides the edge by the net odds: your probability times the net odds, minus the losing probability, all divided by the net odds. At 2.00 with a genuine 55 percent chance, the Kelly criterion calls for ten percent of the bankroll. When the edge disappears the formula turns negative, which simply means do not bet, so the calculator clamps any negative Kelly stake to zero.

The catch is that the Kelly criterion assumes your probability is exact. In the real world it is an estimate with error bars, and full Kelly punishes that error brutally, over-staking and whipping the bankroll up and down. Betting a fraction of the Kelly stake, most commonly a half or a quarter, keeps the large majority of the long-run growth while cutting the swings dramatically. That is why quarter Kelly is the default in the tool, and why the recommended stake it shows is deliberately smaller than the full figure beside it.

What full-Kelly variance actually looks like

The reason quarter Kelly is the calculator's default is not caution for its own sake; it is that the growth you give up is small and the swings you avoid are large. Growth near the Kelly peak is flat, so backing off the full stake barely dents it. Half Kelly captures about three-quarters of the maximum long-run growth rate, and quarter Kelly keeps a little under half of it, while the size of every swing scales straight down with the stake: quarter Kelly bets a quarter of the amount, so it rides a quarter of the volatility.

The default inputs make the trade concrete. A 55 percent chance at odds of 2.00 calls for a full Kelly stake of ten percent of the bankroll. On a 1,000 bankroll that is 100 a bet at full Kelly, 50 at half and 25 at quarter. Now picture a run of five straight losses, which is ordinary in a high-variance sport. At full Kelly, staking ten percent of a shrinking bankroll each time, the bank falls to 590, a drawdown of 41 percent. At quarter Kelly, staking 2.5 percent, the same five losses leave 881, a drawdown of just under 12 percent. Same edge, same bets, and one path is three times as brutal as the other.

Full Kelly also assumes your probability is exactly right, which no estimate ever is, and it punishes the error twice: once for the mistake and again through the outsized stake it sets on it. Betting a fraction absorbs that error rather than amplifying it, which is why the recommended figure in the tool is deliberately smaller than the full Kelly number beside it. The estimate feeding the whole calculation is only ever as good as its source, so sizing down a thin edge and testing the market price first with the odds converter is the honest way to use expected value. Value here is measured on price, the same basis the value scoreboard uses to grade every book.

Worked examples: reading value and stake together

The relationship between probability, price and stake is easiest to see across a spread of cases. The table below runs six of them through the same formulas the calculator uses on live inputs: the expected value from your probability times the decimal price minus one, the full Kelly stake as a fraction of bankroll, and the quarter-Kelly figure the tool recommends by default.

Two patterns stand out. A positive EV always produces a positive Kelly stake, and the quarter-Kelly figure is a quarter of the full one, deliberately smaller to absorb error in your estimate. The row where a 30 percent chance is offered at 3.00 shows the other side: the expected value is negative, so the Kelly stake is clamped to zero. There is no such thing as a small stake on a bad bet in this framework; when the edge disappears, the correct stake is nothing. That single rule protects a bankroll more than any staking trick, which is why the calculator refuses to suggest a stake it cannot justify.

The rows also show why price matters as much as probability. Compare the 40 percent chance at 2.80 with the same 40 percent chance if a tighter book offered 3.00 instead: the higher price lifts the expected value and, with it, the recommended stake, without your view of the outcome changing at all. That is the practical link between expected value betting and the value rankings. A book that shades its prices less leaves more of the edge on the table for the bettor, so where you place a value bet can matter as much as which bet you place.

Common mistakes with the calculator

The first and most damaging mistake is entering the price's own implied probability as your estimate. Do that and your number matches the market's exactly, the EV lands on zero, and the tool correctly reports no edge. Your probability has to come from somewhere other than the odds you are testing, or the whole exercise is circular. The second mistake is backing a thin edge from a shaky estimate; a two percent expected value built on a guess is not really value at all, and the honest response is to size down or pass.

The third mistake is staking full Kelly because the growth looks fastest on paper. Full Kelly assumes a perfect estimate and punishes any error with violent swings, so a fraction is almost always the wiser choice. The last mistake is reading positive EV as a promise about the next bet. Expected value is a long-run average across many wagers, not a forecast of a single result, so a positive-value bet can and often will lose. To sanity-check the market price before you start, the odds converter turns any reference price into an implied probability, and the value scoreboard shows which books price closest to the true chance in the first place.

Model probabilities are estimates, not certainties

No calculator can tell you the true probability of a result; only you can estimate it, and the estimate is where all the risk lives. If the number you feed in is optimistic, every figure downstream is optimistic too, expected value and stake alike. The honest way to use a value bet calculator is to treat a small edge with suspicion, to size down when unsure, and to accept that a positive expected value is a reason to consider a bet rather than a guarantee of anything. Value here is measured on price, the same way the value rankings grade every book, and the bet types guide covers the markets these bets are placed in.