Expected value is the single most important concept in profitable sports betting. Every decision — whether to place a bet, which market to use, how much to stake — flows from whether the bet is positive or negative EV. Here is what it means, how to calculate it, and how to use it.
Expected value (EV) is the average outcome you can expect per bet if you placed the same bet an infinite number of times. It is borrowed from probability theory and is the foundation of every successful betting strategy.
A positive EV (+EV) bet means that, on average, you profit. A negative EV (−EV) bet means you lose money on average. Bookmakers set their margins to ensure virtually every bet on their platform is −EV for the customer. Your job as a bettor is to find the exceptions.
Key insight: A winning bet is not necessarily a good bet, and a losing bet is not necessarily a bad one. What matters is whether the odds offered exceed your true probability estimate for the outcome. That is EV thinking.
The expected value of a bet is calculated as follows:
EV = (Probability of Win × Profit) − (Probability of Loss × Stake)
Where Profit = Stake × (Decimal Odds − 1)
You can also express this per unit staked (for standardisation across different bet sizes):
EV per unit = (p × (decimal odds − 1)) − (1 − p)
Where p = your estimated probability of winning (as a decimal)
Arsenal are priced at 2.20 (decimal) to win at home. After analysis using xG data and Poisson modelling, you estimate their true win probability at 55% (0.55).
Decimal odds
2.20
Implied probability
45.5%
1 ÷ 2.20
Your estimate
55%
EV per unit = (0.55 × (2.20 − 1)) − (0.45 × 1)
EV per unit = (0.55 × 1.20) − 0.45
EV per unit = 0.66 − 0.45
EV = +0.21 per unit staked (+21%)
This means for every £10 staked, you expect to profit £2.10 on average over a large number of identical bets. The bet is strongly +EV — assuming your probability estimate is accurate.
Many bettors judge their ability by their win rate or recent profit. Both are misleading in small samples. EV is what separates skill from luck.
If you back a 1.05 shot at 1.05 odds, you win 95% of the time but your long-run return is −4.76% per bet. Frequent wins ≠ profitable strategy.
If you back a correct outcome at 3.00 odds and it loses this time, but your true probability was 40%, you made a good decision. EV was positive — variance beat you temporarily.
A 10% ROI over 20 bets is almost entirely noise. Over 1,000 bets at the same EV, ROI starts to reflect genuine edge. EV thinking removes the short-term noise.
The bookmaker margin ensures that the average implied probability across all outcomes exceeds 100%. Beating this requires finding bets where your model outperforms the market.
Finding +EV bets requires a probability estimate that differs from the bookmaker's implied probability. Here is a systematic process.
Use xG data, team form, head-to-head records, and a Poisson model to estimate the true probability of each outcome. Do this before looking at the odds.
→ Poisson CalculatorTake the bookmaker's odds and convert them to an implied probability. For decimal odds: Implied P = 1 ÷ Odds. This tells you what probability the bookmaker is pricing in.
→ Implied Probability CalculatorIf your estimated probability exceeds the implied probability, the bet is +EV. The bigger the gap, the higher the expected edge. Use the formula above to calculate the exact EV.
Once you have identified a +EV bet, the Kelly formula tells you the optimal fraction of your bankroll to stake based on your estimated edge and the odds.
→ Kelly CalculatorClosing line value (CLV) — whether your odds were better than the closing price — is the best real-time proxy for whether you are finding genuine +EV.
→ Closing Line Value GuideConfusing a good result with a good bet
Fix: Judge every bet by whether it was +EV at the time you placed it — not by whether it won. A losing +EV bet is still a correct decision.
Using the bookmaker's odds to estimate probability
Fix: The implied probability in bookmaker odds already includes a margin. You must first strip the margin to get the "true" market probability — then compare it to your estimate.
Building EV estimates from too-small samples
Fix: A team winning 3 of their last 4 is not a reliable probability estimate. Use full-season xG data and larger league samples for stable base rates.
Chasing losses after −EV outcomes
Fix: Even +EV bets lose sometimes. Chasing is a −EV response to normal variance — it compounds your losses with worse decisions made under emotional pressure.
Convert any odds format to probability and strip the bookmaker margin.
Model goal probabilities for any match using expected goal averages.
Calculate the optimal stake size given your EV and odds.
Strip the margin from a market to see the no-vig implied probabilities.
Calculate the bookmaker's edge on any market.
The full value betting process — from model to staking plan.
Expected value (EV) is the average return you can expect per unit staked over a large number of identical bets. A positive EV (+EV) bet returns profit on average; a negative EV (-EV) bet loses money on average.
Yes. EV describes the long-run average, not the result of any individual bet. A +EV bet will still lose sometimes — that is normal variance. This is why large sample sizes matter.
You need your own probability estimate for an outcome. If your probability is higher than the implied probability in the odds, the bet is +EV. Tools like KiqIQ's Poisson calculator and implied probability converter help you build those estimates.
