Betting maths guide
What Is Expected Value in Betting?
Expected value, or EV, estimates the average mathematical result of a bet when possible outcomes are weighted by their probabilities. It can explain whether a price looks favourable under a stated probability estimate, but it cannot predict the result of the next bet.
The core idea
A bet has at least two possible financial outcomes: a profit if it wins and a loss if it loses. Expected value combines those outcomes into one probability-weighted average.
Profit if win = stake × (decimal odds − 1)
EV = (win probability × profit if win) − (loss probability × stake)
EV ROI % = EV ÷ stake × 100
EV is therefore not a statement about what will happen once. It is an average implied by the model if the same probability and price could be repeated many times.
A £10 worked example
Suppose the stake is £10, the decimal odds are 2.50 and the estimated chance of winning is 45%.
| Step | Calculation | Result |
|---|---|---|
| Profit if win | £10 × (2.50 − 1) | £15.00 |
| Weighted win profit | 45% × £15 | £6.75 |
| Weighted loss | 55% × £10 | £5.50 |
| Expected value | £6.75 − £5.50 | +£1.25 |
| Expected ROI | £1.25 ÷ £10 | +12.50% |
The calculation is positive only because we supplied a 45% probability estimate. Change that estimate and the answer changes.
Break-even probability
The simplest benchmark for a decimal price is its implied probability. That is also the break-even probability before other costs or settlement complications are considered.
Break-even probability = 1 ÷ decimal odds
At odds of 2.50, the break-even probability is 40%. If your probability estimate is exactly 40%, the simplified EV is zero. Above 40%, the model is positive; below 40%, it is negative.
| Estimated probability | EV on £10 at 2.50 | Interpretation |
|---|---|---|
| 35% | -£1.25 | Negative under the estimate |
| 40% | £0.00 | Break-even |
| 45% | +£1.25 | Positive under the estimate |
| 50% | +£2.50 | More positive under the estimate |
The probability estimate is the hard part
The arithmetic is easy. Estimating the probability accurately is not.
Decimal odds tell you the probability implied by the price, not the objectively correct probability. A bettor can calculate positive EV simply by entering an over-optimistic probability. The calculator will faithfully perform the maths even when the assumption is poor.
That is why Betting Maths treats EV as a transparent model rather than a selection engine. The useful question is not merely “is the EV positive?” but also “where did the probability estimate come from, and how uncertain is it?”
Expected value versus profit
Profit is what actually happened on a settled bet. Expected value is what the model says the average result would be under its assumptions.
| Measure | What it answers |
|---|---|
| Actual profit/loss | What happened on this bet? |
| Expected value | What is the probability-weighted average result under the assumptions? |
| EV ROI | What is that expected result as a percentage of stake? |
| Break-even probability | What win probability would make the simplified EV equal zero at this price? |
A positive-EV bet can lose. A negative-EV bet can win. One result does not validate or invalidate the original probability estimate.
Why repeated bets still do not produce a smooth line
Expected value is a long-run average concept, but real sequences contain variance. Even when a probability model is correct, wins and losses can cluster.
For a 45% event, a sequence of ten trials can easily contain far fewer—or far more—than 4.5 wins. The average only becomes more informative as the number of genuinely comparable independent observations grows. In betting, that ideal is complicated because prices and underlying probabilities usually change from event to event.
EV and bookmaker margin
Bookmaker margin and expected value are related but different ideas. Overround describes how the quoted probabilities in a market add together. EV describes one price relative to a probability estimate.
A market with a 105% book does not mean every selection has exactly -5% EV. Margin can be distributed unevenly between outcomes. That is one reason a no-vig calculation is best treated as a benchmark rather than a direct measurement of true probability.
Common mistakes
- Treating implied probability as true probability. A price is market information, not proof.
- Confusing return with profit. Decimal-odds return includes the returned stake.
- Assuming positive EV guarantees a win. It does not.
- Ignoring settlement rules. Commission, dead heats, voids, deductions and each-way terms can change the actual cash result.
- Using an estimate with false precision. A model saying 43.7% is not necessarily more credible than one saying roughly 44%.
What observed market research can add
Betting Maths now publishes observed bookmaker-market studies alongside theoretical guides. For example, our Premier League opening-weekend research measured actual 1X2 book percentages across 10 matches rather than assuming a made-up margin.
That kind of research does not supply the “true probability” required for an EV calculation, but it helps show how real quoted prices and market margins behave.
When to use the calculator
Use the Expected Value Calculator when you already have a price and a probability estimate and want to see the mathematical consequence of combining them. Change the probability deliberately to see how sensitive the result is.
Do not use the calculator as evidence that your probability estimate is correct.
