Over/Under Goals Calculator
Turn two expected-goals figures into the probability of every total, the chance of over or under each line from 0.5 to 5.5, and the fair price with no bookmaker margin. Add the prices you have been offered to see the margin and the expected value of each side.
In one line: expected total = home xG + away xG; over 2.5 = 1 − P(0) − P(1) − P(2) from a Poisson distribution. With 1.6 + 1.1 = 2.7 goals expected, over 2.5 is 50.64%, a fair price of 1.97.
Expected goals for one match are an estimate: from a model, from recent xG averages or from prices. Decimal odds; convert fractional or American prices with the odds converter. Settlement is on 90 minutes plus stoppage time unless the market says otherwise.
Every line at 2.70 expected goals
| Line | Over | Fair over | Under | Fair under |
|---|---|---|---|---|
| 0.5 | 93.28% | 1.07 | 6.72% | 14.88 |
| 1.5 | 75.13% | 1.33 | 24.87% | 4.02 |
| 2.5 | 50.64% | 1.97 | 49.36% | 2.03 |
| 3.5 | 28.59% | 3.50 | 71.41% | 1.40 |
| 4.5 | 13.71% | 7.29 | 86.29% | 1.16 |
| 5.5 | 5.67% | 17.63 | 94.33% | 1.06 |
Whole-number lines: the push
| Line | Over wins | Push (stake back) | Under wins |
|---|---|---|---|
| 1.0 | 75.13% | 18.15% | 6.72% |
| 2.0 | 50.64% | 24.50% | 24.87% |
| 3.0 | 28.59% | 22.05% | 49.36% |
| 4.0 | 13.71% | 14.88% | 71.41% |
Result
How the Calculator Prices a Goals Line
The Poisson distribution gives the chance of exactly k events when they happen independently at an average rate λ: P(k) = e−λ λk / k!. The sum of two independent Poisson counts is itself Poisson, so the total goals in a match is modelled with λ = home expected goals + away expected goals. Over a half-goal line is one minus the chance of every total below it; under is the rest.
Worked example: 1.6 against 1.1
λ = 2.7. P(0) = e−2.7= 6.72%; P(1) = 2.7 × 6.72% = 18.15%; P(2) = 2.7/2 × 18.15% = 24.50%. Two goals or fewer is 49.36%, so over 2.5 is 50.64% and the fair prices are 1.97 over and 2.03 under. If both sides are offered at 1.90, the book adds up to 105.26% (a 5.26% margin), and the over at 1.90 has an expected value of 50.64% × 1.90 − 1 = −3.8% per stake. At 2.10 it would be +6.3%: that is the price at which your estimate and the bookmaker's disagree enough to matter.
Over 2.5 by expected total
The same function, run for the totals you see most often. A tenth of a goal in the estimate moves the over 2.5 probability by about two and a half points, which is why the input matters more than the formula.
| Expected goals | Over 2.5 | Fair over | Fair under |
|---|---|---|---|
| 1.8 | 26.9% | 3.71 | 1.37 |
| 2.0 | 32.3% | 3.09 | 1.48 |
| 2.2 | 37.7% | 2.65 | 1.61 |
| 2.4 | 43.0% | 2.32 | 1.76 |
| 2.6 | 48.2% | 2.08 | 1.93 |
| 2.8 | 53.1% | 1.88 | 2.13 |
| 3.0 | 57.7% | 1.73 | 2.36 |
| 3.2 | 62.0% | 1.61 | 2.63 |
| 3.4 | 66.0% | 1.51 | 2.94 |
Where the model is weak
Real matches are not perfectly independent: a team that leads often slows down, a team that trails pushes, and red cards change the rate. Studies of league data since Dixon and Coles (1997) find that a plain Poisson model slightly underestimates low-scoring draws. The goal estimate for one match is also uncertain by several tenths. Use the fair price as a range and compare it with the bookmaker margin before reading anything into a small edge. The same goal distribution, split by team, is what prices correct score markets.
Settlement rules that change the answer
Goals markets are normally settled on 90 minutes plus stoppage time; extra time and shoot-outs do not count unless the market says “including extra time”. Own goals count; a goal disallowed by VAR does not. If a match is abandoned, a line already decided (over 2.5 with three goals scored) usually stands and the rest are void. The details are in our extra time and penalties rules and abandoned match rules.
Related Calculators
The same two expected-goals figures also price whether both teams score. Two matches with an identical 2.6-goal total can carry very different odds for that market, depending on how the goals are split, which is what the BTTS calculatorshows.
Every over/under price is a sum of individual scorelines. The correct score calculatorlays out that grid from 0-0 to 5-5, including the "any other score" bucket.
If you prefer betting on the margin rather than the total, the Asian handicap calculatorsettles half and quarter lines, including the half-win and half-loss outcomes.
Frequently Asked Questions
How do you work out the probability of over 2.5 goals?
Add the two teams' expected goals to get the expected total, then treat the number of goals as a Poisson distribution with that mean. Over 2.5 is one minus the chance of 0, 1 or 2 goals. With 1.6 and 1.1 expected goals the total is 2.7, the chance of two goals or fewer is 49.36%, so over 2.5 is 50.64% and the fair price is 1.97.
What are fair odds for over 2.5 goals?
Fair odds are 1 divided by the probability, with no bookmaker margin. They depend on the expected total: about 3.09 when two goals are expected, 2.08 at 2.6, 1.73 at 3.0 and 1.51 at 3.4. A price above the fair figure means the bet has positive expected value if, and only if, your goal estimate is right.
What happens on over/under 2.0 or 3.0 goals if the match ends exactly on the line?
On a whole-number line such as over 2.0 or under 3.0, a total exactly on the line is a push and the stake is returned. Only the half-goal lines (0.5, 1.5, 2.5 and so on) always have a winner. The calculator shows the push probability for the whole lines; quarter lines such as 2.25 split the stake between 2.0 and 2.5.
Do extra time and penalties count in over/under goals bets?
Usually not. Total goals markets are normally settled on 90 minutes plus stoppage time, so goals in extra time do not count and a penalty shoot-out never adds goals. Check the market name and the operator's football rules: some cup markets are offered including extra time and say so.
How accurate is a Poisson model for football goals?
It is a good first estimate, not a forecast. It assumes goals arrive independently at a steady rate, which slightly underestimates 0-0 and 1-1 draws and ignores how game state changes the scoring rate. The bigger error is usually the input: expected goals for a single match carry a lot of uncertainty, so treat the fair price as a range, not a number.
A probability model describes uncertainty; it does not remove it. A 50% bet loses half the time, sometimes several times in a row. 18+. If gambling stops being fun, see our responsible gambling resources.