Betting markets
Over/Under Goals and Both Teams to Score: How Goals Markets Are Priced
How over/under 1.5, 2.5 and 3.5 goals lines and both teams to score are priced, with a worked Poisson example and a short look at Asian quarter lines.
A match is expected to produce 2.6 goals. That sounds like an over 2.5 game, and yet a simple model gives under 2.5 the slight edge: about 52% against 48%. The reason is that goals arrive in whole numbers, and a small expected figure spreads its probability across 0, 1 and 2 more than intuition suggests.
Goals markets ignore who wins. They ask how many goals there will be, or whether both sides will find the net, and that makes them a separate exercise from the match result markets covered in 1X2, double chance and draw no bet.
How over/under lines work
An over/under line sets a total and asks whether the match will finish above or below it. The common ones are 0.5, 1.5, 2.5, 3.5 and 4.5. The half goal is deliberate: no match can finish on 2.5 goals, so every bet either wins or loses.
- Over 1.5 needs two or more goals.
- Over 2.5 needs three or more.
- Over 3.5 needs four or more.
The under is the mirror image. Only goals in normal time plus stoppages count, unless a market states otherwise. Own goals count like any other goal.
The 2.5 line is the most popular because in many leagues it splits matches into two roughly equal halves. That balance is also why its prices usually sit near 1.90 on each side: two near-even outcomes, with a margin shaved off both.
What drives the number of goals
The core input is how many goals each team is expected to score in this particular match. That depends on the attacking strength of one side against the defensive strength of the other, adjusted for home advantage. A side that scores freely at home against a team that concedes plenty away points one way. Two compact, defensive sides point the other.
- Attacking strength of one side against the defensive strength of the other
- Home advantage
- Team news, above all a missing first-choice striker or goalkeeper
- Weather and the state of the pitch
- Motivation, such as a side that only needs a draw on the final day
Other things move it. Team news matters most, especially a missing first-choice striker or goalkeeper. Weather plays a part: heavy rain or a waterlogged pitch tends to slow the game down. Motivation counts as well. A side that only needs a draw on the final day of a season often plays very differently from one that must win.
Expected goals (xG), now standard in data from providers such as Opta, helps separate how well a team is creating chances from how many it happens to have scored. A team scoring far more than its xG over a few weeks is often riding luck in finishing. Rates like that tend to drift back towards the underlying chance quality, which is why many analysts trust a longer xG sample more than a recent goals tally.
The Poisson intuition, with numbers
The classic way to turn an expected goals figure into probabilities is the Poisson distribution, the idea at the heart of Maher’s 1982 model of football scores. It describes how often a rare event happens a given number of times when you know its average rate. The guide to football prediction models covers it in more depth; here is the short version.
Take a hypothetical match where the home side is expected to score 1.5 goals and the away side 1.1, so 2.6 in total. Under a simple Poisson model the chance of exactly k goals is e^(−2.6) × 2.6^k ÷ k!. That gives:
| Total goals | Probability | Running total |
|---|---|---|
| 0 | 7.4% | 7.4% |
| 1 | 19.3% | 26.7% |
| 2 | 25.1% | 51.8% |
| 3 | 21.8% | 73.6% |
| 4 or more | 26.4% | 100.0% |
Reading the lines from that table:
- Over 1.5: 100% − 26.7% = 73.3%, a fair price of about 1.36.
- Over 2.5: 100% − 51.8% = 48.2%, fair price about 2.08. Under 2.5 is 51.8%, about 1.93.
- Over 3.5: 26.4%, fair price about 3.79.
So a match with 2.6 expected goals is close to a coin flip on the 2.5 line, tilted very slightly towards the under. If the posted prices were 1.90 on both sides, the over would be priced a little short by this model and the under close to fair, once the margin is taken into account. The conversion from price to probability is the same one set out in how to read football odds.
A real model would not stop here. The basic Poisson approach treats the two teams’ goals as independent and tends to misjudge low scores like 0–0 and 1–1. Dixon and Coles proposed a correction for exactly that in 1997, along with a way to weight recent matches more heavily. It is a small adjustment in the numbers and a large one in reputation: their paper is still the reference point for goal models more than a quarter of a century later.
Both teams to score
Both teams to score (BTTS, sometimes listed as GG/NG) wins if each side scores at least once. The total does not matter: 1–1 wins, 4–0 loses.
Using the same match, the chance that the home side scores at least once is 1 − e^(−1.5) = 77.7%. For the away side it is 1 − e^(−1.1) = 66.7%. If you treat the two as independent, the chance that both happen is 77.7% × 66.7% = 51.8%, a fair price of about 1.93.
That this matches the under 2.5 figure here is a coincidence of these particular inputs, not a rule. Change the split and the two markets separate quickly. A match with 2.6 expected goals made up of 2.2 for one side and 0.4 for the other has exactly the same over/under probabilities, but its BTTS chance falls to about 29%, because the weaker side often fails to score at all.
That is the main lesson of BTTS: it rewards balance, not volume.
Asian quarter lines, briefly
Asian goal lines add whole numbers and quarter numbers to the familiar half lines. A whole line like over 2.0 refunds the stake if the match ends with exactly two goals. A quarter line like over 2.25 splits the stake into two halves, one on 2.0 and one on 2.5:
- Three or more goals: both halves win.
- Exactly two goals: one half refunded, one half lost.
- One or none: both halves lose.
Over 2.75 works the same way, splitting between 2.5 and 3.0. These lines let the market price fine differences in expectation, and they sit naturally on top of the Poisson table above. Short definitions of each line type are collected in the betting glossary.
- Half line (over 2.5)
- No match can land on it, so the bet wins or loses.
- Whole line (over 2.0)
- Stake refunded if the match ends with exactly two goals.
- Quarter line (over 2.25)
- Stake split in half between 2.0 and 2.5.
- Quarter line (over 2.75)
- Stake split in half between 2.5 and 3.0.
Reading a goals price sensibly
Start from the two teams’ expected goals, not from last week’s scoreline. Build a rough distribution, even a mental one: how likely is a 0–0, how often do these sides produce four goals? Then convert the posted prices into percentages and compare.
You will often find the market already agrees with you, which is useful information too. The interesting cases are the ones where it does not, and the honest question then is who knows more: you, or a price that has absorbed every piece of team news and every stake placed so far.
One detail worth checking every time: a team that has just changed manager. New coaches often change a side’s tempo and pressing within weeks, and goal rates from the previous regime can be a poor guide to the next month.
FAQ
Your questions, answered.
What does over 2.5 goals mean?
It wins if the match produces three or more goals in total, counting both teams, after normal time plus stoppages. A 2–1, a 3–0 or a 4–4 all count. Two goals or fewer and it loses. The half goal exists only so the bet cannot end exactly on the line, which removes any refund.
Is under 2.5 or over 2.5 more common?
It depends on the league and the season. Many top European leagues average somewhere between about 2.5 and 3 goals a game, which puts the two sides of the line close together. Lower-scoring divisions lean towards the under. A league average is only a starting point; the two teams on the day matter much more.
Can both teams to score win in a 1–1 draw?
Yes. BTTS only asks whether each side scored at least once, so 1–1 wins, as does 3–2. A 4–0 loses, even though it is a high-scoring match. That is why BTTS and over 2.5 often move together but are not the same question, and why a one-sided mismatch can be a poor fit for BTTS despite plenty of goals.
What is an over 2.25 Asian line?
It splits the stake into two halves, one on over 2.0 and one on over 2.5. With three or more goals, both halves win. With exactly two goals, the 2.0 half is refunded and the 2.5 half loses, so half the stake is lost. With fewer than two, both halves lose.
How do bookmakers estimate the number of goals in a match?
Usually with models that turn each team's attacking and defensive strength into an expected goals figure for the match, then into a distribution of scorelines. Poisson-based models are the classic starting point. Team news, weather and the market's own reaction to early money then adjust the price before kick-off.
Does a high expected goals number guarantee over 2.5?
No. Even with 3.5 expected goals, a Poisson model gives roughly a 32% chance of two goals or fewer. Goals are rare, random events, and a single save, a missed penalty or a red card can change a match entirely. Expected goals describes a likely range, never a fixed result.