Definitions · with real numbers
Over/under 2.5 goals and both teams to score, explained
Two of the most common football statistics, defined precisely, computed from a scoreline model rather than a hunch, and measured across all 104 matches of the 2026 World Cup.
Both markets sit on every match screen in the app.
Over 2.5 goals means three or more
Total goals by both teams, over the ninety minutes. Over 2.5 means three or more. Under 2.5 means zero, one or two. So 1:1 is under, 2:1 is over, and 0:0 is very much under.
Why the line is a half goal
Nobody scores half a goal, and that is the entire point. If the line were a whole number, say two, a 2:0 match would land exactly on it and there would be no answer to the question. Half goal lines make the question binary: either the total clears 2.5 or it does not. The same logic gives over 1.5, over 3.5 and so on.
Both teams to score means each side gets at least one
Written BTTS. It says nothing about who wins. A 1:1 draw qualifies, a 3:2 win qualifies, and a 4:0 win does not, even though it produced four goals. This is why BTTS and over 2.5 disagree more often than people expect: a heavy one-sided win is high on goals and low on BTTS at the same time.
How both are computed here
Neither is estimated on its own. A Poisson goal model with a Dixon-Coles correction builds one table of every scoreline from 0:0 to 8:8, and both markets are then simply sums of cells in that table.
Because they come off one table, they cannot contradict each other or the win, draw and win split. How that table is built is set out in the Poisson and Dixon-Coles model.
Expected goals here is not shot-based xG
This deserves a warning, because the same two letters mean two different things in football writing.
The expected goals figure shown next to each side in the app is the scoring rate the model solved for, backed out of the betting market before the match. It is a forecast made from prices. The expected goals you see in a post-match report is a different quantity: it is computed from the shots that were actually taken, weighted by how likely each one was to go in. One looks forward from prices, the other looks back from events. Neither is wrong, but comparing them directly is a mistake.
What these markets looked like across a whole tournament
Across all 104 matches of the 2026 World Cup, the market consensus produced this spread.
| Market | Average | Median | Lowest | Highest |
|---|---|---|---|---|
| Over 2.5 goals | 50.5% | 49.2% | 17.9% | 79.2% |
| Both teams to score | 43.6% | 45.5% | 2.3% | 65.2% |
| Draw after 90 minutes | 22.2% | 24.1% | 5.8% | 34.1% |
| Likeliest scoreline | 14.3% | 13.6% | 9.4% | 32.5% |
A few things fall out of that table.
- Over 2.5 is close to a coin flip on average. The mean was 50.5% across the tournament, and the average total scoring rate the model solved for was about 2.72 goals per match. A football match sits almost exactly on the 2.5 line, which is why that line became the standard one.
- The range is enormous. From 17.9% to 79.2%. The lowest belonged to a tie the market expected to be strangled, the highest to Germany against Curaçao on 14 June, where the model solved for a home scoring rate of 3.78 goals.
- BTTS collapses in a mismatch. The floor was 2.3%, in matches such as Mexico against South Africa on the opening day, where the away scoring rate came out at 0.02. Plenty of goals were expected, but only from one end.
- The best available scoreline guess is weak. The likeliest exact scoreline averaged 14.3% and only cleared twenty percent in 8 of the 104 matches.
The five lowest and five highest scoring matches by market view
| Match | Home | Draw | Away | Likeliest score | Over 2.5 | BTTS |
|---|---|---|---|---|---|---|
| Canada v Morocco4 Jul, 17:00 UTC | 1.3% | 25.6% | 73.1% | 0:1 32.5% | 17.9% | 4.3% |
| Belgium v Egypt15 Jun, 19:00 UTC | 16.8% | 33.2% | 50% | 0:1 19.8% | 25.1% | 30.5% |
| Portugal v Spain6 Jul, 19:00 UTC | 23.5% | 34.1% | 42.4% | 0:0 17.4% | 27.2% | 35.3% |
| Algeria v Austria28 Jun, 02:00 UTC | 26.8% | 33.6% | 39.6% | 0:0 16.1% | 29.6% | 38.1% |
| Paraguay v Australia26 Jun, 02:00 UTC | 33% | 33.2% | 33.8% | 0:0 14.8% | 31.9% | 40.7% |
| Match | Home | Draw | Away | Likeliest score | Over 2.5 | BTTS |
|---|---|---|---|---|---|---|
| Germany v Curaçao14 Jun, 17:00 UTC | 92.6% | 5.8% | 1.6% | 3:0 13.1% | 79.2% | 36.6% |
| New Zealand v Belgium27 Jun, 03:00 UTC | 6.9% | 12.5% | 80.6% | 0:3 9.4% | 75.8% | 56.1% |
| Spain v Cape Verde15 Jun, 16:00 UTC | 89.7% | 8.1% | 2.3% | 3:0 14.2% | 71.8% | 35.1% |
| Brazil v Haiti20 Jun, 00:30 UTC | 86.5% | 10% | 3.5% | 3:0 12.7% | 71.4% | 41.7% |
| France v Iraq22 Jun, 21:00 UTC | 89.3% | 8.3% | 2.4% | 3:0 14.1% | 71.4% | 35.6% |
Reading the bar: home windrawaway winPercentages are the fair market consensus after the bookmaker margin is removed.
Related