19:30UPCOMING
Swansea University
Carmarthen TownOver 2.5 Goals
Solid72%
Over 2.5 goals · Wed 28 Oct
Over 2.5 goals predictions, 28 Oct
Every match where the model gives three or more goals a 65% chance or better, with the probability and the bookmaker price. Settled matches stay on the page with the score.
This day
9
picks
Last 90 days
66.0%
landed · 68.2% claimed · 1094 picks
Settled: 5 of 8 landed
DFB Pokal
Germany3 predictions
Cup
Poland1 prediction
UEFA Champions League Women
World3 predictions
The last 90 days on this page's rule: 1094 settled picks, 68.2% claimed on average, 66.0% landed. The page started on 27 Oct; before that, the record replays the same rule on the predictions we stored at the time, frozen at kick-off.
What over 2.5 goals means
An over 2.5 goals bet wins when the match has three or more goals, whoever scores them. 2-1, 3-0 and 2-2 all win; 1-1, 2-0 and 0-0 all lose.
How these predictions are made
A Poisson goal model estimates how many goals each side is likely to score from recent venue form, Elo ratings, rest days and, where we have the data, shots and expected goals. The chance of three or more goals comes from those two numbers, with a correction for low-scoring results, and then passes through a calibration layer fitted on a season the model never trained on. This page lists every match where that chance is 65% or more. The methodology page has the detail.
Where the model is weakest
Over 2.5 is the model's least reliable market: its confident over 2.5 calls have landed a few points below what they claimed. The record on this page shows that gap as it stands, updated every day, and the track record compares it with the other markets.
Frequently asked questions
- What's the difference between over 2.5 goals and both teams to score?
- Over 2.5 needs three goals from anyone; both teams to score needs at least one from each side. A 3-0 wins over 2.5 and loses BTTS, and a 1-1 wins BTTS and loses over 2.5. Our BTTS predictions have their own page.
- Why are there more over 2.5 picks on some days than others?
- The page lists every match that clears 65%, so a full weekend card produces more than a quiet midweek. There is no quota and no padding.
- Does extra time count?
- No. Bookmakers settle over 2.5 on the score after 90 minutes plus stoppage time, and that is what the model predicts.
- Why isn't a 70% pick a sure thing?
- Because 70% means it loses three times in ten. Across ten picks at that level, expect about three to lose. That is what the probability says, not the model failing.