Mathematical Football Predictions & Match Statistics
Find football matches through goal probabilities, team form and transparent pre-match forecasts. Explore eight daily insights, four matches in focus and your own saved scanner filters.
Daily statistical insights
Patterns from upcoming matches, supported by recorded results and model probabilities. Each insight shows the sample or scoring inputs behind it.
Data: Established sample17 home / 17 away venue games
1 · Home win
57.3%
X · Draw
23.6%
2 · Away win
19.1%
Over 2.5 goals
48.7%
Both teams to score
48.4%
Most likely score: 1-0(12.6%)
Why this forecast?
Fluminense score 1.82 and concede 1.00 goals per game at home, based on 17 matches.
Mirassol score 0.88 and concede 1.53 goals per game away from home, based on 17 matches.
Rowdie adjusts the observed scoring and conceding rates towards competition averages to reduce the influence of small samples. The resulting goal estimates are 1.73 home and 0.89 away.
Season in this competition · 35 home-team games / 35 away-team games. Competition baseline: 353 matches. Valid score records in the stored sample: home 100.0%, away 100.0%, competition 100.0%. Source completeness is not independently verified.
Rowdie adjusted model rowdie-shrunk-1 · Calculated .
Goal estimates use results, not shot-based xG. Data quality describes the sample, not the chance of winning.
Data: Established sample13 home / 13 away venue games
1 · Home win
58.2%
X · Draw
20.4%
2 · Away win
21.3%
Over 2.5 goals
65.8%
Both teams to score
62.7%
Most likely score: 2-1(9.7%)
Why this forecast?
Colo Colo score 2.62 and concede 0.85 goals per game at home, based on 13 matches.
U. Catolica score 1.69 and concede 1.54 goals per game away from home, based on 13 matches.
Rowdie adjusts the observed scoring and conceding rates towards competition averages to reduce the influence of small samples. The resulting goal estimates are 2.15 home and 1.24 away.
Season in this competition · 27 home-team games / 27 away-team games. Competition baseline: 216 matches. Valid score records in the stored sample: home 100.0%, away 100.0%, competition 100.0%. Source completeness is not independently verified.
Rowdie adjusted model rowdie-shrunk-1 · Calculated .
Goal estimates use results, not shot-based xG. Data quality describes the sample, not the chance of winning.
Data: Established sample13 home / 13 away venue games
1 · Home win
65.7%
X · Draw
20.8%
2 · Away win
13.4%
Over 2.5 goals
50.2%
Both teams to score
44.8%
Most likely score: 1-0(13.3%)
Why this forecast?
Palestino score 2.08 and concede 1.08 goals per game at home, based on 13 matches.
Union La Calera score 0.54 and concede 1.69 goals per game away from home, based on 13 matches.
Rowdie adjusts the observed scoring and conceding rates towards competition averages to reduce the influence of small samples. The resulting goal estimates are 1.94 home and 0.74 away.
Season in this competition · 27 home-team games / 27 away-team games. Competition baseline: 216 matches. Valid score records in the stored sample: home 100.0%, away 100.0%, competition 100.0%. Source completeness is not independently verified.
Rowdie adjusted model rowdie-shrunk-1 · Calculated .
Goal estimates use results, not shot-based xG. Data quality describes the sample, not the chance of winning.
Data: Established sample10 home / 10 away venue games
1 · Home win
55.3%
X · Draw
25.8%
2 · Away win
18.9%
Over 2.5 goals
40.6%
Both teams to score
42.3%
Most likely score: 1-0(15.2%)
Why this forecast?
Cobán Imperial score 1.70 and concede 1.30 goals per game at home, based on 10 matches.
Deportivo San Pedro score 0.40 and concede 1.40 goals per game away from home, based on 10 matches.
Rowdie adjusts the observed scoring and conceding rates towards competition averages to reduce the influence of small samples. The resulting goal estimates are 1.53 home and 0.78 away.
Season in this competition · 20 home-team games / 20 away-team games. Competition baseline: 122 matches. Valid score records in the stored sample: home 100.0%, away 100.0%, competition 100.0%. Source completeness is not independently verified.
Rowdie adjusted model rowdie-shrunk-1 · Calculated .
Goal estimates use results, not shot-based xG. Data quality describes the sample, not the chance of winning.
Legacy AnWP model. These value estimates use the original model. A positive model edge compares estimated probability with the available price. The displayed bookmaker average currently uses one source, 1xBet.
Rowdie adjusted model uses a Poisson goal model with scoring and conceding rates stabilised towards the competition averages. Small samples receive more adjustment; observed form remains separate from forecast probabilities.
Each forecast shows the number of team and venue matches, the competition sample and the calculation time. Established sample means at least 10 results per team, 5 at the relevant venue and 30 in the competition, with valid scores in the stored sample. It describes the available data, not a guarantee of model accuracy.
The initial stabilisation uses the weight of eight prior matches; it is an explicit starting parameter, not a claim of measured calibration. Forecasts are recorded before kickoff and assessed in Prediction Results. The archive starts when recording is enabled; earlier predictions are not reconstructed.
Open Match Scanner to combine filters. Value Bets and the native AnWP panels use separately labelled legacy estimates; they can differ from the adjusted Rowdie forecast. Goal estimates are based on results, not shot-based xG.