The model projects 3.58 goals for FSV Mainz 05 and calculates a 94% chance they score at least once. It is featured because it ranks among the three highest estimated probabilities in the eligible eight-pick shortlist; this is not a measure of betting value.
★ #294%
Over 1.5 goals
The model projects 4.95 combined goals. Its estimated probability of at least two goals is 94%. It is featured because it ranks among the three highest estimated probabilities in the eligible eight-pick shortlist; this is not a measure of betting value.
★ #382%
Over 2.5 goals
The combined goal projection is 4.95. The Poisson distribution assigns 82% to three or more total goals. It is featured because it ranks among the three highest estimated probabilities in the eligible eight-pick shortlist; this is not a measure of betting value.
More selections
4
FSV Mainz 05 or Draw
81%
5
Werder Bremen to score
78%
6
FSV Mainz 05 2+ goals
78%
7
Both teams to score
74%
8
FSV Mainz 05 to win
61%
Featured picks may overlap (for example a team scoring and over 1.5 goals). A high probability does not mean good betting value.
League baseline. Expected goals start from this competition’s real scoring averages at home (2.28) and away (1.41) per match.
Attack and defence ratings. Each team is rated against that baseline from the season table and its last 20 matches, with older matches fading (half-life ~90 days).
Opponent adjustment. Goals against weak defences count less, goals against strong ones count more.
Regression to the mean. Small samples are pulled toward the league average so one big result does not distort the estimate.
Poisson + Dixon–Coles. Expected goals generate a probability for every scoreline; a low-score correction fixes the known underestimation of draws and 1-0 / 0-1 / 1-1.
Markets. 1X2, totals and both-teams-to-score are read from that scoreline grid.
1.70 / 0.84FSV Mainz 05 attack / defence
1.14 / 1.08Werder Bremen attack / defence
Ratings are relative to the league average (1.00). Attack above 1 scores more than average; defence above 1 concedes more.
A statistical model, not a trained AI. It does not know about injuries, lineups, motivation or odds. Use it as one more data point.