BundesPredict
Win, draw and loss probabilities for every Bundesliga fixture, from a model refit every day. When news breaks, tell the agent what changed and watch the numbers move.
3 free chats a day · no card neededBring the news the model hasn't seen
Lineups, injuries, a storm over the stadium. The agent checks what you said against squad data and the forecast, then reruns the model with small, capped changes.
- You say what changedIn plain words, about one fixture.
- The agent checks itSquad, form and weather lookups, shown as it works.
- The odds move, with receiptsModel odds next to the new ones, and every adjustment listed.
DortmundvHoffenheim
Guirassy is out for Dortmund, how does that change Dortmund v Hoffenheim?
3 agent steps
- Dortmund 61%, draw 23%, Hoffenheim 16% predict_matchcompleted
- Serhou Guirassy, striker, high importance lookup_playercompleted
- Home win 61% to 53% with 1 adjustment predict_match_with_contextcompleted
xG 1.97 – 0.92 → 1.62 – 0.92
The base model has Dortmund at 61% to beat Hoffenheim. Guirassy is their main goal source and a penalty taker, so I took 0.35 expected goals off the Dortmund attack. That moves the home win to 53%, with the draw and away win taking the difference. The rest of the squad is unchanged.
The model does the maths. The agent does the words.
A fitted score grid
A Dixon-Coles model rates every team's attack and defence from Bundesliga results since 2020/21 and is refit daily. Each fixture becomes one grid of scoreline chances, so win, draw, loss and goal markets always agree.
Small, capped nudges
The agent reads your news, checks it with squad, form and weather tools, and asks for expected-goals changes. It has no way to set a probability, and each change is capped at ±0.60 xG.
Every answer shows its work
You see the model's probabilities beside the adjusted ones, what each adjustment asked for, and where the cap cut it down. Every prediction is stored with the exact model run behind it.
All 18 clubs, every matchday
Close to the market, and honest about it
Every number below is from predictions the model made before seeing the result. Lower ranked probability score (RPS) is better.
Free to look. Pro to talk all day.
Free
- Probabilities for every fixture
- 3 agent chats a day
- Saved chat history
- Public track record
Pro
- Everything in Free
- 100 agent chats a day (fair use)
- Matchday email every Thursday
- Cancel any time from your account