How a pick is made
No tips, no gut feel. Each match runs through the same six steps, every day.
- 01
Team strength
Every club carries an attack index and a defence index measured against a league average of 1.00. Higher attack means more goals scored; higher defence means more goals conceded.
- 02
Expected goals
Home expected goals = 1.45 × home attack × away defence. Away expected goals = 1.15 × away attack × home defence. The gap between 1.45 and 1.15 is the home-advantage edge.
- 03
Recent form
Points from the last five results shift each side's expected goals by up to about 12%. A side on five wins plays above its baseline; a side on five losses below it.
- 04
Goal simulation
A Poisson distribution prices every scoreline from 0-0 to 7-7. Adding the cells gives home, draw and away probabilities, over 2.5 goals and both teams to score.
- 05
Final prediction
The highest-probability result becomes the pick, with its probability as the confidence. If home and away are nearly level and a draw is plausible, the draw is called.
- 06
Grading
Each prediction is saved before kickoff. When the match finishes it is graded against the real score and counted in the record, hits and misses alike.
The daily run
At 00:00 UTC the engine creates the next fixtures, grades yesterday's matches, refreshes form and writes new predictions for today and the next two days. The board is always ready when you open it.
Demo data notice. Fixtures, ratings and results in this build are generated by the engine itself to show the full daily workflow. Connect a live fixtures and results feed to run it on real matches.
Football is low-scoring and uncertain; even strong models are right about half the time on three-way results. Treat probabilities as estimates, not promises.