Ideas worth keeping.
Explore how a probability-backed forecast becomes a useful piece of football intelligence, from the first signal to the final audit.
The Bundesliga title race is over by April
Ten seasons of Bundesliga data show the title is usually decided two to three matchdays before the end. Here is why the pattern is structural, not coincidental.
What makes the three European club competitions different
Champions League, Europa League, Conference League: how promotion and relegation between UEFA's three club competitions changes how a prediction model has to handle each one.
Cup football is a different game
Why the FA Cup, Copa del Rey, DFB-Pokal, and Coupe de France break every prediction model trained on league football.
Which European league is hardest to call
Premier League, La Liga, Serie A, Bundesliga, Ligue 1: a 2024-25 season comparison of home wins, draws, away wins, and goals per match.
How to read a probability-backed football forecast
A practical guide to looking past a single result and understanding what a probability-backed forecast is really saying.
Why predictions should be locked before kickoff
Timing is part of the evidence. A forecast is more useful when its inputs are fixed before the match can influence the story.
What a useful football model can, and cannot, tell you
Models are powerful tools for organizing information, but a useful football model should make uncertainty easier to see, not hide it.