wk2026
Prices every 2026 World Cup match, then measures the gap between the model’s number and the bookmaker’s.
A bookmaker’s odds are a price, and a price can be wrong. The only way to know is to hold your own estimate of the same probability — one the bookmaker being tested didn’t set.
wk2026 builds that estimate for the 2026 World Cup with a Dixon-Coles goal model in Python, so every quoted line can be met with a number of the tournament’s own.
The model turns team strength into goal rates and goal rates into match probabilities. Set against the bookmaker’s odds, each fixture yields an expected value — positive when the price is generous, negative when it isn’t.
The claim doesn’t stop at the estimate. A bet log and a closing-line-value view record each call, so the model is scored against the market rather than trusted on tone.
the model’s number set against the bookmaker’s — value is the gap between them
A live value board for the tournament, read against Unibet NL: matches, the model’s edge on each, and tabs for the bet log and closing line value.
Match by match, the question stays the same and stays checkable — was the price fair, and did the model know before the market did.