The 2026 World Cup
forecast match by match
Win/draw/loss odds, predicted scorelines, the bracket all the way to the trophy, and a live "who's going to win" forecast that updates after every result. Built on 150 years of international football data. Free to play.
Free to enter. 100,000 WC Points, real World Cup matches. See the leaderboard →
(1872 → today)
predicted
forecasts stored
kick-off
Who lifts the trophy?
Title odds from 5000 simulations of the full tournament. Updates after every result.
Higher = more likely champion · full forecast →
What you get
Every prediction. Every fixture. Every round. For the whole tournament.
Match predictions
For every fixture: chance each team wins, the most likely scoreline, and how many goals the model expects from each side.
Predicted bracket
The model's path through the knockouts to the trophy — drawn out as a proper bracket, and re-drawn the moment a result lands.
Where the bookies are wrong
Compare what the model thinks each result is worth against what the bookies are paying. Bigger gap = bigger gift from the market.
Teams & squads
Every team's page: ranking, title chance, fixtures, recent results, and the full 26-player squad with shirt numbers and ages.
Browse the 48 teams →Did the model get it right?
After every game: what the model predicted, what actually happened, and a running scoreboard of how many it called correctly.
Open the scoreboard →Why this isn't just vibes
We don't just pick a favourite. We run three different models that look at football in three different ways — then combine them.
Strength rankings (Elo)
Like a chess rating, but for national teams. Win a big match against a strong opponent and you go up; lose to a minnow and you tumble. We've replayed every international ever since 1872 to get the current numbers. Top 5 right now:
- Spain2216
- Argentina2188
- France2123
- England2085
- Brazil2068
How many goals?
The second model thinks about each match as a scoreline. Given how good each team's attack and defence is, what's the chance of 1-0? 2-1? 3-3? Add those up to get win/draw/loss — and you get a real most-likely scoreline too.
Pattern-matching (XGBoost)
The third model is a machine-learning algorithm that's seen tens of thousands of matches and learned which factors actually move the needle — strength gap, recent form, rest days, home advantage, head-to-head history.
Mix & calibrate
We take all three models' votes and blend them — leaning more on whichever one has been most accurate. That blended forecast is what you see on every page.
Honest about the model
International football is hard. Squads change between tournaments, friendlies tell you less than club games, and draws are more common than at club level. So we're upfront: the model is right about the outcome 60% of the time on history it has never seen — clearly better than guessing, but a long way from perfect.
Every page on this site shows the workings, so you can see exactly how the model arrived at every number. These are predictions, not certainties.
100,000 WC Points
to back your calls.
Two free competitions: bet your points on every match, or predict the scores in the tipping ladder. Biggest tally at the final wins.
Join free — get 100,000 WC Points →