How it works

The visualisation is organised like the tournament itself. Each group is a petal, and each team is a bubble inside it.

A bigger bubble means a higher chance of winning the whole World Cup. That is not simply that a team is “better,” but that the model thinks they are more likely to lift the trophy from here.

When you press Play, bubbles animate as matches resolve and everyone's chances shift in real time.

The control bar at the bottom of the main page shows where you are in the tournament: pre-tournament, group-stage matchdays, or a knockout round. Tap the day label to open a picker and jump to any point in the calendar.

When you are not simulating, changing the day updates the visualisation to reflect all real results played up to that point. Win chances and bubble sizes adjust to match that snapshot of the tournament.

To run a simulation, pick the day you want to start from and press Play from. The model will play out the remaining fixtures from there. You can only start from days where the required match results are already in place.

After each real result, we ask: given what is still uncertain, what are all the plausible ways the tournament could finish? Those paths are combined into one win-chance number per team.

For individual matches, we use Elo ratings derived from FIFA ranking. Stronger teams are more likely to win head-to-head. Group-stage uncertainty, third-place qualification, and the knockout bracket all feed into the final number.

These chances are calculated exactly from the model. We are not running thousands of random tournaments and counting winners.

Team 1AlgeriaArgentinaAustraliaAustriaBelgiumBosnia and HerzegovinaBrazilCanadaCape VerdeColombiaCroatiaCuraçaoCzech RepublicDR CongoEcuadorEgyptEnglandFranceGermanyGhanaHaitiIranIraqIvory CoastJapanJordanMexicoMoroccoNetherlandsNew ZealandNorwayPanamaParaguayPortugalQatarSaudi ArabiaScotlandSenegalSouth AfricaSouth KoreaSpainSwedenSwitzerlandTunisiaTurkeyUnited StatesUruguayUzbekistan

Team 2AlgeriaArgentinaAustraliaAustriaBelgiumBosnia and HerzegovinaBrazilCanadaCape VerdeColombiaCroatiaCuraçaoCzech RepublicDR CongoEcuadorEgyptEnglandFranceGermanyGhanaHaitiIranIraqIvory CoastJapanJordanMexicoMoroccoNetherlandsNew ZealandNorwayPanamaParaguayPortugalQatarSaudi ArabiaScotlandSenegalSouth AfricaSouth KoreaSpainSwedenSwitzerlandTunisiaTurkeyUnited StatesUruguayUzbekistan

BrazilElo 2052

vs

Elo 2088Argentina

44.8% Brazil55.2% Argentina

$$P(\text{Brazil wins}) = \frac{1}{1 + 10^{(2088 - 2052) / 400}} = 44.8\%$$

Play picks one random path through the remaining fixtures, weighted by each team's Elo rating. After every match, ratings update and everyone's win chances are recalculated. That is why bubbles jump during a simulation.

The randomness is seeded, so starting from the same day always produces the same run. You are watching one possible tournament unfold, while the numbers on screen always reflect the full range of what could still happen.

1. You press Play from a chosen day.
2. Each match day, a winner is picked based on Elo.
3. Win chances are recalculated for every team.
4. Bubbles resize to match the new numbers.

The bubbles show each team's chance of winning the World Cup. Those chances come from a model that weighs every plausible path through the bracket. Play runs one random path through the rest of the tournament so you can watch how a single storyline changes the picture.