PresairaPresairaWorld Cup 2026
ForecastMatchesGroupsPredictPlayers

An independent, calibrated AI forecast. Not affiliated with FIFA or any official World Cup organisation.

For interest and analysis only: not betting advice.

Methodology · how the forecast is built.

Built by Mohammed Ehab Elnomany · probabilistic, not an oracle.

  • Forecast
  • Matches
  • Groups
  • Predict
  • Players

June 11 – July 19, 2026 · United States · Canada · Mexico

World Cup 2026 - every stage played

48 nations, 104 matches, one champion. The road from kick-off to the Final, complete.

See how the AI stacked up against the humans.

See the final scores→

The 2026 World Cup is the first with 48 teams and 104 matches, the largest in the tournament's history.

Tournament Forecast

A calibrated, probabilistic AI forecast.

World Cup 2026 Champion

final result
Champion crown
Spain
Champions

From day one, the model’s top two were the eventual finalists: Spain at 10.9%, Argentina at 9.6%. It called the exact Final.

Backtested on 2018 & 2022 · Updated 19 Jul 2026 · Final run, all 104 results in

Title-odds over time

champion probability
100.0%50.0%0.0%
Spain>99.9%
Argentina<0.1%
Brazil<0.1%
England<0.1%
France<0.1%
Morocco<0.1%
KickoffFinal

Each point is a model forecast published during the tournament, conditioned on every result completed at that time; the line connects successive published forecasts (recalcs ran in batches, so some matches share a segment).

How the model saw it

predicted vs actual
#TeamActual finishvsPre-cup
  1. 1SpainChampions=Finished where the model ranked it, among these eight10.9%
  2. 2ArgentinaRunners-up=Finished where the model ranked it, among these eight9.6%
  3. 3BrazilRound of 16▼22 places worse than the model's rank, among these eight7.0%
  4. 4FranceFourth place=Finished where the model ranked it, among these eight6.6%
  5. 5EnglandThird place▲22 places better than the model's rank, among these eight5.4%
  6. 6PortugalRound of 16▲11 place better than the model's rank, among these eight4.2%
  7. 7GermanyRound of 32=Finished where the model ranked it, among these eight4.1%
  8. 8NetherlandsRound of 32▲11 place better than the model's rank, among these eight4.0%

The model's pre-tournament champion odds (before a ball was kicked) against where each side actually finished. Ordered by that pre-tournament number; the green/red chip is how many places better or worse a side finished than the model ranked it, among these eight.

Golden Boot race

actual / exp
  • 1Kylian Mbappé10 / 2.6 exp
  • 2Lionel Messi8 / 2.4 exp
  • 3Jude Bellingham7 / - exp
  • 4Erling Haaland7 / 2.6 exp
  • 5Ousmane Dembélé6 / - exp
See all scorers →

AI vs Humans

11 brackets
  1. 1rank 1Ahmad Hassan141 / 172 pts
  2. 2rank 2Presaira modelSpainBenchmark138 / 172 pts
  3. 3rank 3Seif Tamer116 / 172 pts
  4. 4rank 4M7mdEhab113 / 172 pts
  5. 5rank 5Mohammed Tarik98 / 172 pts

Round-weighted points vs the actual bracket. Full comparison →

See the final AI vs humans board.
Methodology

How the forecast is built: Elo, Dixon–Coles, Monte Carlo, and backtests.

Read →

June 11 – July 19, 2026 · United States · Canada · Mexico

World Cup 2026 - every stage played

48 nations, 104 matches, one champion. The road from kick-off to the Final, complete.

See how the AI stacked up against the humans.

See the final scores→

The 2026 World Cup is the first with 48 teams and 104 matches, the largest in the tournament's history.

Tournament Forecast

A calibrated, probabilistic AI forecast.

World Cup 2026 Champion

final result
Champion crown
Spain
Champions

From day one, the model’s top two were the eventual finalists: Spain at 10.9%, Argentina at 9.6%. It called the exact Final.

Backtested on 2018 & 2022 · Updated 19 Jul 2026 · Final run, all 104 results in

How the model saw it

predicted vs actual
#TeamActual finishvsPre-cup
  1. 1SpainChampions=Finished where the model ranked it, among these eight10.9%
  2. 2ArgentinaRunners-up=Finished where the model ranked it, among these eight9.6%
  3. 3BrazilRound of 16▼22 places worse than the model's rank, among these eight7.0%
  4. 4FranceFourth place=Finished where the model ranked it, among these eight6.6%
  5. 5EnglandThird place▲22 places better than the model's rank, among these eight5.4%
  6. 6PortugalRound of 16▲11 place better than the model's rank, among these eight4.2%
  7. 7GermanyRound of 32=Finished where the model ranked it, among these eight4.1%
  8. 8NetherlandsRound of 32▲11 place better than the model's rank, among these eight4.0%

The model's pre-tournament champion odds (before a ball was kicked) against where each side actually finished. Ordered by that pre-tournament number; the green/red chip is how many places better or worse a side finished than the model ranked it, among these eight.

Golden Boot race

actual / exp
  • 1Kylian Mbappé10 / 2.6 exp
  • 2Lionel Messi8 / 2.4 exp
  • 3Jude Bellingham7 / - exp
  • 4Erling Haaland7 / 2.6 exp
  • 5Ousmane Dembélé6 / - exp
See all scorers →

Title-odds over time

champion probability
100.0%50.0%0.0%
>99.9%
<0.1%
<0.1%
<0.1%
<0.1%
<0.1%
KickoffFinal

Each point is a model forecast published during the tournament, conditioned on every result completed at that time; the line connects successive published forecasts (recalcs ran in batches, so some matches share a segment).

AI vs Humans

11 brackets
  1. 1rank 1Ahmad Hassan141 / 172 pts
  2. 2rank 2Presaira modelSpainBenchmark138 / 172 pts
  3. 3rank 3Seif Tamer116 / 172 pts
  4. 4rank 4M7mdEhab113 / 172 pts
  5. 5rank 5Mohammed Tarik98 / 172 pts

Round-weighted points vs the actual bracket. Full comparison →

See the final AI vs humans board.
Methodology

How the forecast is built: Elo, Dixon–Coles, Monte Carlo, and backtests.

Read →