Probability Lab · v1Snapshot · 2026-05-14

World Cup 2026, by the numbers.

Match-by-match probabilities for all 104 fixtures of the 2026 FIFA World Cup, generated by two transparent statistical models. We publish the methodology, the input ratings, and the predictions before kick-off — and the accuracy data after. This is applied probability, not a tip sheet.

104

Matches modelled

Teams

Models, compared

Open

Methodology

How the model works·See how it did on past tournaments·Track our calibration as results come in

Framing note

Probabilities, not tips.

Every match in this tool has three probabilities — home win, draw, away win — that sum to one. We report the predicted scoreline as the expected goals from each model, rounded. We do not show odds, we do not show payouts, we do not show stakes, and we do not link to any betting venue.

The point of this exercise is the opposite of a tip sheet. We publish the model openly, the input ratings openly, and the predictions before kick-off. After each match we publish what happened. Over 104 matches that gives a calibration test: are events we said were 70% likely happening about 70% of the time? That is the discipline an honest probabilistic forecaster signs up to. The accuracy page tracks it in real time.

The tools you see below are a working example of applied probability theory: the same Bayesian and Poisson machinery used in sovereign-default modelling, election forecasting, and epidemiology. Football is just a clean, public proving ground.

Tournament outlook · 10,000-path Monte Carlo

Who the model thinks lifts the trophy.

We simulated the entire tournament — 104 matches, group stage through final — ten thousand times, using the Dixon-Coles ensemble. The percentages below are the share of simulations in which each team won the World Cup. Seed: 42. Frozen pre-kickoff in the predictions-snapshot.json file committed to the repository.

Spain

23.3%

to win

SF 41%

QF 52%

R16 66%

Argentina

15.3%

to win

SF 32%

QF 42%

R16 54%

France

12.3%

to win

SF 33%

QF 46%

R16 61%

England

7.3%

to win

SF 24%

QF 36%

R16 55%

Brazil

5.1%

to win

SF 21%

QF 35%

R16 52%

Colombia

4.7%

to win

SF 18%

QF 31%

R16 49%

Portugal

4.7%

to win

SF 19%

QF 32%

R16 51%

Netherlands

4.1%

to win

SF 19%

QF 34%

R16 51%

Audit trail

2026-05-14

Pre-kickoff snapshot frozen

Every prediction on this page was generated and locked in on 2026-05-14 — before any 2026 World Cup match was played. The full set of probabilities, the simulation parameters, and 10,000 simulated brackets are preserved in a version-controlled snapshot that cannot be edited retroactively. If we miss, the miss stays in the record.

01Group stage

72 matches across 12 groups.

For each group we show the model's expected-points projection, then every match with the ensemble probability and predicted scoreline. Click a match to see the two models compared.

Group AHost: MEX

Projected standings

#	Team	xPts	P(advance)P(adv)
1	MexicoCONCACAF	5.7	77%
2	Korea RepublicAFC	4.1	63%
3	CzechiaUEFA	4.0	26%
4	South AfricaCAF	2.6	14%

Match predictions (ensemble)

Matchday 1
Mexico v South Africa
2–1
predicted
61% MEXD 25%RSA 14%
Pick: Mexico(61% conf.)⚠ Models disagree
Matchday 1
Korea Republic v Czechia
1–1
predicted
39% KORD 26%CZE 35%
Pick: Korea Republic(39% conf.)
Matchday 2
Mexico v Korea Republic
2–1
predicted
52% MEXD 25%KOR 23%
Pick: Mexico(52% conf.)⚠ Models disagree
Matchday 2
Czechia v South Africa
2–1
predicted
49% CZED 26%RSA 24%
Pick: Czechia(49% conf.)⚠ Models disagree
Matchday 3
Czechia v Mexico
1–2
predicted
22% CZED 25%MEX 53%
Pick: Mexico(53% conf.)⚠ Models disagree
Matchday 3
South Africa v Korea Republic
1–2
predicted
23% RSAD 26%KOR 51%
Pick: Korea Republic(51% conf.)⚠ Models disagree

Group BHost: CAN

Projected standings

#	Team	xPts	P(advance)P(adv)
1	CanadaCONCACAF	5.4	75%
2	SwitzerlandUEFA	5.2	73%
3	Bosnia and HerzegovinaUEFA	3.4	21%
4	QatarAFC	2.4	12%

Match predictions (ensemble)

Matchday 1
Canada v Switzerland
2–1
predicted
39% CAND 25%SUI 36%
Pick: Canada(39% conf.)
Matchday 1
Qatar v Bosnia and Herzegovina
1–1
predicted
26% QATD 27%BIH 47%
Pick: Bosnia and Herzegovina(47% conf.)⚠ Models disagree
Matchday 2
Canada v Qatar
2–1
predicted
62% CAND 25%QAT 14%
Pick: Canada(62% conf.)⚠ Models disagree
Matchday 2
Bosnia and Herzegovina v Switzerland
1–2
predicted
20% BIHD 26%SUI 54%
Pick: Switzerland(54% conf.)⚠ Models disagree
Matchday 3
Bosnia and Herzegovina v Canada
1–2
predicted
19% BIHD 25%CAN 56%
Pick: Canada(56% conf.)⚠ Models disagree
Matchday 3
Switzerland v Qatar
2–1
predicted
60% SUID 26%QAT 15%
Pick: Switzerland(60% conf.)⚠ Models disagree

Group CNeutral venue

Projected standings

#	Team	xPts	P(advance)P(adv)
1	BrazilCONMEBOL	5.5	75%
2	MoroccoCAF	4.4	65%
3	ScotlandUEFA	4.0	26%
4	HaitiCONCACAF	2.7	15%

Match predictions (ensemble)

Matchday 1
Brazil v Morocco
2–1
predicted
47% BRAD 26%MAR 27%
Pick: Brazil(47% conf.)⚠ Models disagree
Matchday 1
Haiti v Scotland
1–2
predicted
25% HAID 26%SCO 48%
Pick: Scotland(48% conf.)⚠ Models disagree
Matchday 2
Brazil v Haiti
2–1
predicted
59% BRAD 26%HAI 16%
Pick: Brazil(59% conf.)⚠ Models disagree
Matchday 2
Scotland v Morocco
1–1
predicted
33% SCOD 26%MAR 41%
Pick: Morocco(41% conf.)
Matchday 3
Scotland v Brazil
1–2
predicted
24% SCOD 26%BRA 50%
Pick: Brazil(50% conf.)⚠ Models disagree
Matchday 3
Morocco v Haiti
2–1
predicted
52% MARD 26%HAI 21%
Pick: Morocco(52% conf.)⚠ Models disagree

Group DHost: USA

Projected standings

#	Team	xPts	P(advance)P(adv)
1	TürkiyeUEFA	4.6	67%
2	United StatesCONCACAF	4.2	63%
3	ParaguayCONMEBOL	4.0	27%
4	AustraliaAFC	3.6	23%

Match predictions (ensemble)

Matchday 1
United States v Paraguay
2–1
predicted
38% USAD 25%PAR 36%
Pick: United States(38% conf.)
Matchday 1
Australia v Türkiye
1–2
predicted
29% AUSD 26%TUR 45%
Pick: Türkiye(45% conf.)⚠ Models disagree
Matchday 2
United States v Australia
2–1
predicted
42% USAD 25%AUS 33%
Pick: United States(42% conf.)
Matchday 2
Türkiye v Paraguay
1–1
predicted
42% TURD 26%PAR 32%
Pick: Türkiye(42% conf.)
Matchday 3
Türkiye v United States
2–2
predicted
41% TURD 25%USA 34%
Pick: Türkiye(41% conf.)⚠ Models disagree
Matchday 3
Paraguay v Australia
1–1
predicted
40% PARD 26%AUS 34%
Pick: Paraguay(40% conf.)

Group ENeutral venue

Projected standings

#	Team	xPts	P(advance)P(adv)
1	EcuadorCONMEBOL	5.3	73%
2	GermanyUEFA	5.2	72%
3	Côte d'IvoireCAF	3.6	23%
4	CuraçaoCONCACAF	2.5	12%

Match predictions (ensemble)

Matchday 1
Germany v Curaçao
2–1
predicted
60% GERD 26%CUW 15%
Pick: Germany(60% conf.)⚠ Models disagree
Matchday 1
Côte d'Ivoire v Ecuador
1–2
predicted
22% CIVD 26%ECU 52%
Pick: Ecuador(52% conf.)⚠ Models disagree
Matchday 2
Germany v Côte d'Ivoire
2–1
predicted
52% GERD 26%CIV 22%
Pick: Germany(52% conf.)⚠ Models disagree
Matchday 2
Ecuador v Curaçao
2–1
predicted
60% ECUD 26%CUW 14%
Pick: Ecuador(60% conf.)⚠ Models disagree
Matchday 3
Ecuador v Germany
1–1
predicted
38% ECUD 26%GER 36%
Pick: Ecuador(38% conf.)
Matchday 3
Curaçao v Côte d'Ivoire
1–2
predicted
25% CUWD 26%CIV 49%
Pick: Côte d'Ivoire(49% conf.)⚠ Models disagree

Group FNeutral venue

Projected standings

#	Team	xPts	P(advance)P(adv)
1	NetherlandsUEFA	5.2	72%
2	JapanAFC	4.8	69%
3	SwedenUEFA	3.5	22%
4	TunisiaCAF	2.9	17%

Match predictions (ensemble)

Matchday 1
Netherlands v Japan
1–1
predicted
41% NEDD 26%JPN 33%
Pick: Netherlands(41% conf.)
Matchday 1
Tunisia v Sweden
1–1
predicted
31% TUND 26%SWE 42%
Pick: Sweden(42% conf.)
Matchday 2
Netherlands v Tunisia
2–1
predicted
55% NEDD 26%TUN 19%
Pick: Netherlands(55% conf.)⚠ Models disagree
Matchday 2
Sweden v Japan
1–2
predicted
25% SWED 26%JPN 49%
Pick: Japan(49% conf.)⚠ Models disagree
Matchday 3
Sweden v Netherlands
1–2
predicted
23% SWED 26%NED 52%
Pick: Netherlands(52% conf.)⚠ Models disagree
Matchday 3
Japan v Tunisia
2–1
predicted
53% JPND 26%TUN 21%
Pick: Japan(53% conf.)⚠ Models disagree

Group GNeutral venue

Projected standings

#	Team	xPts	P(advance)P(adv)
1	BelgiumUEFA	5.1	72%
2	IranAFC	4.4	65%
3	EgyptCAF	3.8	25%
4	New ZealandOFC	3.1	18%

Match predictions (ensemble)

Matchday 1
Belgium v Egypt
2–1
predicted
48% BELD 26%EGY 26%
Pick: Belgium(48% conf.)⚠ Models disagree
Matchday 1
Iran v New Zealand
2–1
predicted
48% IRND 26%NZL 26%
Pick: Iran(48% conf.)⚠ Models disagree
Matchday 2
Belgium v Iran
1–1
predicted
44% BELD 26%IRN 30%
Pick: Belgium(44% conf.)
Matchday 2
New Zealand v Egypt
1–1
predicted
30% NZLD 27%EGY 44%
Pick: Egypt(44% conf.)
Matchday 3
New Zealand v Belgium
1–2
predicted
21% NZLD 26%BEL 53%
Pick: Belgium(53% conf.)⚠ Models disagree
Matchday 3
Egypt v Iran
1–1
predicted
32% EGYD 26%IRN 42%
Pick: Iran(42% conf.)

Group HNeutral venue

Projected standings

#	Team	xPts	P(advance)P(adv)
1	SpainUEFA	6.1	81%
2	UruguayCONMEBOL	4.7	68%
3	Saudi ArabiaAFC	2.8	16%
4	Cabo VerdeCAF	2.7	15%

Match predictions (ensemble)

Matchday 1
Spain v Cabo Verde
2–1
predicted
63% ESPD 25%CPV 12%
Pick: Spain(63% conf.)⚠ Models disagree
Matchday 1
Saudi Arabia v Uruguay
1–2
predicted
19% KSAD 26%URU 55%
Pick: Uruguay(55% conf.)⚠ Models disagree
Matchday 2
Spain v Saudi Arabia
2–1
predicted
63% ESPD 25%KSA 12%
Pick: Spain(63% conf.)⚠ Models disagree
Matchday 2
Uruguay v Cabo Verde
2–1
predicted
56% URUD 26%CPV 18%
Pick: Uruguay(56% conf.)⚠ Models disagree
Matchday 3
Uruguay v Spain
1–2
predicted
21% URUD 25%ESP 53%
Pick: Spain(53% conf.)⚠ Models disagree
Matchday 3
Cabo Verde v Saudi Arabia
1–1
predicted
35% CPVD 27%KSA 38%
Pick: Saudi Arabia(38% conf.)

Group INeutral venue

Projected standings

#	Team	xPts	P(advance)P(adv)
1	FranceUEFA	5.5	75%
2	NorwayUEFA	4.4	65%
3	SenegalCAF	4.1	28%
4	IraqAFC	2.4	12%

Match predictions (ensemble)

Matchday 1
France v Senegal
2–1
predicted
50% FRAD 26%SEN 25%
Pick: France(50% conf.)⚠ Models disagree
Matchday 1
Norway v Iraq
2–1
predicted
54% NORD 26%IRQ 20%
Pick: Norway(54% conf.)⚠ Models disagree
Matchday 2
France v Norway
2–1
predicted
48% FRAD 26%NOR 27%
Pick: France(48% conf.)⚠ Models disagree
Matchday 2
Iraq v Senegal
1–2
predicted
21% IRQD 26%SEN 53%
Pick: Senegal(53% conf.)⚠ Models disagree
Matchday 3
Iraq v France
1–2
predicted
15% IRQD 25%FRA 60%
Pick: France(60% conf.)⚠ Models disagree
Matchday 3
Senegal v Norway
1–1
predicted
35% SEND 26%NOR 39%
Pick: Norway(39% conf.)

Group JNeutral venue

Projected standings

#	Team	xPts	P(advance)P(adv)
1	ArgentinaCONMEBOL	5.8	78%
2	AustriaUEFA	4.1	62%
3	AlgeriaCAF	3.5	22%
4	JordanAFC	3.1	18%

Match predictions (ensemble)

Matchday 1
Argentina v Algeria
2–1
predicted
57% ARGD 26%ALG 18%
Pick: Argentina(57% conf.)⚠ Models disagree
Matchday 1
Austria v Jordan
2–1
predicted
46% AUTD 26%JOR 28%
Pick: Austria(46% conf.)⚠ Models disagree
Matchday 2
Argentina v Austria
2–1
predicted
54% ARGD 26%AUT 21%
Pick: Argentina(54% conf.)⚠ Models disagree
Matchday 2
Jordan v Algeria
1–1
predicted
33% JORD 26%ALG 40%
Pick: Algeria(40% conf.)
Matchday 3
Jordan v Argentina
1–2
predicted
16% JORD 25%ARG 58%
Pick: Argentina(58% conf.)⚠ Models disagree
Matchday 3
Algeria v Austria
1–1
predicted
31% ALGD 26%AUT 42%
Pick: Austria(42% conf.)

Group KNeutral venue

Projected standings

#	Team	xPts	P(advance)P(adv)
1	PortugalUEFA	5.1	72%
2	ColombiaCONMEBOL	5.1	71%
3	UzbekistanAFC	3.4	21%
4	Congo DRCAF	2.9	16%

Match predictions (ensemble)

Matchday 1
Portugal v Uzbekistan
2–1
predicted
52% PORD 26%UZB 22%
Pick: Portugal(52% conf.)⚠ Models disagree
Matchday 1
Colombia v Congo DR
2–1
predicted
55% COLD 26%COD 19%
Pick: Colombia(55% conf.)⚠ Models disagree
Matchday 2
Portugal v Colombia
1–1
predicted
38% PORD 26%COL 37%
Pick: Portugal(38% conf.)
Matchday 2
Congo DR v Uzbekistan
1–1
predicted
32% CODD 26%UZB 42%
Pick: Uzbekistan(42% conf.)
Matchday 3
Congo DR v Portugal
1–2
predicted
19% CODD 26%POR 55%
Pick: Portugal(55% conf.)⚠ Models disagree
Matchday 3
Uzbekistan v Colombia
1–2
predicted
22% UZBD 26%COL 52%
Pick: Colombia(52% conf.)⚠ Models disagree

Group LNeutral venue

Projected standings

#	Team	xPts	P(advance)P(adv)
1	EnglandUEFA	5.5	74%
2	CroatiaUEFA	4.9	69%
3	PanamaCONCACAF	3.9	25%
4	GhanaCAF	2.3	11%

Match predictions (ensemble)

Matchday 1
England v Croatia
2–1
predicted
43% ENGD 26%CRO 31%
Pick: England(43% conf.)
Matchday 1
Ghana v Panama
1–2
predicted
21% GHAD 26%PAN 52%
Pick: Panama(52% conf.)⚠ Models disagree
Matchday 2
England v Ghana
2–1
predicted
61% ENGD 26%GHA 14%
Pick: England(61% conf.)⚠ Models disagree
Matchday 2
Panama v Croatia
1–2
predicted
28% PAND 26%CRO 46%
Pick: Croatia(46% conf.)⚠ Models disagree
Matchday 3
Panama v England
1–2
predicted
23% PAND 26%ENG 51%
Pick: England(51% conf.)⚠ Models disagree
Matchday 3
Croatia v Ghana
2–1
predicted
58% CROD 26%GHA 16%
Pick: Croatia(58% conf.)⚠ Models disagree

How to read this page

A few honest words on uncertainty.

A 70% probability is not a guarantee. It means: if we ran this scenario 10 times, we'd expect this outcome roughly 7 times. The remaining 3 are not surprises; they are part of the model. The same logic applies to every probability on this page.

When the two models disagree by more than 10 percentage points on any outcome, we flag the match with a ⚠ Models disagree label. That is not a hedge — it is a confession: when two reasonable models disagree, the truth has wider error bars than either suggests on its own.

The model inputs are FIFA Men's World Ranking points (April 2026) and approximate Elo ratings from eloratings.net. We do not use bookmaker odds, transfer-market values, or fitness data — both for transparency and to keep the model citable. The methodology page walks through every equation.

Accountability

Hold us to the calibration plot.

Once group-stage matches start on 11 June 2026, the accuracy page will be live with running Brier scores, log-loss, and a calibration plot of predicted vs realised frequencies. A model that says things confidently and is wrong is not a useful model — and we will be the first to report it.

Methodology →Accuracy tracking →

Elsewhere in the Lab.

The Probability Lab is one part of a wider platform. Three nearby corners worth a look.

Lab hub

All Probability Lab projects

The umbrella for applied-probability work — current and forthcoming projects with open methodology.

Models

Economic models, interactive

15 sliderable economic models — AD-AS, IS-LM, Mundell-Fleming, Solow, Cournot, Phillips, Laffer, and more.

Data

Data Store

Curated CSV and JSON downloads on African public finance, mobile payments, and macroeconomic indicators.

Analysis

What a slider can do that a chalkboard can't

The argument behind the Probability Lab — and the interactive models hub it sits alongside.