worldcup.infratek.ai

A calibrated forecast for the 2026 World Cup knockout stage — probabilities, not certainty. We show our work, and our misses.
P(win cup) · ensemble[strength+market] v0.1.0 · 10000 simsPhase 0· model 0.1.0· updated 28 Jun 2026, 22:22 UTC
Unverified seed. The “market” input is a strength-based proxy until the collection layer lands a vetted odds snapshot. Read the numbers as structurally honest but not yet sourced-final.
First prediction — expect it rough. This is the opening forecast of the run, before any knockout match has informed the model. The intervals are wide on purpose; they narrow as matches are played and the model learns — and you'll be able to watch that happen.

Next match

the pre-match forecast
R32 · next matchforecast · pre-match · ensemble[strength+market] v0.1.0
🇧🇷
Brazil
65%
56%–73%
draw12%8.9%–16%
🇯🇵
Japan
23%
18%–28%
Read: Brazil clearly favored. Each number is a range, not a lock.
🕑 Mon 29 Jun · 16:00 UTC📍 hou · neutral knockout

Who wins the cup?

P(win) · 32 teams · 10,000 sims · interval shown for every team
#TeamP(win cup)ReadIntervalRange
1🇦🇷Argentina32%co-favorite31%–33%
2🇫🇷France13%leading contender12%–14%
3🇪🇸Spain12%leading contender12%–13%
4🏴󠁧󠁢󠁥󠁮󠁧󠁿England8.4%in the mix8.0%–8.9%
5🇧🇷Brazil7.9%in the mix7.5%–8.3%
6🇵🇹Portugal4.5%outside shot4.2%–4.8%
7🇳🇱Netherlands4.2%outside shot3.9%–4.5%
8🇩🇪Germany3.7%outside shot3.3%–4.0%
9🇧🇪Belgium2.1%outside shot1.9%–2.4%
10🇺🇸United States1.7%longshot1.5%–1.9%
11🇨🇴Colombia1.7%longshot1.5%–1.9%
12🇲🇽Mexico1.3%longshot1.1%–1.5%
13🇭🇷Croatia1.2%longshot1.0%–1.4%
14🇨🇦Canada1.2%longshot1.0%–1.4%
15🇲🇦Morocco1.1%longshot0.9%–1.3%
16🇨🇭Switzerland0.9%longshot0.7%–1.0%
17NOR0.6%rank outsider0.5%–0.7%
18🇯🇵Japan0.4%rank outsider0.3%–0.5%
19🇸🇳Senegal0.4%rank outsider0.3%–0.5%
20🇪🇨Ecuador0.2%rank outsider0.1%–0.3%
21ALG0.1%rank outsider<0.1%–0.2%
22🇨🇮Côte d'Ivoire0.1%rank outsider<0.1%–0.2%
23🇦🇺Australia0.1%rank outsider<0.1%–0.2%
24GHA0.1%rank outsider<0.1%–0.2%
25SWE0.1%rank outsider<0.1%–0.2%
26🇦🇹Austria0.1%rank outsider<0.1%–0.1%
27🇵🇾Paraguay<0.1%rank outsider<0.1%–0.1%
28BIH<0.1%rank outsider<0.1%–0.1%
29🇪🇬Egypt<0.1%rank outsider<0.1%–0.1%
30COD<0.1%rank outsider0%–<0.1%
31CPV<0.1%rank outsider0%–<0.1%
32RSA0%rank outsider0%–0%

Road to the final

the bracket shape — results fill in as matches are played
Round of 32 · 32Round of 16 · 16Quarter-finals · 8Semi-finals · 4Final · 2
How this forecast works — and what "calibrated" means

The headline number is P(win the cup): out of many simulated run-throughs of the remaining bracket, the share a team lifts the trophy. It is not a prediction that any one team will win — a 16% favorite is still far likelier to go home than to win it all. That is the point.

The model is an ensemble. A strength model (Elo-style team ratings) gives per-match win/draw/loss odds; a goals model (Dixon–Coles bivariate Poisson) gives scorelines; market-implied odds act as a prior and a benchmark to beat. The blend feeds a Monte-Carlo simulation over the fixed knockout bracket, which tallies how often each team reaches each round and wins the cup.

Every number carries an interval. We sample the model's own parameter uncertainty before simulating, so the range you see is real spread, not decoration. We never publish a point without its band, and we never animate the band — a quivering needle reads as panic; a static range reads as honesty.

"Calibrated" means our 70%s should happen about 70% of the time. Once matches are played we score ourselves on exactly that — Brier score and reliability — and we publish every miss in a public error log, with a one-line cause. The model learns ratings from results as the tournament runs, but its feature weights are tuned on history (past World Cups, Euros, Copas), never on these few games.

Not a betting service. Predictions only. Markets are read-only signal — no betting.