This is a model scenario, updated as results land — not a settled outcome. The 2026 World Cup is in progress; everything below is the Lemeister baseline's read, re-computed against real results. Analysis and education, not betting advice.
Why the model backs Portugal
The Lemeister baseline rates every team on three transparent inputs — an Elo strength rating, recent form, and an expected-goals (xG) proxy — then turns them into match probabilities. Portugal's profile is title-grade:
- Elo 2010 — the model's elite tier, among the strongest sides in the field.
- Form 0.74 — a high recent-results signal (the scale runs 0 to 1).
- xG 1.9 created / 1.0 conceded — a near-two-goals attack over a contained defence.
That is exactly the shape the model rewards: a side that should out-create what it concedes in almost every match it plays. It is why, even after a slow start, Portugal remain the baseline's pick to lift the trophy.
Group K so far
The group has not been the procession the priors implied — which is precisely why the model updates on results rather than reputations.
| Match | Status | Model (pre-match) | Result / projection | MeisterIQ |
|---|---|---|---|---|
| Portugal v DR Congo | Played | POR 66% · draw 20% · 14% | 1–1 (held to a draw) | 60 |
| Portugal v Uzbekistan | Upcoming | POR 71% · draw 18% · 11% | proj 2–0 | 61 |
| Colombia v Portugal | Upcoming | POR 45% · draw 25% · COL 30% | proj 1–0 POR | 53 |
The opener was a warning: the model rated Portugal at 66% to beat DR Congo, and they were held to 1–1. That is the kind of result that trims a group-stage cushion without changing the underlying read — Portugal still profile as comfortable favourites against Uzbekistan (71%) and as narrow favourites in the group decider against Colombia (45% / 25% / 30%), the lowest-conviction call of the three at MeisterIQ 53. Colombia, on Elo 1920 and form 0.70, is the one Group K side the model genuinely respects.
Elsewhere in the field, the bracket has already lost a name: Türkiye are out, eliminated at the group stage after a 0–1 loss to Paraguay. The model drops eliminated sides from the knockout projection the moment the result settles — no team lingers in the bracket on reputation.
The path to the title
Run the model autofill through the knockouts and Portugal's half opens up, while the other half throws up the tournament's story so far — a Netherlands run all the way to the final.
| Round | Match | Model call |
|---|---|---|
| Quarter-final | Argentina v Portugal | POR 57% · proj 1–2 |
| Quarter-final | France v Netherlands | Netherlands through |
| Semi-final | Brazil v Portugal | POR 62% · proj 1–2 |
| Semi-final | Netherlands v Spain | Netherlands through |
| Final | Netherlands v Portugal | POR 56% · proj 1–2 |
Portugal's route is the model's own: past Argentina (57%) in the quarters and Brazil (62%) in the semis — two heavyweight ties the baseline still leans Portuguese on the strength of that 1.9/1.0 xG profile. Opposite them, the bracket sends Netherlands surging through France and Spain to the final — the dark-horse run that makes the showpiece. And in the final, the model holds: Portugal 56%, projected 1–2. Champions.
Open the full interactive bracket on the Lemeister Terminal — every tie carries its own win probability, and it re-fills as real results come in.
Read it live, ask it anything
These are not static numbers. They live in the product and update with the tournament:
- Lemeister Terminal — the live World Cup bracket above, plus Portugal's fixtures, model read and lineups on the World Cup hub and the Portugal team page.
- Ask MeisterQuery — interrogate any of this in plain language at MeisterQuery: try "what's the forecast for Colombia v Portugal", "who is out of the World Cup", or "who does the model have winning the final". Every answer is sourced.
The model has no favourites and no agenda — only the inputs and the math, re-run every time a result lands. Right now, through one draw, one elimination and a Dutch surge, the math still ends the same way: Portugal, champions.
