Group K is the most predictable in the tournament: Portugal (83.1% qual) and Colombia (79.8%) are nearly tied on qualification probability and nearly tied on Δ xG (+1.07 vs +0.96). It is the only group in the World Cup where two teams sit at 80%+ qualification — the rest of the group (Uzbekistan 27.9% and DR Congo 9.2%) is decoration. The decisive match is the MD3 Colombia–Portugal in Miami (0.31/0.40/0.29): the match of the group, possibly the match of the entire group stage. It is Cristiano Ronaldo’s real farewell at 41.
The Ensemble v1 (DIBP + Dual SPI + Elo + Bayesian Poisson, calibrated with temperature T=0.85 over 364 validation matches, post-calibration ECE 3.3%) produces probability predictions for each of the 103 World Cup fixtures. The table below shows the average per-match win probability for each nation within the group (not the probability of finishing 1st — that is covered in section 4):
| Nation | Per-match win prob (avg) | xG for | xG against | Δ xG |
|---|---|---|---|---|
| 🇵🇹 Portugal | 54.9% | 2.03 | 0.96 | +1.07 |
| 🇨🇴 Colombia | 52.0% | 1.88 | 0.92 | +0.96 |
| 🇺🇿 Uzbekistan | 24.7% | 1.04 | 1.57 | -0.53 |
| 🇨🇩 DR Congo | 11.2% | 0.74 | 2.24 | -1.50 |
Portugal and Colombia combined add up to 163 qual points — the group with the highest concentration of top-2 strength in the entire World Cup. The difference between Colombia (79.8%) and Uzbekistan (27.9%) is 52 points, the second-largest intra-group step in the tournament after Argentina-Algeria.
83.1% qualification, 47.9% to top the group. Δ xG +1.07 (2.03 for / 0.96 against). Cristiano Ronaldo at 41, in his real farewell. The Bernardo Silva / Bruno Fernandes / Vitinha / Bernardo generation gives continuity post-2026. xP 5.79.
Model strengths — Strengths: offensive depth — Bernardo, Bruno, Vitinha, Rafael Leão, Diogo Jota. Δ xG +1.07 says it all. Back-four led by Rúben Dias. Consistent history: 9 consecutive World Cups since 2002.
Risks — Risks: managing Cristiano Ronaldo — if Martínez benches him in MD1 there’s drama; if he plays him there’s a slowdown on the left side of the attack. The MD3 Colombia–Portugal can become a runner-up fixture if MD2 produces a surprise.
Realistic path — Most likely path: 1st in group (47.9%) or 2nd (35.3%). Round of 16 against 2nd of Group H (Uruguay most likely). Tough crossover either way.
79.8% qualification, 41.9% to top the group. Δ xG +0.96 (1.88 for / 0.92 against). Best Colombian side since the Pekerman generation of 2014. James Rodríguez at 35, Luis Díaz as the offensive reference. xP 5.57.
Model strengths — Strengths: Luis Díaz at peak, James as creator, defense with Lerma and Cuesta. Lorenzo deploys a very direct 4-2-3-1. Δ xG +0.96 — 2nd best in the group and top-12 in the World Cup. History: quarterfinals in 2014.
Risks — Risks: 1st vs 2nd depends almost entirely on MD3 vs Portugal. If Colombia draws, they need head-to-head goal difference to favor them — and it likely will, because Colombia should beat DR Congo more handily.
Realistic path — Most likely path: 2nd (37.9%) or 1st (41.9%). The split is almost 50-50. Round of 16 against the 1st/2nd of Group H — Spain or Uruguay.
27.9% qualification. World Cup debut. Δ xG −0.53. Generation raised at Pakhtakor and now spread across Europe: Shomurodov, Khanaev. The best debutant Asian side in World Cup history.
Model strengths — Strengths: World Cup debut doesn’t scare this generation — they live in Europe, Shomurodov at Roma and Khanaev at Ferencváros. Disciplined mid block. The best Uzbek generation in history.
Risks — Risks: 1.57 xG against. Δ xG −0.53. Defense collapses against teams with 2.0+ xG per match — and that’s exactly what Portugal generates (2.03). Only 8.2% chance of topping the group.
Realistic path — Most likely path: 3rd in group. To qualify they need to beat DR Congo in MD3 and have Portugal or Colombia stumble against them. The 28% qual is the highest among any debutant in the tournament.
9.2% qualification. Δ xG −1.50: worst in the group. 2.24 xG against per match. Generation with European pros (Bolasie, Wissa, Mbemba). History: Round of 16 group stage in Brazil 2014. This time will be 4th.
Model strengths — Strengths: European experience (Bolasie historic, Wissa, Mbemba). African coaching staff that knows World Cup conditions. Generation trained with Belgian discipline.
Risks — Risks: 2.24 xG against per match — worst defense in the tournament alongside Cape Verde. Δ xG −1.50. Only 2.0% chance of topping the group. Will lose all 3 in projection — the 9.2% qual is generous.
Realistic path — Most likely path: 4th. The only scenario where they qualify is a wild table upset: beat Portugal or Colombia + draw the other + beat Uzbekistan by +3 goals. Real probability <2%.
The six group games are spread across FIFA-assigned host cities. Probabilities reflect the model's 1X2 distribution for each match, expressed as percentages.
Up to this point we have shown the average per-match win probability (how strong each team is in any given individual match). But "who finishes 1st in the group?" is a different question: it requires winning ENOUGH matches AND scoring MORE points than the other 3 by the end of matchday 3. For that we simulate the entire group 100,000 times and count how often each team finishes 1st, 2nd, 3rd or 4th.
Tournament probabilities (Portugal +1.07 Δ xG, Colombia +0.96) show an extreme two-tier group. The MD3 between the top two is the only group-stage fixture in the 2026 World Cup where two top-15 teams face each other already qualified — and fighting for first place.
Simulating the six group matches 100,000 times under the model's probability distributions, we obtain the following probabilities of finishing the group in each position. These numbers are the most relevant probabilistic input for informed Battle Royale decisions.
| Nation | P (1st) | P (2nd) | P (qualify) | P (eliminated) |
|---|---|---|---|---|
| 🇵🇹 Portugal | 48% | 35% | 83% | 17% |
| 🇨🇴 Colombia | 42% | 38% | 80% | 20% |
| 🇺🇿 Uzbekistan | 8% | 20% | 28% | 72% |
| 🇨🇩 DR Congo | 2% | 7% | 9% | 91% |
Portugal wins the group in 48% of the 100k simulations. Colombia wins in 42%. Joint probability of both advancing: 84%. The most balanced 1st-2nd split in the World Cup: 6 points of difference between top finishers.
Another powerful metric: the expected number of points each nation will accumulate across its three group fixtures, weighted by the model's probabilities. Results:
| Nation | XP | Expected goals (for) | Expected goals (against) |
|---|---|---|---|
| 🇵🇹 Portugal | 5.79 | 5.29 | 2.65 |
| 🇨🇴 Colombia | 5.57 | 4.77 | 2.58 |
| 🇺🇿 Uzbekistan | 3.19 | 2.82 | 3.94 |
| 🇨🇩 DR Congo | 1.73 | 2.20 | 5.91 |
xP Portugal 5.79 vs Colombia 5.57: only 0.22 expected points apart — the tightest top-tier pairing in the entire tournament. Uzbekistan 3.19 sits within the competitive tier. DR Congo 1.73 is one of the lowest xP in the World Cup.
Uzbekistan take a point off Colombia in MD1 Probability 31.2%. If Uzbekistan draw, their qual jumps from 28% to ~52%. Colombia then arrives at MD3 vs Portugal under real pressure — they can’t draw, they need to win to guarantee 1st place.
Portugal put 3+ on DR Congo in MD1 Probability ~40% (rout+). If Portugal accumulate goal diff +5 across MD1+MD2 against the two weak sides, an MD3 draw is enough for 1st place via better goal difference. Colombia would then be forced to win MD3.
The MatchScan Ensemble v1 combines four independent predictive models weighted by their historical accuracy over the consolidated 87,000+ match database:
DIBP (37.5%) — Dynamic Inter-team Bayesian Poisson. Goal-generation model with inter-team rating adjustment.Dual SPI (25%) — Soccer Power Index with separate attack/defense adjustment.Elo v0 (37.5%) — Dynamic Elo system with opponent-strength adjustment.Bayesian Poisson (0% in final blend) — Bayesian model showing instability on asymmetric matches.
The resulting probability set is calibrated via scalar temperature T=0.85 (Platt scaling) with Bayesian correction b_H=-0.24 and b_D=+0.38. Calibration is verified over 364 cross-validation matches, yielding pre-calibration ECE of 4.68% and post-calibration ECE of 3.26% — very low expected calibration error, indicating published probabilities are reliable as frequentist expectations.
World Cup 2026 Battle Royale on MatchScan. Pick 8 teams for the tournament. One per matchday. If your team loses or draws, eliminated. Last survivor wins €100 + 1 year MatchScan VIP. Free entry, no card, no spam.
Sign up →