FootyPulse World Cup 2026 Round 2 Predictions — The Full Accuracy Report
Published 26 June 2026 · FootyPulse AI Analysis
8 min read
FootyPulse World Cup 2026 Round 2 Predictions — The Full Accuracy Report
Published June 2026 · FootyPulse Analysis
Round 2 of the 2026 World Cup is complete. All 24 group stage second fixtures have been played — though two matches (Jordan vs Algeria and Colombia vs Congo DR) had no predictions generated in time, so this report covers the 22 matches where we had pre-match AI predictions.
Here's the full honest scorecard.
The headline numbers
- 22 matches predicted
- 15 correct outcomes (68.2%) — we called the right W/D/L result
- 4 exact scorelines (18.2%) — we called the precise final score
- 7 matches wrong (31.8%)
- 20 half-time re-predictions generated — the first full round using the live stats-enhanced HT model
Compared to Round 1:
| Metric | Round 1 | Round 2 | Change |
|---|---|---|---|
| Correct outcomes | 62.5% (15/24) | 68.2% (15/22) | +5.7% |
| Exact scorelines | 8.3% (2/24) | 18.2% (4/22) | +9.9% |
| Wrong predictions | 37.5% | 31.8% | -5.7% |
Both metrics improved meaningfully in Round 2. The exact score rate more than doubled from Round 1 to Round 2 — from 2 exact scores to 4. The model is getting sharper as real tournament data replaces pre-tournament assumptions.
The four exact scores
Brazil 3-0 Haiti — predicted 3-0 at 78% confidence. Exact. ✅
Argentina 2-0 Austria — predicted 2-0 at 68% confidence. Exact. ✅ Messi missed a penalty in the first half, then scored twice. Our AI re-predicted 2-0 at half-time despite the miss — the first-half stats (Argentina's complete dominance, Austria's zero attacking threat) confirmed the scoreline was coming.
USA 2-0 Australia — predicted 2-0 at 62% confidence. Exact. ✅
Scotland 0-1 Morocco — predicted 0-1 at 45% confidence. Exact. ✅ Our lowest-confidence exact score of the round — and arguably our most impressive call. Scotland vs Morocco was a tight match on paper, and predicting the precise 0-1 scoreline at only 45% confidence shows the model finding the right call even when the data is genuinely close.
The best correct calls
France 3-0 Iraq (78% confidence, predicted 4-0) The biggest correct call of the round by confidence. France won comfortably despite a two-hour lightning delay that halted play at half-time — the longest weather interruption in World Cup 2026 so far. The model correctly identified France's superiority regardless of conditions. We were one goal high on the margin.
Portugal 5-0 Uzbekistan (72% confidence, predicted 3-0) Portugal smashed five past Uzbekistan — two goals more than predicted. The direction was emphatic and correct; we underestimated how badly Uzbekistan would be outclassed. Our half-time re-prediction saw Portugal leading 3-0 and correctly called the final score: 5-0. Exact at HT.
Spain 4-0 Saudi Arabia (68% confidence, predicted 2-0) Spain continued their tournament dominance. At half-time: 73% possession, 5 shots on target to zero for Saudi Arabia. The enhanced HT model re-predicted 5-0. Final score: 4-0. One goal off — but the most vivid demonstration of what live stats add to the half-time re-prediction.
Canada 6-0 Qatar (68% confidence, predicted 2-0) Canada were ruthless. We predicted 2-0 — the actual scoreline was 6-0. The model correctly identified Canada's superiority but couldn't anticipate the scale of the rout. HT re-prediction after Canada led 3-0 called the final at 5-0. Actual: 6-0. One goal off at HT.
The biggest misses
Ecuador 0-0 Curaçao (78% confidence, predicted 3-0) This is our single biggest miss of Round 2 — matching Spain vs Cape Verde Islands from Round 1 for sheer surprise factor. Ecuador, ranked 52 places above Curaçao, failed to score despite dominating. Curaçao set up defensively and frustrated a side that struggled to break down a low block. The same pattern that caught us out with Spain in Round 1: highly favoured team, organised underdog, 0-0. This is a systematic gap in how the model weights defensive organisation against expected dominance.
England 0-0 Ghana (68% confidence, predicted 2-1) England's failure to score against Ghana — our most high-profile miss given the UK audience. England had 81% possession at half-time with 6 shots but couldn't convert. The HT re-prediction, seeing that dominance, called 2-0 for England. The model was right about England controlling the match — wrong about them converting that control into goals. Ghana's defensive discipline was exceptional and England paid the price for poor finishing, not poor tactics.
Uruguay 2-2 Cape Verde Islands (68% confidence, predicted 2-0) Uruguay led, conceded, and couldn't hold on against a Cape Verde side that's proving to be the tournament's most resilient underdog. Cape Verde have now drawn 0-0 with Spain and 2-2 with Uruguay. The HT model saw Uruguay leading with dominant stats and predicted 4-1. Final: 2-2. The model correctly read Uruguay's dominance but completely missed Cape Verde's ability to convert rare chances.
Belgium 0-0 Iran (58% confidence, predicted 2-0) Belgium dominated possession and shots but failed to score. The HT model saw Belgium's complete first-half dominance and predicted 2-0. Final: 0-0. The Iran goalkeeper had an exceptional game — exactly the kind of individual performance that no statistical model can predict.
The draw problem — still present but improving
In Round 1, we predicted 2 draws and 9 happened. In Round 2:
- Draws predicted: 1
- Draws that happened: 5
- Draws correctly predicted: 0
The draw underweighting is the single most consistent finding across both rounds. We've now predicted 3 draws across 46 matches and 14 draws have actually occurred. AI models systematically favour decisive outcomes — it's an inherent limitation of prediction models trained on historical form data where draws are the least predictable outcome.
The five Round 2 draws — Ecuador 0-0 Curaçao, Belgium 0-0 Iran, England 0-0 Ghana, Czech Republic 1-1 South Africa, Uruguay 2-2 Cape Verde Islands — share a common thread: a higher-ranked team failing to break down a well-organised, motivated underdog.
The half-time re-prediction model — first full round results
Round 2 was the first full round using the stats-enhanced half-time re-prediction — incorporating live possession, shots on target, corners, and cards rather than just the scoreline. Results across 20 matches:
- 4 exact HT re-predictions (correct final score called at half-time)
- 11 correct outcomes (right team won or correctly called draw)
- 5 wrong outcomes
The standout HT moments:
- Germany vs Ivory Coast: HT predicted 2-1, actual 2-1. Exact. ✅
- France vs Iraq: HT predicted 3-0, actual 3-0. Exact. ✅ (called after 2-hour weather delay)
- Portugal vs Uzbekistan: HT predicted 5-0, actual 5-0. Exact. ✅
- Argentina vs Austria: HT predicted 2-0, actual 2-0. Exact. ✅ (after Messi penalty miss)
The England vs Ghana HT re-prediction is the most instructive miss: England had 81% possession and 6 shots at half-time — dominant by every metric — and the model correctly read that England should win the second half. They didn't. When elite attackers can't convert dominant possession, no statistical model can override human finishing failures.
Two matches with no prediction
Jordan vs Algeria and Colombia vs Congo DR did not have predictions generated before kickoff. Both results are excluded from this accuracy report. This is a data gap we're working to close — the daily prediction cron and match page auto-generation should prevent this in future rounds.
What changes for Round 3
Round 3 features a significant new variable: squad rotation. Teams that have already secured their group position before the final round will likely rest key players to preserve fitness for the knockout stage. Our Round 3 predictions now explicitly account for this — assessing each team's qualification status and adjusting win probabilities and predicted scorelines for matches where rotation is likely.
Spain, France, Argentina, Portugal, and Brazil all enter Round 3 as group leaders with qualification secure. Expect our predictions for their final group games to reflect the likelihood of weakened XIs rather than treating them as full-strength sides.
The full Round 2 scorecard
| Match | Predicted | Actual | Result |
|---|---|---|---|
| Scotland vs Morocco | 0-1 | 0-1 | ✅ Exact score |
| USA vs Australia | 2-0 | 2-0 | ✅ Exact score |
| Brazil vs Haiti | 3-0 | 3-0 | ✅ Exact score |
| Argentina vs Austria | 2-0 | 2-0 | ✅ Exact score |
| Switzerland vs Bosnia & Herzegovina | 2-1 | 4-1 | ✅ Correct |
| Canada vs Qatar | 2-0 | 6-0 | ✅ Correct |
| Mexico vs South Korea | 2-1 | 1-0 | ✅ Correct |
| Netherlands vs Sweden | 2-1 | 5-1 | ✅ Correct |
| Germany vs Ivory Coast | 3-0 | 2-1 | ✅ Correct |
| Tunisia vs Japan | 1-2 | 0-4 | ✅ Correct |
| Spain vs Saudi Arabia | 2-0 | 4-0 | ✅ Correct |
| France vs Iraq | 4-0 | 3-0 | ✅ Correct |
| Norway vs Senegal | 2-1 | 3-2 | ✅ Correct |
| Portugal vs Uzbekistan | 3-0 | 5-0 | ✅ Correct |
| Panama vs Croatia | 1-2 | 0-1 | ✅ Correct |
| Ecuador vs Curaçao | 3-0 | 0-0 | ❌ Missed |
| Belgium vs Iran | 2-0 | 0-0 | ❌ Missed |
| Uruguay vs Cape Verde Islands | 2-0 | 2-2 | ❌ Missed |
| England vs Ghana | 2-1 | 0-0 | ❌ Missed |
| Czech Republic vs South Africa | 2-0 | 1-1 | ❌ Missed |
| Türkiye vs Paraguay | 2-1 | 0-1 | ❌ Missed |
| New Zealand vs Egypt | 1-1 | 1-3 | ❌ Missed |
| Jordan vs Algeria | No prediction | 1-2 | — |
| Colombia vs Congo DR | No prediction | — | — |
Running tournament totals (Rounds 1 + 2)
Across 46 predicted matches:
- Correct outcomes: 30/46 = 65.2%
- Exact scorelines: 6/46 = 13.0%
- Draws predicted: 3 | Draws actual: 14 | Draws correct: 0
View live results and predictions →
Half-time re-predictions available for Pro subscribers →
FootyPulse predictions are for entertainment purposes only. We are 100% independent — no betting affiliates, no sponsored content, no commercial influence on our predictions.
Share this article: