# Brazil vs Norway Postgame Audit

Prediction lock: 2026-07-05 10:57 AM EDT

Four-score card: 2-0, 2-1, 1-1

Score 4: 1-2

Support: Over 1.5 YES primary threshold, Under 3.5 secondary threshold.

Final score: Brazil 1-2 Norway

Source: https://www.theguardian.com/football/live/2026/jul/05/brazil-v-norway-world-cup-2026-last-16-live

## Settlement

- Four-score card result: MISS. The locked score set was 2-0, 2-1, 1-1, and the final score was 1-2.
- Four-score result: HIT. The locked score 4 was 1-2.
- Support settlement: Over 1.5 YES hit; Over 3.5 NO hit; Over 2.5 had no clean support recommendation.
- Public finance note: no personal staking ledger is recorded here. Turtle records prediction protocol, not private execution.

## Match Timeline

- Half time: Brazil 0-0 Norway.
- 79': Haaland scored for Norway, making Brazil 0-1 Norway.
- 89': Haaland scored again, making Brazil 0-2 Norway.
- 90+10': Neymar scored a penalty, making Brazil 1-2 Norway.
- Full time: Brazil 1-2 Norway.

The shape matters. score 4 was not hit by a random stoppage-time consolation in a 0-2 model failure. Norway's star branch created the decisive two-goal away score, and Brazil's late penalty supplied the BTTS component that made 1-2 the correct tail instead of 0-2.

## What Went Right

- The full-sweep revision kept Norway alive instead of deleting the underdog because of star distortion. Norway SDI was 0.81 and the star-scoring veto was active before kickoff.
- The score 4 rule did exactly what it was designed to do. It did not select the fourth-highest count cell; it selected the best unrepresented failure branch. That branch was 1-2.
- The exact final score landed in the documented containment set, not in an ad hoc live rewrite.
- The support layer was coherent with the score cloud. All locked scores were over 1.5 and under 3.5, and the match finished with three goals.
- Aaditya's Norway-upset intuition was data-informed: Haaland/Odegaard/Nusa plus Norway's emotional-pressure index created a visible upset branch before the result.

## What Went Wrong

- Turtle still overpromoted the Brazil result thread into the four-score card. Brazil win was 0.736, but the away-tail event was not small enough to be treated as merely decorative.
- The Norway-vs-Brazil realm was sparse but high-signal: only usable, not strong, yet it contained enough star/tactical evidence to deserve a harder promotion test.
- The model treated 1-2 as containment rather than as a possible official branch. That preserved honesty, but it left the public C4 one branch too favorite-heavy.
- Brazil clean control was over-respected by the first score score. The final Neymar goal showed Brazil scoring was real, but Norway scoring twice was the sharper branch.

## Smallest Rule Change

Add a score 4 promotion audit before lock:

If the score 4 is an underdog BTTS score, TailRole(s) >= 0.85, star_distortion_index >= 0.80, emotional_pressure_index_underdog - emotional_pressure_index_favorite >= 0.08, and the favorite-win thread is driven more by opponent-strength realm than by current-match attacking evidence, then compare the score 4 against the weakest four-score card branch by branch diversity, not by raw result probability alone.

For this match, 1-2 would have triggered review because TailRole(1-2)=0.912, Norway SDI=0.81, Norway EPI=0.73, Brazil EPI=0.63, and the weakest official branch was the draw branch 1-1. The rule does not automatically promote every emotional underdog; it only forces a written comparison before the public card is locked.

## Carry Forward

This is a Project Turtle win by containment, not by four-score card. The next model should preserve the pride and the correction together: the tail worked, the score card was still too favorite-biased, and the public record must show both.