Brazil two-goal control is the strongest opponent-threshold family.
Project Turtle / full-sweep lock / 2026-07-05
Brazil vs Norway
Brazil remained the correct favorite in the result thread. The revision is that Norway cannot be erased as a generic underdog: Haaland, Odegaard, and Nusa activate a bounded scoring veto, so Turtle must keep one Brazil-BTTS score and one normal upset score.
Full-sweep readiness: opponent-threshold strong for Brazil, usable for Norway; lineup and player layers are proxy-rated, not official XI.
Locked four-score card
Card: 2-0, 2-1, 1-1, 1-2.
Brazil still wins, but Norway scoring branch survives the star-scoring veto.
The draw/block branch is weaker than Brazil win but stronger than Norway win.
Star/emotional upset score: Norway scoring twice completes C4 as the best danger branch.
This is a pre-kickoff full-sweep revision of the earlier team-threshold card. The old clean-control duplicate slot was removed by the star scoring veto.
Pregame / postgame analysis
The same file will hold the before and after.
Pregame analysisBrazil are favored; Norway scoring cannot be discarded.
The old card selected 1-0 over 2-1 by raw exact-count. Full-sweep rejects that as under-modeled because Norway SDI=0.81, BTTS=0.443, and Brazil clean sheet=0.513. That is enough to keep 2-1 public.
Market prices are not part of the score proof. They can only decide whether the locked cloud has a support line worth taking.
Postgame analysisFinal: Brazil 1-2 Norway. C4 hit.
Brazil scored late; Norway had already built the upset branch.
The locked score set was 2-0, 2-1, 1-1, 1-2.
The fourth normal score was exactly 1-2.
Over 1.5 YES and Over 3.5 NO both settled correctly.
The public lesson is precise. Turtle landed the four-score card because the full-sweep revision preserved the one score that broke the favorite script. That is the purpose of score 4: not decorative, not a random long shot, but a normal C4 member that keeps the best unrepresented football story alive when the score card is too concentrated.
Timeline: the match was 0-0 at half time, Haaland scored at 79' and 89', and Neymar's stoppage-time penalty made the final 1-2. That sequence validates the score-four branch: Norway's star branch produced the upset, while Brazil's scoring gate made 1-2 better than a Norway clean-sheet score.
What went right: Norway's star layer was not erased. The file had Norway SDI=0.81, an active star-scoring veto, and an emotional-pressure index of 0.73 against Brazil's 0.63. The tail comparison gave 1-2 a TailRole of 0.912, far above 3-1 at 0.330, because 3-1 duplicated already represented Brazil branches while 1-2 covered the missing Norway-upset branch.
What still needs discipline: Brazil win probability was modeled at 0.736, so the underdog BTTS branch was easy to under-respect. The next correction is a score-four promotion audit: when TailRole is at least 0.85, underdog SDI is at least 0.80, and emotional pressure favors the underdog by at least 0.08, the weakest C4 branch must be compared against the underdog score before lock.
Objective model specification
The proof begins with definitions, not confidence.
The key correction is opponent thresholding. Brazil is not judged by all last-50 rows equally; Brazil is judged primarily by rows near Norway's strength realm, and Norway is judged primarily by rows near Brazil's strength realm. The formula below makes the data weighting visible before any exact score is selected.
| Symbol | Definition | Value Used | Why It Matters |
|---|---|---|---|
| R_i(j, eps) | Rows in team i ledger whose opponent rating lies within eps of opponent j rating. | eps=0.75; widen only if sparse | Prevents flat last-50 averages from treating all opponents as equal. |
| w_m | w_rec(m) * w_comp(m) * w_squad(m) * exp(-(r_opp(m)-r_target)^2/(2*sigma^2)) | sigma=0.75; unknown squad overlap defaults to bounded proxy | Rows closer to today's opponent strength and seriousness count more. |
| GF_R(i|j) | sum_{m in R_i(j)} w_m GF_m / sum_{m in R_i(j)} w_m | Brazil vs Norway realm GF=2.3078; Norway vs Brazil realm GF=1.2189 | This is the opponent-threshold attack estimate before match-state gates. |
| GA_R(i|j) | sum_{m in R_i(j)} w_m GA_m / sum_{m in R_i(j)} w_m | Brazil concession realm=0.3930; Norway concession realm=1.9560 | This is the opponent-threshold defensive weakness estimate. |
| lambda_BRA_raw | (A_BRA + D_NOR) / 2 | (2.3078 + 1.9560) / 2 = 2.1319 | Brazil expected goals are anchored in Brazil attack and Norway defensive allowance. |
| lambda_NOR_raw | (A_NOR + D_BRA) / 2 | (1.2189 + 0.3930) / 2 = 0.8059 | Norway expected goals are anchored in Norway attack and Brazil defensive allowance. |
| lambda_final | lambda_raw * m_knockout * m_lineup * m_tactical * m_emotion, with each multiplier bounded | lambda_BRA=2.1121; lambda_NOR=0.7827 | Context can move the mean slightly; it cannot invent an unsupported score family. |
| P_count(h,a) | Pois(h;lambda_BRA) * Pois(a;lambda_NOR) | Pois(h;2.1121) * Pois(a;0.7827) | This is the exact-score count kernel. It is an input to selection, not the whole proof. |
| TailRole(s) | 0.40 U(s) + 0.25 G(s) + 0.20 S(s) + 0.15 Q(s) | U=unrepresented family, G=gate survival, S=support compatibility, Q=probability normalized among tail comparisons | score 4 is not the fourth-highest score; it is the best unrepresented failure branch. |
Mathematical derivation
Not a display table: the card changes only when a rule beats the count model.
The raw Poisson grid is only the count kernel. Turtle first writes the kernel, then asks whether the top cells duplicate the same football story. A score can lose despite higher raw probability only when a declared rule supplies a numerical inequality.
\[
P_{\mathrm{count}}(h,a)
=
e^{-2.1121}\frac{2.1121^h}{h!}
\cdot
e^{-0.7827}\frac{0.7827^a}{a!}.
\]
The lambdas come from opponent-threshold attack and defensive-weakness estimates, then bounded context gates. This kernel ranks score cells; it does not by itself decide the four-score card.
\[
P(2,0)=0.123367
>
P(1,0)=0.116821
>
P(2,1)=0.096561
>
P(1,1)=0.091438.
\]
A pure count model would initially prefer 1-0 over 1-1. Turtle does not accept that automatically because 2-0 and 1-0 are both clean-control claims.
\[
\mathrm{SDI}_{NOR}=0.81\ge 0.70,\quad
P(\mathrm{BTTS})=0.443\ge0.400,\quad
P(\mathrm{BRA\ clean})=0.513<0.620.
\]
Since Norway has an active star-scoring veto, the second clean-sheet score cannot occupy a public slot after 2-0. This removes 1-0 from C4; it does not claim 1-0 is impossible.
\[
I(2,1)=0.354174>I(1,0)=0.349197.
\]
Once the duplicate clean slot is vetoed, 2-1 is the Brazil-win score that preserves the Norway scoring branch and beats 1-0 under the branch-aware Turtle index.
\[
P(\mathrm{draw})=0.187>P(\mathrm{NOR\ win})=0.077,\quad
P(\mathrm{over\ 1.5})=0.815.
\]
The low draw branch is stronger than the Norway-win branch. The over-1.5 gate rejects 0-0, so the draw representative is 1-1.
\[
P_{\mathrm{count}}(3,1)=0.067981
>
P_{\mathrm{count}}(1,2)=0.035785,
\qquad
\mathrm{TailRole}(1,2)=0.912>\mathrm{TailRole}(3,1)=0.330.
\]
This is the whole distinction. If score 4 meant fourth-highest probability, 3-1 wins. But score 4 means best unrepresented failure branch. Since 2-0 already covers Brazil cushion and 2-1 covers Brazil-BTTS, 3-1 is represented. The missing story is Norway upset with both teams scoring, so 1-2 is the score 4.
Objective claims
Every football claim is tied to an inequality.
This table is deliberately plain. A claim survives only if the evidence and inequality are visible; the caution column records what the claim does not prove.
| Claim | Evidence | Inequality | Conclusion | Caution |
|---|---|---|---|---|
| Brazil control is the primary branch. | Brazil realm GF=2.304 and Norway elite-realm GA=2.385; result thread Brazil win=0.736. | P(Brazil by 2+)=0.546 > P(Brazil by 1)=0.197 > P(draw)=0.187 > P(Norway win)=0.077 | The first score representative must come from Brazil cushion; 2-0 is selected because it is the largest cell in that family. | This does not prove a clean sheet by itself; it only proves the dominant result-spread family. |
| A duplicate clean-control card is overconfident. | Norway SDI=0.81, BTTS=0.443, Brazil clean-sheet thread=0.513. | Brazil clean sheet 0.513 < duplicate-clean threshold 0.620 while BTTS 0.443 >= scoring-veto threshold 0.400 | 1-0 cannot occupy a score slot after 2-0; 2-1 becomes the scoring-branch representative. | 1-0 remains a live football score, but Turtle refuses to spend two public slots on the same clean-sheet story. |
| The draw branch is stronger than the Norway-win branch. | Result thread draw=0.187 and Norway win=0.077; over-1.5=0.815. | P(draw)=0.187 > P(Norway win)=0.077 and P(over 1.5)=0.815 | Keep 1-1 as the draw representative; reject 0-0 because it contradicts the total thread. | This is not a claim that Brazil fail often. It is a branch-preservation rule after Brazil control is already represented. |
| The score 4 must protect the missing story, not the fourth probability. | 3-1 has P_count=0.067981; 1-2 has P_count=0.035785; TailRole(3-1)=0.330; TailRole(1-2)=0.912. | P_count(3-1)>P_count(1-2), but TailRole(1-2)>TailRole(3-1) | 3-1 is rejected as represented by 2-0 and 2-1; 1-2 is kept as the best unrepresented Norway-upset failure mode. | If the chosen philosophy were pure containment probability, 3-1 would be the fourth score. Turtle's declared four-score protocol is different. |
Full-sweep layers
What changed after the old thread model.
Star scoring veto active.
Real, but below duplicate-clean threshold.
Enough to preserve one scoring branch.
Medium; opens tail, not chaos.
| Family | Representatives | Probability Index | Status | Reason |
|---|---|---|---|---|
| Brazil cushion clean | 2-0, 3-0, 4-0 | 0.330 | score_represented_by_2-0 | strongest spread branch and best S10 cell |
| Brazil win BTTS | 2-1, 3-1 | 0.240 | score_represented_by_2-1 | star scoring veto makes this non-duplicate branch necessary |
| Low draw | 1-1, 0-0 | 0.150 | score_represented_by_1-1 | draw thread beats Norway win and 0-0 fails total gate |
| Norway upset one goal | 1-2, 0-1 | 0.075 | score4_represented_by_1-2 | upset tail preserved by star/emotional gate; 0-1 fails Brazil scoring |
| High chaos | 3-2, 2-2, 4-1 | 0.055 | excluded | volatility and five-goal gates below full activation |
| Metric | Value | Threshold | Status | Model Effect |
|---|---|---|---|---|
| game_state_volatility | 0.60 | 0.65 for high chaos | medium | BTTS and upset tail visible; 3-2 not admitted |
| five_goal_mass | 0.167 | 0.20 for 3-2 activation unless other gates all fire | below_gate | rejects 3-2 |
| btts_yes | 0.443 | 0.40 for scoring-branch preservation | active | 2-1 enters C4 |
Candidate table
Every S10 score and opened tail receives one status.
| Rank | Score | Poisson Probability | Turtle Index | Status | Rule | Deduction |
|---|---|---|---|---|---|---|
| 1 | 2-0 | 0.123367 | 0.782287 | SCORE | R2 strongest result-spread family | Brazil home_by_2_plus mass 0.545636 is the strongest spread state; 2-0 has the largest P_count inside that family. |
| 2 | 1-0 | 0.116821 | 0.349197 | EXCLUDED_BY_BTTS_CLEAN_SHEET_THREAD | R_control_double_slot + R_star_scoring_veto | 1-0 is a second clean-control slot. Since Norway SDI=0.81, BTTS=0.442967, and Brazil clean sheet=0.513335<0.62, the duplicate slot loses to 2-1. |
| 3 | 2-1 | 0.096561 | 0.354174 | SCORE | R_star_scoring_veto | 2-1 shares home_by_1 with 1-0, has I=0.354174>I(1-0)=0.349197, and preserves the active Norway scoring branch. |
| 4 | 1-1 | 0.091438 | 0.316635 | SCORE | R_draw_branch_preservation | Draw probability 0.187368 beats Norway win 0.076546, and 1-1 beats 0-0 because Over 1.5=0.815159. |
| 5 | 3-0 | 0.086852 | 0.578840 | EXCLUDED_BY_INDEX_COMPARISON | R_family_representative | 3-0 shares the Brazil clean-cushion family with 2-0 and loses P_count 0.086852<0.123367. |
| 6 | 3-1 | 0.067981 | 0.530733 | EXCLUDED_BY_INDEX_COMPARISON | R_family_representative | 3-1 is a Brazil cushion-BTTS child, but 2-1 already represents the BTTS win branch with lower total risk. |
| 7 | 0-0 | 0.055312 | 0.212618 | EXCLUDED_BY_TOTAL_THREAD | R_total_gate | 0-0 contradicts Over 1.5=0.815159 and Brazil scored-rate gates. |
| 8 | 4-0 | 0.045859 | 0.431436 | EXCLUDED_BY_INDEX_COMPARISON | R_chaos_cap | 4-0 shares Brazil clean-cushion with 2-0 and has lower P_count 0.045859. |
| 9 | 0-1 | 0.043293 | 0.197881 | EXCLUDED_BY_BTTS_CLEAN_SHEET_THREAD | R_brazil_scoring_gate | Norway clean-sheet state is 0.084521 and Brazil attack realm is too strong to make 0-1 the upset representative. |
| 10 | 2-2 | 0.037790 | 0.197382 | EXCLUDED_BY_SPREAD_THREAD | R_draw_high_gate | 2-2 loses draw representation to 1-1; high draw stays below tail priority. |
| 12 | 1-2 | 0.035785 | 0.281000 | SCORE_4 | R_upset_tail_preservation | Norway win mass is too small for C4 but star/emotional gates preserve the best upset tail; 1-2 beats 0-1 by Brazil scoring gate. |
| 13 | 3-2 | 0.026605 | 0.281000 | EXCLUDED_BY_TOTAL_THREAD | R_chaos_3_2_gate | 3-2 requires five-goal chaos. Five-goal mass=0.167 and game-state volatility=0.60 are below the declared activation gates. |
Score 4 comparison
Why 1-2 remains the score 4, not 3-1.
If Turtle simply chose the fourth-highest count probability, 3-1 would win. The four-score rule is different: score 4 must cover the best unrepresented failure branch after C4. Here TailRole(s)=0.40U+0.25G+0.20S+0.15Q. Because 2-0 already covers Brazil cushion and 2-1 already covers Brazil-BTTS, 3-1 has TailRole 0.330. The Norway upset branch 1-2 has TailRole 0.912 and becomes the score 4.
| Candidate Score | Poisson Probability | Probability Norm | Unrepresented Family Score | Gate Survival Score | Support Compatibility Score | Tail Role Score | Decision | Proof Reason |
|---|---|---|---|---|---|---|---|---|
| 3-0 | 0.086852 | 1.000 | 0.00 | 0.20 | 0.70 | 0.340 | reject | Highest excluded raw cushion cell, but TailRole=0.340 because the family is already represented by 2-0. |
| 3-1 | 0.067981 | 0.783 | 0.00 | 0.45 | 0.50 | 0.330 | reject_as_fourth_probability_not_tail | If score 4 were the fourth-highest count cell, 3-1 would win. Under TailRole it loses because Brazil cushion and Brazil BTTS are already represented. |
| 2-2 | 0.037790 | 0.435 | 0.00 | 0.35 | 0.50 | 0.253 | reject | Draw family already has 1-1; high draw lacks enough volatility to replace the low draw. |
| 0-1 | 0.043293 | 0.498 | 1.00 | 0.20 | 0.25 | 0.575 | reject | It covers Norway upset, but it requires Brazil not scoring; Norway clean-sheet state is only 0.084521. |
| 1-2 | 0.035785 | 0.412 | 1.00 | 1.00 | 1.00 | 0.912 | select_score 4 | It is the best unrepresented upset branch after rejecting 0-1, and it is compatible with both Over 1.5 and Under 3.5 support. |
| 3-2 | 0.026605 | 0.306 | 0.00 | 0.20 | 0.50 | 0.196 | reject | It needs five-goal chaos; five-goal mass 0.167 and volatility 0.60 miss the activation gate. |
Lineup and star layer
Expected-player context is bounded and auditable.
These are not official lineups. They are proxy inputs used only to prevent obvious star-player distortion from being lost in a flat last-50 pool.
| Team Code | Player | Position | Expected Starter | Expected Minutes Bucket | Confidence | Notes |
|---|---|---|---|---|---|---|
| BRA | Vinicius Junior | LW | 1 | 70-90 | medium | Proxy star layer; lineup not official in local file. |
| BRA | Neymar | AM/FW | 1 | 60-90 | medium | Public preview context only. |
| BRA | Rodrygo | RW/FW | 1 | 60-90 | medium | Public preview context only. |
| NOR | Erling Haaland | ST | 1 | 80-90 | medium | Star scoring veto anchor. |
| NOR | Martin Odegaard | AM | 1 | 70-90 | medium | Creative support proxy. |
| NOR | Antonio Nusa | W | 1 | 60-80 | medium | Transition threat proxy. |
Support market after lock
Lean: Over 1.5 YES first; Under 3.5 second.
| Market | Side | Model Probability | Fair Price | Max Entry Price | Suggested Stake Usd | Portfolio Role | Decision |
|---|---|---|---|---|---|---|---|
| Over 1.5 goals | YES | 0.815159 | 0.82 | 0.76 | 10-15 | Protects all C4 scores because every locked score has at least two total goals. | primary support lean if executable price is below threshold |
| Over 3.5 goals | NO | 0.700981 | 0.70 | 0.64 | 5-10 | Protects all C4 scores because all locked scores are under four total goals. | secondary support; do not over-size because Brazil 3-1 is an excluded but live branch |
| Over 2.5 goals | NO | 0.473329 | 0.47 | 0.41 | 0 | Conflicts with 2-1 and 1-2; not a clean support market. | no support bet |
| BTTS | YES | 0.442967 | 0.44 | 0.38 | 0-5 | Supports 2-1, 1-1, 1-2 but conflicts with 2-0. | optional only at large discount; not primary |
Formal proof ledger
Definitions, data, rules, lemmas, theorem.
D1Universe
Let G={0,...,7}x{0,...,7}. S10 is the first ten cells of G sorted by P_count(h,a)=Pois(h;2.1121)Pois(a;0.7827), with exact ties broken lexicographically.
D2Threads
Each score inherits result, spread, total, BTTS/clean-sheet, and score-family evidence. No market price is used before lock.
D3Star scoring veto
If an underdog has SDI_team>=0.70, BTTS>=0.40, and favorite clean-sheet probability<0.62, a duplicate favorite-clean slot must yield to the best scoring-branch representative.
D4Chaos gate
A five-goal score such as 3-2 may enter score 4 only if five-goal mass>=0.20 or game-state volatility>=0.65 with BTTS and star gates active.
D5Score 4 objective
score 4 is not defined as the fourth-highest probability cell. score 4 is the best unrepresented failure branch after C4, unless no unrepresented branch passes its gates.
D6Tail role formula
For inspected tail candidate s, TailRole(s)=0.40U(s)+0.25G(s)+0.20S(s)+0.15Q(s), where U is unrepresented-family score, G is gate-survival score, S is support-compatibility score, and Q is P_count normalized within tail comparisons.
D7Expected-goal construction
lambda_BRA_raw=(A_BRA+D_NOR)/2=(2.3078+1.9560)/2=2.1319 and lambda_NOR_raw=(A_NOR+D_BRA)/2=(1.2189+0.3930)/2=0.8059; bounded context gates give final lambdas 2.1121 and 0.7827.
Data 1Threshold realms
Brazil vs Norway-realm rows are strong: effective rows=12.685, GF=2.304, GA=0.395. Norway vs Brazil-realm rows are usable but sparse: effective rows=4.229, GF=0.490, GA=2.385.
Data 2Thread facts
Brazil win=0.736, draw=0.187, Norway win=0.077. Brazil by 2+=0.546, Brazil by 1=0.197. BTTS=0.443 and Brazil clean sheet=0.513.
Data 3Advanced gates
Norway SDI=0.81, Brazil SDI=0.76, game-state volatility=0.60, five-goal mass=0.167. Thus the star scoring veto is active and the chaos gate is inactive.
R1First score
Select the best score in the strongest spread family. Since home_by_2_plus=0.546 is largest and P_count(2-0)=0.123367 exceeds the other clean-cushion cells, 2-0 is selected.
R2Duplicate clean-control rule
A second clean-control score is allowed only if D3 is false. D3 is true, so 1-0 cannot occupy the second slot when 2-1 has I=0.354174>I(1-0)=0.349197.
R3Draw branch
Because draw=0.187 exceeds Norway win=0.077, keep one draw/block representative. 1-1 beats 0-0 under the total gate because Over1.5=0.815.
R4Score 4
The score 4 is the best thesis-compatible unrepresented failure branch. 1-2 is selected over 0-1 because Brazil clean-sheet loss states are weak and Brazil scoring is not rejected.
L1Inclusion 2-0
2-0 is in S10, has the largest P_count, and belongs to the strongest spread branch. Therefore 2-0 is SCORE_1.
L2Inclusion 2-1
2-1 is in S10, preserves Norway scoring, and beats the duplicate clean slot by Turtle index: 0.354174>0.349197. Therefore 2-1 is SCORE_2.
L3Inclusion 1-1
1-1 is in S10, represents the draw branch, and 0-0 fails Over1.5 evidence. Therefore 1-1 is SCORE_3.
L4Inclusion 1-2
Norway win mass is too small for C4, but score 4 compares unrepresented failure branches. 1-2 has TailRole=0.912, which exceeds 0-1 at 0.575 and all represented-score tails. Therefore 1-2 is SCORE_4.
L5Comparison 3-1 vs 1-2
3-1 has higher P_count than 1-2 (0.067981>0.035785), so it would win a fourth-probability rule. It loses Turtle score 4 because TailRole(3-1)=0.330<TailRole(1-2)=0.912 after the unrepresented-family term is applied.
L6Exhaustion
1-0 is removed by the star scoring veto; 3-0, 3-1, and 4-0 lose family representation; 0-0 fails the total gate; 0-1 fails Brazil scoring; 2-2 loses draw representation; 3-2 fails the chaos gate.
TheoremFull-sweep lock
Under Definitions D1-D7, Data 1-3, and Rules R1-R4, the unique disclosed four-score card is C4={2-0,2-1,1-1,1-2}, in that order.
QEDC4 = {2-0, 2-1, 1-1}, score 4 = {1-2}Full-sweep pre-kickoff lock.
Sources and artifacts