RELEASE 4.8 TEXTBOOK EDITION (EN + PL)

Version: 4.8
Date: 2026-03-03
Branch: main

ENGLISH VERSION

1) Main Breakthrough of Release 4.8

The main breakthrough is not a kernel replacement.
The breakthrough is a robust readout repair in the GW channel:

• bounded signed coupling term (kappa_t) plus monotonic compression (gamma_c),
• which removes the historical fold-2 random-null instability,
• while keeping one frozen kernel and one shared mass+flavor branch.

This is what moved the project from fragile/partial robustness to:

• BOUNDED_GW_TRIAD_LOCKABLE_PASS (QW-2001),
• then formal closure gate pass: SINGLE_KERNEL_TRIPLE_SECTOR_CLOSURE_PASS_V3 (QW-2002).

2) Did Kernel Parameters Change?

No. Kernel parameters did not change in Release 4.8.

Frozen kernel:

• omega = 0.37341399972174283
• phi = -1.310233577483508
• beta = 0.6159380564131874
• eta = 2.8

What changed was the GW operator/readout layer, not the kernel ontology.

3) Frozen Parameter Set Used in Lockable Closure

3.1 Shared mass+flavor parameters

• chi_im = 0.31277657543084514
• diag_iso = 0.011133623526631775
• diag_q_coeff = 0.023231150712776396
• diag_sector = 0.11340639008162685
• lambda_mix = 0.6558721036754207
• p_amp = 0.9252148940096931
• phase_q = 0.26672804099980674
• r_dist = 0.9758497827019121
• rho_gap = 1.0065574010654268
• theta_iso = 0.3453065240958936
• theta_sector = 0.16772593255657867

3.2 GW bounded operator parameters (repair)

• xi1 = 0.00164905739
• xi3 = 0.001712654074
• xi4 = 0.000052824299
• gamma_c = 0.796713
• kappa_t = 1.4

4) Core Formula Set (Textbook-Level)

4.1 Kernel

[ K(d)=\frac{\cos(\omega d+\phi)}{1+\beta d^{\eta}} ]

This is the single frozen structural function used across sectors.

4.2 Cyclic topological distance (mod 24)

[ d_{ij}=1+\min\left(|q_i-q_j|\bmod 24,;24-(|q_i-q_j|\bmod 24)\right) ]

4.3 Mass branch (noncircular chain)

Effective charge:

[ q_{\text{eff}}=q_{\text{base}}+s,\Delta_{\text{info}} ]

Mass law:

[ m_{\text{pred}}=m_{\text{top}}\cdot 4^{-\gamma q_{\text{eff}}/4} ]

With Release-4.8 locked branch values:

• gamma = 1.5066666666666666
• Delta_info = 0.17702715186107637
• mean mass relative error (reference branch): 12.0508%

4.4 Flavor Hamiltonian (shared operator)

Base amplitude:

[ \text{base}{ij}=\operatorname{sgn}(K{ij}),|K_{ij}|^{p_{amp}},d_{ij}^{r_{dist}} ]

Near-gap damping:

[ \text{near}{ij}=e^{-\rho{gap}|i-j|} ]

Phase:

[ \Phi_{ij}=\phi_q(q_i-q_j)+\theta_{iso},\sigma_{iso}(i-j)+\theta_{sector},\sigma_{sec}(i-j) ]

Complex coupling:

[ S_{ij}=\lambda_{mix},\text{base}{ij},\text{near}{ij} ]

Hermitian Hamiltonian form:

[ H=\frac{1}{2}(H+H^{\dagger})=\operatorname{Re}+i,\chi_{im},\operatorname{Im}_{asym}+\operatorname{Diag} ]

Mixing matrices:

[ V_{CKM}=|U_u^{\dagger}U_d|,\qquad V_{PMNS}=|U_\nu^{\dagger}U_\ell| ]

4.5 GW baseline score (before bounded repair)

Kernel-derived weights:

[ w_n=\frac{|K(n)|^{p_{amp}}n^{r_{dist}}}{\sum_{k=1}^{4}|K(k)|^{p_{amp}}k^{r_{dist}}} ]

Score:

[ S_{base}=w_1 f_{max}+w_2 f_{mean}+w_3 f_{0ms}+w_4 f_{10ms} ]

4.6 Signed coupling channels (GW structural terms)

With pair_sign = +1 (H1-V1), -1 (L1-V1), 0 (H1-L1), and

[ \tau = \text{best_lag_ms}\cdot 10^{-3} ]

a) first channel:

[ c_1\propto pair_sign\cdot\sin(\omega\tau+\phi)\cdot(corr_{0}-corr_{10}) ]

b) second channel:

[ c_3\propto pair_sign\cdot\cos(\omega\tau+\phi)\cdot(mean_abs+|corr_0|+|corr_{10}|) ]

c) third channel:

[ c_4\propto pair_sign\cdot\sin(\omega\tau+\phi)\cos(\omega\tau+\phi)\cdot(|corr_0|+|corr_{10}|+mean_abs) ]

Each channel is standardized by its own standard deviation.

4.7 Release-4.8 breakthrough operator (bounded coupling)

Signed combination:

[ t=\xi_1 c_1+\xi_3 c_3+\xi_4 c_4 ]

Bounded/saturated term:

[ t_{eff}=\operatorname{clip}\big(t,-\kappa_t\sigma_t,+\kappa_t\sigma_t\big) ]

Raw score:

[ S_{raw}=S_{base}+t_{eff} ]

Monotonic compression:

[ z=\frac{S_{raw}-\operatorname{median}(S_{raw})}{\sigma_{S}},\qquad S=\operatorname{median}(S_{raw})+\sigma_S,\operatorname{sgn}(z)|z|^{\gamma_c} ]

4.8 Statistical validation formulas

Control threshold:

[ q_{90}=Q_{0.9}(S_{control}) ]

GW diagnostics:

[ ADV=P(S_{HL}>q_{90})-P(S_{control}>q_{90}) ]

[ SEP=\operatorname{median}(S_{HL})-\operatorname{median}(S_{control}) ]

[ GAP=|\operatorname{median}(S_{HV})-\operatorname{median}(S_{LV})| ]

Triad bootstrap pass-rate:

[ R_{boot}=\frac{#{\text{bootstrap samples passing all flags}}}{N_{boot}} ]

5) Thresholds Used in Closure Gates

• CKM mean rel error <= 15%
• PMNS mean rel error <= 15%
• GW SEP >= 0.002
• GW ADV >= 0.30
• GW AUC >= 0.75
• GW control gap <= 0.0025

Mass branch constraints (fixed from source branch):

• mass mean rel error <= 15%
• mass max rel error <= 35%
• tau/charm ratio rel error <= 20%

6) Breakthrough Timeline (Key Gates)

1. QW-1967 ISOSPIN_LOCAL_REFINEMENT_PASS
   first deterministic triad pass in this branch.
2. QW-1968 FRAGILE_PASS_NOT_YET_LOCKABLE
   pass existed but robustness volume was weak.
3. QW-1969 INSUFFICIENT_BOOTSTRAP_ROBUSTNESS
   bootstrap lockability still insufficient.
4. QW-1970 STRUCTURAL_CONTROL_TERM_INSUFFICIENT
   improved but still not lockable (boot5000=0.7994).
5. QW-1999 BOUNDED_COUPLING_ROBUST_HARD_PASS
   bounded coupling removed hard null-tail instability.
6. QW-2000 BOUNDED_COUPLING_DEEP_AUDIT_PASS
   deep audit confirmed robustness (null_random_p90_mean_max=31.33%).
7. QW-2001 BOUNDED_GW_TRIAD_LOCKABLE_PASS
   full lockability achieved (boot2500/5000/10000 = 1.0).
8. QW-2002 SINGLE_KERNEL_TRIPLE_SECTOR_CLOSURE_PASS_V3
   Stage-C internal closure status reached.
9. QW-2003 FROZEN_LOCKABLE_PACKAGE_READY
   frozen external package exported with SHA256.

7) Final Internal Status (Release 4.8)

• Single frozen kernel: preserved.
• Single shared mass+flavor branch: preserved.
• GW channel robustness: repaired and lockable.
• Stage-C status: internally closed (TOE_STAGE_C_SINGLE_KERNEL_CLOSED_LOCKABLE_INTERNAL).

What remains:

• external, independent, blind confirmatory execution on frozen package.

WERSJA POLSKA

1) Główny przełom Release 4.8

Głównym przełomem nie była zmiana kernela.
Przełomem była naprawa operatora odczytu w kanale GW:

• ograniczenie (saturacja) sprzężenia przez kappa_t,
• zachowanie monotonicznej kompresji przez gamma_c,
• usunięcie historycznej niestabilności fold-2 random-null,
• przy zachowaniu jednego zamrożonego kernela i jednej wspólnej gałęzi masa+flavor.

To właśnie podniosło system do:

• BOUNDED_GW_TRIAD_LOCKABLE_PASS (QW-2001),
• a następnie do formalnego zamknięcia bramki Stage-C: SINGLE_KERNEL_TRIPLE_SECTOR_CLOSURE_PASS_V3 (QW-2002).

2) Czy wartości parametrów kernela się zmieniły?

Nie. Parametry kernela w Release 4.8 się nie zmieniły.

Kernel zamrożony:

• omega = 0.37341399972174283
• phi = -1.310233577483508
• beta = 0.6159380564131874
• eta = 2.8

Zmianie uległa warstwa operatora GW (odczyt/statystyka), a nie rdzeń kernela.

3) Zamrożony zestaw parametrów użyty do domknięcia

3.1 Wspólne parametry masa+flavor

• chi_im = 0.31277657543084514
• diag_iso = 0.011133623526631775
• diag_q_coeff = 0.023231150712776396
• diag_sector = 0.11340639008162685
• lambda_mix = 0.6558721036754207
• p_amp = 0.9252148940096931
• phase_q = 0.26672804099980674
• r_dist = 0.9758497827019121
• rho_gap = 1.0065574010654268
• theta_iso = 0.3453065240958936
• theta_sector = 0.16772593255657867

3.2 Parametry operatora GW (naprawa)

• xi1 = 0.00164905739
• xi3 = 0.001712654074
• xi4 = 0.000052824299
• gamma_c = 0.796713
• kappa_t = 1.4

4) Kluczowe wzory (poziom podręcznikowy)

4.1 Kernel

[ K(d)=\frac{\cos(\omega d+\phi)}{1+\beta d^{\eta}} ]

To jedna wspólna funkcja strukturalna używana we wszystkich sektorach.

4.2 Cykliczna odległość topologiczna (mod 24)

[ d_{ij}=1+\min\left(|q_i-q_j|\bmod 24,;24-(|q_i-q_j|\bmod 24)\right) ]

4.3 Gałąź masowa (łańcuch niekołowy)

Ładunek efektywny:

[ q_{\text{eff}}=q_{\text{base}}+s,\Delta_{\text{info}} ]

Prawo masy:

[ m_{\text{pred}}=m_{\text{top}}\cdot 4^{-\gamma q_{\text{eff}}/4} ]

Wartości gałęzi użytej jako źródło:

• gamma = 1.5066666666666666
• Delta_info = 0.17702715186107637
• średni błąd względny mas: 12.0508%

4.4 Hamiltonian flavor (wspólny operator)

Amplituda bazowa:

[ \text{base}{ij}=\operatorname{sgn}(K{ij}),|K_{ij}|^{p_{amp}},d_{ij}^{r_{dist}} ]

Tłumienie odległości pokoleniowej:

[ \text{near}{ij}=e^{-\rho{gap}|i-j|} ]

Faza:

[ \Phi_{ij}=\phi_q(q_i-q_j)+\theta_{iso},\sigma_{iso}(i-j)+\theta_{sector},\sigma_{sec}(i-j) ]

Sprzężenie zespolone:

[ S_{ij}=\lambda_{mix},\text{base}{ij},\text{near}{ij} ]

Postać hermitowska:

[ H=\frac{1}{2}(H+H^{\dagger})=\operatorname{Re}+i,\chi_{im},\operatorname{Im}_{asym}+\operatorname{Diag} ]

Macierze mieszania:

[ V_{CKM}=|U_u^{\dagger}U_d|,\qquad V_{PMNS}=|U_\nu^{\dagger}U_\ell| ]

4.5 Bazowy score GW (przed naprawą)

Wagi z kernela:

[ w_n=\frac{|K(n)|^{p_{amp}}n^{r_{dist}}}{\sum_{k=1}^{4}|K(k)|^{p_{amp}}k^{r_{dist}}} ]

Score:

[ S_{base}=w_1 f_{max}+w_2 f_{mean}+w_3 f_{0ms}+w_4 f_{10ms} ]

4.6 Kanały sprzężeń podpisanych

Dla pair_sign: +1 (H1-V1), -1 (L1-V1), 0 (H1-L1), oraz

[ \tau = \text{best_lag_ms}\cdot 10^{-3} ]

mamy:

[ c_1\propto pair_sign\cdot\sin(\omega\tau+\phi)\cdot(corr_{0}-corr_{10}) ]

[ c_3\propto pair_sign\cdot\cos(\omega\tau+\phi)\cdot(mean_abs+|corr_0|+|corr_{10}|) ]

[ c_4\propto pair_sign\cdot\sin(\omega\tau+\phi)\cos(\omega\tau+\phi)\cdot(|corr_0|+|corr_{10}|+mean_abs) ]

Każdy kanał jest standaryzowany przez swoje odchylenie standardowe.

4.7 Wzór przełomowy Release 4.8 (bounded coupling)

Kombinacja podpisana:

[ t=\xi_1 c_1+\xi_3 c_3+\xi_4 c_4 ]

Ograniczenie amplitudy:

[ t_{eff}=\operatorname{clip}\big(t,-\kappa_t\sigma_t,+\kappa_t\sigma_t\big) ]

Score surowy:

[ S_{raw}=S_{base}+t_{eff} ]

Kompresja monotoniczna:

[ z=\frac{S_{raw}-\operatorname{median}(S_{raw})}{\sigma_{S}},\qquad S=\operatorname{median}(S_{raw})+\sigma_S,\operatorname{sgn}(z)|z|^{\gamma_c} ]

4.8 Wzory testów statystycznych

Próg kontrolny:

[ q_{90}=Q_{0.9}(S_{control}) ]

Diagnostyki GW:

[ ADV=P(S_{HL}>q_{90})-P(S_{control}>q_{90}) ]

[ SEP=\operatorname{median}(S_{HL})-\operatorname{median}(S_{control}) ]

[ GAP=|\operatorname{median}(S_{HV})-\operatorname{median}(S_{LV})| ]

Bootstrap triady:

[ R_{boot}=\frac{#{\text{próbki bootstrap spełniające wszystkie flagi}}}{N_{boot}} ]

5) Progi użyte przy domykaniu

• CKM mean rel error <= 15%
• PMNS mean rel error <= 15%
• GW SEP >= 0.002
• GW ADV >= 0.30
• GW AUC >= 0.75
• GW control gap <= 0.0025

Dla gałęzi masowej (stałej):

• mass mean rel error <= 15%
• mass max rel error <= 35%
• tau/charm ratio rel error <= 20%

6) Chronologia przełomów (kluczowe bramki)

1. QW-1967 ISOSPIN_LOCAL_REFINEMENT_PASS
   pierwszy deterministyczny pass triady na tej gałęzi.
2. QW-1968 FRAGILE_PASS_NOT_YET_LOCKABLE
   pass był kruchy pod rygorem odporności.
3. QW-1969 INSUFFICIENT_BOOTSTRAP_ROBUSTNESS
   bootstrap lockability nadal za słaba.
4. QW-1970 STRUCTURAL_CONTROL_TERM_INSUFFICIENT
   poprawa, ale nadal bez lockability (boot5000=0.7994).
5. QW-1999 BOUNDED_COUPLING_ROBUST_HARD_PASS
   naprawa ogona null-tail w kanale GW.
6. QW-2000 BOUNDED_COUPLING_DEEP_AUDIT_PASS
   deep-audit potwierdził stabilność (null_random_p90_mean_max=31.33%).
7. QW-2001 BOUNDED_GW_TRIAD_LOCKABLE_PASS
   pełna lockability (bootstrap 2500/5000/10000 = 1.0).
8. QW-2002 SINGLE_KERNEL_TRIPLE_SECTOR_CLOSURE_PASS_V3
   formalne zamknięcie Stage-C (wewnętrzne).
9. QW-2003 FROZEN_LOCKABLE_PACKAGE_READY
   zamrożony pakiet z hash: f5123046189f7f137a0f2cd2c715eea424d230e2352e75f6e80c483b8f069c02.

7) Końcowy status wewnętrzny (Release 4.8)

• Jeden zamrożony kernel: zachowany.
• Jedna wspólna gałąź masa+flavor: zachowana.
• Kanał GW: naprawiony i lockable.
• Status Stage-C: TOE_STAGE_C_SINGLE_KERNEL_CLOSED_LOCKABLE_INTERNAL.

Co jeszcze pozostaje:

• niezależny, zewnętrzny, ślepy confirmatory run na zamrożonym pakiecie.