Predictions & results

The numbers, by sector

Each row is a closed form in the cell integers, its status tier, and its distance from experiment. Tiers are honest: a theorem is not a postdiction, and a suggestive pattern is not a theorem.

Tier 1 mathematical theorem Tier 2 derived given identifications Tier 3 >1.5σ tension Tier 4 suggestive / speculative

The postdictions across roughly 60 observables are under an ongoing look-elsewhere-corrected joint-χ² audit; the earlier "60+ observables, zero free parameters" headline has been retired as not defensible against a trials-factor analysis. What is pre-registered and sharp are the eleven falsifiable predictions below.

Sharpest test

δPMNS / δCKM = 3

This ratio follows from the colour factor \(C_A = 3\) alone. It is exact, parameter-free, and the PMNS CP phase is exactly what DUNE is built to measure (around 2035). If the ratio is confirmed ≠ 3 at >3σ, the framework is falsified. No other approach makes this particular prediction.

Sector 1

Electromagnetism & electroweak

ObservableFormula (cell integers)StatusDistance from experiment
α⁻¹ (fine structure)(4π)³ᐟ²·π·[47/48 + 10/(3·48³) + 22/(3·48⁵)]Tier 2137.035999055 · 0.3σ from Cs (2018)
sin²θ_W (geometric)(17 − 3√17)/20Tier 10.14%
α_s(M_Z)C_A² − C_A·ln(C_A)/(2π)Tier 20.01σ
m_H / M_Z18/(9+√17)Tier 20.14%
α_GUT⁻¹25Tier 2novel — not directly measurable
Sector 2

Lepton & neutrino masses

ObservableFormulaStatusDistance from experiment
Koide relationQ = (Σm)/(Σ√m)² = 2/3Tier 1exact (structural)
m_er₁·M_P·exp(−(E−F)(2Δ+√Δ)/16)Tier 20.002%
m_μKoide from m_e, θ = 2/9Tier 20.004%
m_τKoide from m_e, θ = 2/9Tier 20.009%
m₁ (lightest ν)= 0Tier 1exact theorem
m₃ (ν)m_e·exp(−(11+13√17)/4)Tier 20.075%
Mass hierarchynormal ordering forcedTier 1theorem
|Δm²₃₂| / Δm²₂₁= 33 (Eisenstein norm, 33 = S² − C_A·P)Tier 4≈0.5σ · pending formal derivation
Sector 3

Quark mixing (CKM / Wolfenstein)

ObservableFormulaStatusDistance from experiment
λ (Cabibbo, NLO)sin(π/14)·(1 + √17/363)Tier 20.22505 · 0.07σ
A (Wolfenstein)(F − r₁)/F = (19+√17)/28Tier 20.82583 · −0.015σ
R_b (CKM triangle)r₁²/(r₁r₂ − 1) = (49 − 9√17)/30Tier 10.39640 · 0.36σ
ρ̄R_b·cos(δ_CKM)Tier 20.15898 · −0.002σ
η̄R_b·sin(δ_NLO)Tier 30.3478 · lever-arm phase residual
δ_CKMinter-type torsion operator, (C_A−1):1Tier 266.36° · 0.25σ
m_d / m_ssin²(π/14)Tier 21.0%
Sector 4

Neutrino mixing (PMNS)

ObservableFormulaStatusDistance from experiment
tan²θ₁₂(√17/9)(1 − √17/144)Tier 20.44501 · 0.074σ
sin²θ₂₃1/2 + √17/81Tier 20.5509 · 0.2σ
sin²θ₁₃(√17/27)²(1 − √17/162)²Tier 20.02215 · 0.2σ
δ_PMNS3πR = 200.7°Tier 20.15σ (±25°)
δ_PMNS / δ_CKM= 3 (colour factor C_A)Tier 1novel — testable by DUNE ~2035
Sector 5

Gravity & cosmology

ObservableFormula / originStatusDistance from experiment
Ω_DM / Ω_b3(1 + 2√3)/2⁴ᐟ³ = 5.3147Tier 20.92% (Planck 2018)
ρ_Λ (dark energy)ρ₀(l_P/R_U)² × 6/7Tier 21.4% (Planck 2018)
r_p (proton radius)4ℏ/(m_p c)Tier 20.02%
Bekenstein area quantum k= C_A = 3Tier 1exact
Schwarzschild g_tt, g_rrcovariant vacuum density, ν = 1/2Tier 2exact match to GR
Einstein equationsunimodular, from foam actionTier 2consistent with all GR tests
Maxwell equations (all 4)□D = 0 + Helmholtz + VolterraTier 2exact
Friedmann equations (both)foam energy conservationTier 2exact
Cosmological constantΛ = integration constant (dissolved)Tier 4resolves the 10¹²³ discrepancy
Falsifiability

What would end it

The framework is rigid: changing any cell integer by ±1 moves a headline number by ≥1%, so there is nowhere to hide. Any one of these falsifies it.

  1. A confirmed dark-matter particle (LZ / XENON / ADMX) — UFFT says dark matter is lattice anisotropy.
  2. A confirmed supersymmetric partner at any energy — fourteen faces leave no room.
  3. α⁻¹ measured outside 137.035999055 ± ~4×10⁻⁹, or the rubidium value confirmed past 3σ.
  4. δ_PMNS / δ_CKM confirmed ≠ 3 at >3σ (DUNE).
  5. Vacuum structure shown not to be the truncated octahedron — it is the unique Kelvin solution, so this removes the geometric inputs entirely.
  6. The altitude-dependent decoherence ratio Γ(r)/Γ(∞) = 1 − 2GM/rc² returning a null result at the predicted magnitude.

The full eleven pre-registered predictions and all derivations are in the Core Framework. To recompute the spectrum these rest on, see Verify.