A theory that cannot fail explains nothing. This page lists every output of the TFPT 4.5 series together with the experiment or structural argument that would kill the construction. The no-knobs audit underneath records what the theory may consume as input and what it may not.
Each output is killed by any single satisfied criterion. The construction does not survive on most outputs being right.
Carrier theorem, joint discrete solve, electromagnetic closure, strong-CP, Higgs sector, axion haloscope, cosmic birefringence, EHT residual intercept, and the cosmology comparison surface — each row is sufficient on its own to falsify the construction.
The 3+2 carrier split is not derivable without importing Standard-Model representation data — i.e. dim E_+ = 2 or dim E_- = 3 turn out to require an SM-side input rather than the compact Higgs index and the primitive Yukawa type.
An alternative admissible discrete solution survives all primitive constraints (boundary kernel, Higgs index, primitive Yukawa, seam normalization). The closed branch would no longer be unique.
The carrier-form closure equation F_U(1)(α) = 0 fails to admit a unique positive root, or the root differs from CODATA 2022 by more than the declared interface uncertainty (currently ≈ 4 × 10⁻⁸ in α⁻¹).
A stable nonzero hadronic EDM signal under the declared comparison convention. The admissibility/determinant-line argument forces θ_eff = 0 at theorem level, so any robust EDM detection kills the construction.
Robust discovery of a second light seam-even Higgs doublet. The compact bosonic index forces N_Φ = 1; an additional doublet kills the determinant class.
Calibrated haloscope exclusion of the 15.764 GHz ± 50 MHz window at the coupled sensitivity. A clean exclusion of the predicted band kills the cosmology readout chain.
Externally calibrated cosmic-rotation analysis statistically consistent with β = 0 within ±0.05° (without EB self-nulling). Determinant-line response would not be in the data.
Calibrated achromatic residual χ₀^res(x) = χ₀^obs − χ₀^GRMHD statistically consistent with zero across the horizon-scale image, or no 1/r² profile, or no E·B sign flip, or measurable λ² dependence.
Robust inconsistency between the closed-branch readouts (Λ_IR, η_B, Σ m_ν, Ω_b) and the comparison data under the declared convention, beyond stated interface uncertainty.
An assumption first becomes invisible — i.e. an empirical input enters the proof chain without being declared as such, or a downstream comparison feeds back into the primitive branch.
A claim of 'no fitted constants' is only as strong as the audit table behind it. For each TFPT output, the table below records the inputs the construction may use, the inputs it explicitly may not use, and the number of free parameters available for absorption. The free-knob count is the bar to clear.
| Output | Inputs allowed | Inputs forbidden | Free knobs |
|---|---|---|---|
| α⁻¹(0) | Primitive kernel (Paper 1), carrier packet (Paper 2), exact seam opening φ_seam(α) | Fitting against CODATA / atom-recoil values; freezing φ_seam at φ₀ inside the root equation | 0 |
| λ_C (Cabibbo angle) | Retained seed φ₀ from the joint discrete solve | Global CKM-fit residual absorbed into λ_C | 0 |
| sin²θ₁₃ (reactor angle) | Retained seed φ₀ + γ = 5/6 | Oscillation-fit central value pre-loaded as input | 0 |
| δ_CKM (CP phase) | Holonomy transport on the rigid branch; lower critical point of the cusp cubic | Adjustable CP dial in the transport sector | 0 |
| N_Φ = 1 (Higgs index) | Compact bosonic index on the seam-even line bundle | Observed Higgs count used as primitive input | 0 |
| θ_eff = 0 (strong-CP null) | Admissibility selector P_adm = P_prim · P_sing · P_Θ, determinant-line phase | Tuned θ-phase, hidden flavor-side cancellation | 0 |
| β = 0.2424° (cosmic birefringence) | Determinant-line / Chern–Simons response, retained seed φ₀ | Calibration absorbed into the predicted angle | 0 |
| β_BH(r) (achromatic dyonic intercept) | Coupling 1/(256π⁴) = 16 c₃⁴ from the same admissibility data as α; geometric weights from the GRMHD source model | Free coupling rescaling; spatial profile fitted to the residual map | 0 (coupling) · model-dependent (geometry, emission radius) |
| ξ = c₃ / φ₀ (Einstein-limit normalizer) | c₃ = 1/(8π) (Paper 1), φ₀ from the retained seed | Deriving SI G_N from two dimensionless numbers | 0 |
| ν_a ≈ 15.764 GHz (axion haloscope) | Seam transfer, determinant-line phase, downstream cosmology budget | Coupling rescaling to fit a haloscope window | 0 (coupling) · model-dependent (cosmology) |