Falsifiable predictions: the TQNT 2026 preregistration
The citable document
These predictions are frozen, timestamped and citable in CA9-WP-2026-003, “Pre-registration of dated predictions” (PDF · doi:10.5281/zenodo.22282250), published in the Cahiers de physique théorique CA9 working paper series. Any prediction missed within its window counts against the theory.
This register freezes, before measurement, TQNT's dated quantitative predictions with their refutation conditions. Any prediction missed within its window counts against the theory, with no requalification possible.
✓ The declared antecedent
One prediction of the same programme is already confirmed: Ωcc ~ 3,727 MeV (LHCb, 06/2026), within the predicted range [3,715, 3,788] MeV. Honest chronology: the announcement preceded the final consolidation of the scale — a blind derivation in practice, not a formal preregistration. Hence precisely this register for what follows.
The confrontation register — verdicts to date
Each line points to a prediction deposited with a date and a DOI; the verdict is updated when the measurement lands, in either direction.
| Prediction | Deposited | Verdict |
|---|---|---|
| Ωcc ∈ [3,715, 3,788] MeV | 07/2026 (blind derivation, declared chronology) | ✓ verified — LHCb 06/2026: ~3,727 MeV |
| S₈ = Planck's σ₈ (the tension must regress) | 27/07/2026 | ✓ ongoing, favourable — KiDS-Legacy 2025: 0.74σ from Planck |
| Silence of the freeze GW background (Ω ~ 10⁻⁴⁴) | 07/2026 | ✓ verified — NANOGrav = astrophysics (the predicted absence) |
| H₀: the calibrators must converge downwards | 27/07/2026 | ongoing, favourable — JWST TRGB/JAGB: 68.4 ± 1.7; next milestone Gaia DR4 (late 2026) |
| w = −1 (DESI's w(z) preference must regress) | 27/07/2026 | pending — DESI DR3 |
| Ξcc* = 3,659 MeV (vs ~3,690 lattice/QM) | 04/08/2026 | pending — LHCb Run 3 |
| Forward p-O: neutron ratio p-O/pp = 1.9 ± 0.15 | 03/09/2026 — before any LHCf publication | pending — LHCf p-O (07/2025 data) |
| H₀(standard sirens) → 67.4 ± 1 | 02/09/2026 | pending — O5 (~2028) |
| r ≤ 0.01 | 27/07/2026 | pending — LiteBIRD (~2032) |
The owned refutations — the register cuts both ways
The programme records its failures in the same register as its successes: F1 (pure topological inflation: ns refuted — closed by the Starobinsky realization); F2 (V5.16 reheating: BBN tension — closed in 07/2026 by the exact w = 1/3 cascade); F3 (class-E γ lines at the historical energies: excluded by Fermi-LAT — lifted in 07/2026 after the canonical recomputation; the six true lines are 1.07–1.64 GeV, listed below); the lithium-7 "ashes" mechanism (refuted by the framework itself at the efficiency wall); and several application concepts (refuted by the framework's own theorems). A missed prediction stays on the register — no requalification.
1 · Hadronic spectroscopy (LHC, Run 3+)
| Prediction | TQNT value | Refuted if |
|---|---|---|
| Ξcc* (J = 3/2 partner) | 3,659 MeV | measured ~3,690 (lattice/QM) |
| Ωb* | 6,067 MeV (+22 vs Ωb) | outside ± a few MeV |
| Ωb(1P, j = 0, J = ½) — the quartet's 5th | 6,297–6,304 MeV | absence or outside the window |
| B₀*/B₁' (bottom jK = 1/2 doublet) | centroid 5,665 MeV (members ~5,660/5,666, split 6.6 MeV) | 5,591 or 5,517 (−½/−1 readings); ~5,700–5,750 (other models) |
| Σb(1P) quintet | j₀ 6,076; j₁ 6,092/6,102; j₂ 6,111/6,128 MeV | j₀/j₂ levels absent in Λbπ |
| Σc(1P) triplet | 2,772 / 2,821 / 2,869 MeV (± a few MeV) | outside the windows |
| Ξb'(1P, j = 0) | ~6,243 MeV | — (declared Ξb(6227) reservation) |
| Bottom 1D doublet | 5,903 / 5,972 MeV (gap ~69) | BJ(5970) ≠ 1D top (JP ≠ 3⁻) |
| Ωccc | 4,782 MeV (strict) | clear deviation |
| Ωbbb | outside the framework (negative binding expected) | standard bound state found |
| Hcc⁺⁺ (0⁺ ground state, the fused hexaknot) | 4,497 ± 10 MeV, BOUND (cτ ~ 30 μm); channels Λc⁺pK⁻π⁺, Λc⁺Λπ⁺ | no displaced-vertex signal in the listed channels |
| Hcc excited 1⁺ | 4,572 ± 10 (pinned at the ΛcΛc threshold) | any peak > 4,573 MeV refutes |
| Hbb⁰ (0⁺) | 11,212 ± 10 MeV, bound (cτ ~ 220 μm) | — |
| Light H (uuddss dibaryon) | ABSENT (2,481, +250 above threshold) | any detection |
| Strange-baryon orbital bands | shift s = 0.358(2) + full dilution | s = 0.376 (ΔWr/3) |
| Δ(2950) (JP = 15/2⁺, 2-star) — added 2026-08-04 | true mass ~2,850 MeV (inverse Regge: 15/2 exact at 2,849; PDG 2,950 ± 100 from old analyses) | mass confirmed at 2,950 ± 30 with JP = 15/2⁺ in modern partial waves |
| jK = 3/2 splitting | κ = 0.0650 GeV² universal (×0.813 strange) | non-universality |
| Z/H (asthenions) | exact degeneracy of the mirror doublets | any splitting (= EW CP-odd signature) |
2 · Families, neutrinos, weak currents
| Prediction | Refuted if |
|---|---|
| Never a 4th family (derived Z₃ saturation) | any 4th generation |
| No sterile neutrino, at any mass | any confirmation (MicroBooNE, DUNE, KATRIN…) |
| 0νββ: rate = 0 exactly (Dirac neutrinos) | any 0νββ signal |
| Σmν ≈ 66 meV | very different cosmological value |
| Charged-current saturation: BRlep/mode ≥ ~8% | clear violation |
3 · Dark matter (the eremnions)
| Prediction | Refuted if |
|---|---|
| Absence in standard direct detection (> 5 GeV, LZ/XENONnT) | standard WIMP detection |
| Absence also at 1.7–2.2 GeV (DS-20k, XLZD): σeq ~ 10⁻⁴⁸ cm² | any detection in this window |
| 6 narrow γ lines at 1.07/1.10/1.22/1.35/1.47/1.64 GeV (~1 dex below the 2015 limits; deep Fermi-LAT target) | smeared continuum at these energies (= tension); absence at ×10 sensitivities |
| No BH post-merger regenerated dark matter (P22 WKB) | residual post-merger signal |
4 · Cosmology and gravitation
| Prediction | Refuted if |
|---|---|
| Ωtot = 1 exact (flatness theorem) | any confirmed curvature |
| w = −1 constant (dw/dz = 0 by construction) | w(z) confirmed (the DESI preference must regress) |
| H₀ and S₈ tensions: must regress (systematics/local void in GR) | new physics confirmed |
| Large-scale CMB anomalies: no cross-replication (E modes, lensing, 21 cm) | any cross-replication |
| Standard sirens at low z → global H₀ | local amplification confirmed |
| Metrology of G: downward convergence (Gpredicted = 6.6710×10⁻¹¹) | consolidation at 6.674 (refutes at 5×10⁻⁴) |
| No GRB dispersion (EQG > 9.3 EP protected) | dispersion detected |
| No drift of the constants (c, α: integer lock k = 135) | confirmed drift |
| No primordial anti-helium (co-localized mirror pairs, universal CP crumb — v3 addition, 2026-09-04) | any confirmed primordial anti-He (AMS-02 candidates) |
| r ≤ 0.01 (Starobinsky) | large r measured (LiteBIRD) |
5 · Forward p-O at the LHC (LHCf) — deposited 2026-09-03, before any publication
The July 2025 p-O run reproduces the first interaction of an air shower; LHCf's forward spectra are unpublished at deposit time. Derived chain, zero fitted number (leading exchange rate q = (2−√2)/(2(3−√2)) = 0.185).
| Prediction | Value | Refuted if |
|---|---|---|
| Pn(pp) — probability that the leading baryon is a neutron | 0.185 (closed form) | outside [0.14, 0.23] |
| Pn(p-O) — per-contact compounding over wounded nucleons | 0.36 ± 0.03 | outside [0.30, 0.42] |
| Ratio Pn(p-O)/Pn(pp) (acceptances cancel) | 1.9 ± 0.15 | ≈ 1 (no compounding) or outside [1.6, 2.2] |
| Neutron softening ⟨1−kn⟩(p-O) | > 0.60 (vs 0.536 measured in pp) | ≤ 0.536 |
| Forward photon energy fraction | reduced vs pre-2025 generators | consistent with pre-2025 generators |
6 · The internal null test (B8) — v3 addition, 2026-09-04
A rare case: the depositor pre-registers a null result on his own device — if the framework is right, it predicts the failure of its own former patent.
| Prediction | Refuted if |
|---|---|
| T₂(trefoil quantum dot)/T₂(rectangle) = 1 on the same wafer (± ~30% systematics) | reproducible ratio > 10 (would refute indiscernibility) |
| No dependence of T₂ on the etched-pattern topology | any reproducible topological dependence |
Frozen source document (dated versions, sources included): DOI 10.5281/zenodo.22282250 · all publications
⚖ Accountability conditions
- The values are frozen at the deposit date (2026-07-27; Δ(2950) added 2026-08-04).
- A prediction missed within its window counts against the theory and will be recorded in the monograph's falsification register (F1–F3, whose history shows the practice).
- Absence predictions are refutable by any confirmed detection.
- Any update will go through a new timestamped version, without withdrawal of the present one.
🔖 Reference
TQNT 2026 preregistration (FR + EN, as of 2026-07-27), backed by the TQNT V6 monograph (From nothing to particles, DOI 10.5281/zenodo.19131538) and its scripts.