Honest status
What changed, and why
A theory's credibility depends on its record of retractions as much as its record of hits. This is BPR's.
Current claim, September 2026
The goal is a theory of everything. The current claim is smaller. BPR 2.0 is a flavor-sector organizing framework with a proposed, untested particle sector. Its flavor formulas agree with data and were, in some cases, registered before the measurements. Its fermion mass labels are conjectural. The number of matter families is an input. Its particle sector, a gauged dihedral point group on the lattice, is a legitimate two-dimensional anyon model that has not been shown to give the strong force, chiral matter, or any spectrum. Its coherence result is conditional. It has not been peer reviewed.
Timeline
- 2025BPR 1.0. A scalar, Abelian phase field on a two-sphere boundary. Claimed to derive electromagnetism, gravity, quantum mechanics, the Standard Model constants, and the strong force from two integers. Headline claim: a theory of everything.
- April 2026Gap audit. Eight gaps triaged. Some closed (p-selection over-determination, CP phase extraction), some scoped as research problems (the GUT residual, quantum gravity), some identified as not BPR-specific. The three-generation derivation was written at this time and later found invalid.
- 2026Sealed glueball benchmark. Targets locked before any spectrum. The 1.0 boundary content provably cannot host a light parity-odd state; the 2⁺⁺/0⁺⁺ ratio fell outside the pass band; an extra light vector state appeared. Verdict: FAIL. All three proposed loopholes were tested and closed.
- August 2026BPR 2.0. The theory-of-everything claim withdrawn as premature; the goal kept. The lattice's rotation-and-mirror symmetry promoted from decoration to a gauge force (Postulate 0d, PROPOSED). Flavor formulas verified numerically unchanged under the relabeling. A Wilson-type dynamics frozen; Monte Carlo locates its phase transition; a derived coupling places the substrate in the deconfined phase, so the physical excitations would be anyons rather than confined flux loops. Benchmark v3 sealed and unopened.
- 10 September 2026Foundation repair. An independent review found two invalid derivations. The SU(3) index argument behind ℓt = 283 (the index of an ordinary SU(N) bundle on S² is zero, not three). The compact-boson proof of exactly three generations (conformal spin is an integer; no group action was constructed; fusion products were misread as descendants). Both withdrawn. All nine flavor labels relabelled conjectural; ngen = 3 relabelled as input. The condensate "coherence guaranteed" verdict downgraded to conditional. The 2.0 particle-physics bridge recorded as open with no partial credit.
The negative-findings registry
Everything determined not to work is indexed in one document so that reviewers do not have to scrape the commit history. The main entries:
| Item | Status | One-line reason |
|---|---|---|
| Theory-of-everything headline claim (1.0) | withdrawn as premature | The 1.0 particle sector failed the sealed glueball benchmark, so the claim to have reached the goal was set aside. The goal itself is unchanged; see the conditions below. |
| SU(3) color-bundle index behind ℓt = 283 | withdrawn | c₁ = 0 for an SU(N) bundle; index is zero; a U(1) twist would be an extra input. |
| Compact-boson proof of three generations | withdrawn | Integer conformal spin; no representation-theory argument; arithmetic slip. |
| Condensate "coherence guaranteed / robust" | downgraded | Classical fixed-norm minimum only; quantum vacuum not computed; T* conditional on assumed dispersion, cutoff, density, and an underived temperature. |
| BPR 2.0 particle-physics bridge | open | 2D non-Abelian anyons do not give 4D SU(3), chiral matter, anomaly cancellation, or a spectrum. |
| M4 parity-odd loop operator (2.0 benchmark machinery) | defective, superseded | The 8-link "L" was not a contiguous path (not gauge invariant) and its shape was achiral, so the operator vanished identically. Replaced by chiral polyomino loops, verified gauge invariant and parity-odd. Benchmark v3 still sealed. |
| Flavor-label operator scan | negative | Eight frozen-ingredient operators, preregistered scoring, Bonferroni-corrected: none reproduces a flavor sector. Labels stay conjectural. |
| Glueball sector, all branches | closed | Blind benchmark failed; interaction, lump, and phason loopholes each dispositioned. |
| Riemann-zeta / GUE spectral statistics | closed | No natural prime-modular operator class reproduces Wigner–Dyson statistics. |
| Casimir exponent δ ≈ 1.37 | superseded | Was a fit; replaced by derived δ = 2, and 1.37 would now refute the projection postulate. |
| Lab-scale parameters used for cosmology (pre-v0.7) | fixed | Produced absurd values for a₀, η, Tc; different physics probes different scales. |
Full registry: CLOSED_AND_DEPRECATED.md.
What would earn the claim back
The project will describe BPR as a theory of everything again when, and only when, the following are on record and reproducible:
- A particle sector that yields four-dimensional SU(3) colour, chiral Standard Model matter, and anomaly cancellation from the boundary structure.
- A pass on the sealed glueball benchmark v3, with the same targets and blindness the 1.0 sector failed.
- At least one fermion mass label derived as an eigenvalue of an explicit Hamiltonian, with its uncertainty frozen before comparison.
- The number of matter families derived rather than taken as input.
- A survived falsification test from the experiments list that standard physics did not also predict.
Until then the claim stays narrowed, and this page stays honest about the gap.
Work in progress
What is actually being done against the list above, with dates, so that "in progress" means something.
| Condition | Step taken | Outcome |
|---|---|---|
| Pass the sealed benchmark v3 | 10 Sep 2026: vectorized Monte Carlo (about 40× faster per link), longer time axis, 3-operator variational basis per channel, GEVP effective masses. Found and fixed a defective parity-odd operator in the earlier machinery. | running first production run at 8×8×16 near the transition; gate status to be recorded whichever way it falls |
| Derive one fermion label | 10 Sep 2026: preregistered scan of every operator the framework already has, fixed scoring rule. | negative no existing operator produces a sector. Next step is constructive: write down a fermionic boundary action. |
| Particle-physics bridge | Not started. | open |
| Family count derived | Only conditional route on record is the twisted-Dirac index with flux as input. | open |
Open problems, in priority order
- A particle-physics bridge. A specified fermionic action on the boundary, a map from internal modes to four-dimensional families, gauge representations and chirality for every species, and anomaly cancellation. Nothing in Path B supplies this yet.
- A derivation of any one flavor label. Pick one of the nine, derive it as an eigenvalue of an explicit Hamiltonian with every assumption exposed, freeze its uncertainty, and compare. One success would change the status of the whole sector.
- The quantum vacuum. Show whether the interacting substrate's ground state is the coherent condensate its classical analysis predicts, and derive or bound the substrate temperature.
- A consistent renormalization scheme for quark mass comparisons.
- The dimensionful anchor and the GUT-scale residual, which are shared with every unification program and are not BPR-specific.
The rules the project holds itself to
- Targets for any spectrum test are sealed before computation, and the seal is enforced mechanically by tests that fail if a target value appears in a solver.
- A failed benchmark narrows the claim; it does not get argued around.
- Every label on this site is reproducible from a function in the repository.
- Withdrawn derivations stay on record with their corrections, not deleted.