2 anchors → 19+ outputs Hardest row resolved: +0.058σ
Test 1 — Two anchors fix nineteen-plus (the quark, lepton, and neutrino sector)
Standard Model flavor physics — the masses and mixings of the quarks, charged leptons, and neutrinos — normally takes roughly a dozen and a half independently-inserted numbers. Here, two measured anchors — the top Yukawa and |Vus| — go in, and nineteen-plus independent flavor observables come back out of the frozen geometry: |Vcb| lands 0.005σ from experiment, the Jarlskog CP invariant 0.21σ, and the electron, muon, and tau masses 0.01–0.07σ with zero lepton inputs. A handful in, twenty-plus out.
What was done
The internal geometry carries a fixed "flavor chamber" — a set of frozen operators built from the geometry's own structure (its modular fixed point, its sector projectors, its Wilson-line data), not a free 3×3 matrix per particle sector. Before any measured value is read, this chamber already has almost everything shaped; it has exactly two open blanks: an overall up-sector mass scale, and a single mixing angle.
Two measured numbers are declared, in order, to fill those two blanks:
| Step | Anchor read | Value | What it fixes |
|---|---|---|---|
| Fix 1 — the scale | Top-quark Yukawa coupling yt(MZ) | 0.9665 | The up-sector normalization; once pinned, the charm and up masses are forced through fixed geometric ratios, not fit independently. The down-sector scale then arrives from the geometry itself, from a separate frozen calculation — not from a third measurement. |
| Fix 2 — the angle | CKM mixing element |Vus| | 0.22436 | The chamber's single mixing angle; once pinned, every other quark-mixing magnitude (|Vcb|, |Vub|, |Vtd|, |Vts|, |Vtb|, |Vcd|, |Vcs|, |Vud|) is forced, along with the CP-violating phase. |
After these two fixes, the freeze record closes: nothing in the flavor pipeline is adjustable afterward. Every other flavor quantity reported is a frozen output, not a further input.
Independent?
The discipline enforced here is over-determination, openly counted: two declared calibration inputs, nineteen-plus independent outputs — the arithmetic runs the wrong way for a fit. The chamber's operators are frozen by the geometry before any comparison to data; per-entry tuning is formally forbidden. Two of the most striking blocks (the charged-lepton masses and every neutrino observable) use zero further anchors — eight further observables including all three neutrino mixing angles and a leptonic CP phase come out with no additional measured input whatsoever.
The count also survives the strictest version of itself. Six to eight more outputs follow from the program's second calibration pair, {MPl, αi}. Counted strictly across the whole construction — every measured or fitted real the machinery consumes, roughly 12–14 in all, against roughly 22 independent full-precision outputs — the ratio is a labeled ~4× over-determination. That strict metric is always stated alongside the headline, never in place of it, and it is over-determination given declared anchors, not derivation from nothing: the anchors are named exactly as physics names the meter and the second, and the predictions live in the dimensionless ratios and patterns the anchors do not contain.
Honest result
| Sector | Independent frozen outputs | Calibration inputs used |
|---|---|---|
| Quark, within-sector (up & down) | mt/mc, mc/mu, mb/ms, ms/md | — |
| Quark, between-sector | |yt/yb|(MZ) | — |
| Quark mixing (CKM) | |Vcb|, |Vub|, |Vcd|, |Vcs|, |Vtd|, |Vts|, |Vtb| | — |
| Quark CP violation | δCKM / Jarlskog invariant JCKM | — |
| Charged leptons | me, mμ, mτ | — |
| Neutrinos | two mass-squared splittings, three PMNS mixing angles, leptonic CP phase | — |
| Total | 19 or more independent outputs | 2 (yt, |Vus|) |
Selected agreements: |Vcb| comes out 0.0408 against a measured 0.04079 — a 0.005σ pull, on a quantity the mechanism had no freedom to aim at; |Vub| comes out 0.00378 against 0.00382; the Jarlskog invariant comes out 2.92×10⁻⁵ against a measured 3.00×10⁻⁵ (0.21σ); the CKM CP-violating phase comes out at +60.0° (from the geometry's raw holonomy of −120°, aligned to the standard Wolfenstein convention) against a measured 65.5° ± 1.5°, a sub-1σ pull at the declared ~10% structural precision. The leptonic CP phase comes out around 260.2°, inside the NuFIT global-fit band of 195°–270°. The charged-lepton masses land at 0.01σ (mτ), 0.02σ (mμ), and 0.07σ (me) — among the tightest agreements anywhere in the ledger, on quantities measured to many decimals, with zero lepton inputs.
The hardest row — published, then resolved. The same rigid mass-ratio ladder first forced the up-quark to ~3.16 MeV against a measured 1.27 ± 0.43 MeV — a ~4.4σ miss the program printed on its own front pages rather than tuning away (there is no separate up-quark anchor to tune). It was then resolved target-blind: the full 13-dimensional three-layer transport supplies a dimensionless factor 1/√6 = 1/√|S₃|, fixed purely by the order of the six-element Weyl symmetry group of the flavor shape — machine-checked to have used only that group order, with negative controls directly rejecting the alternative that a ~0.40 was fitted — giving mu = 1.2948 MeV, a pull of +0.058σ. The factor only exists in the full 13-D transport; the earlier 4-D shadow calculation was both harder and wrong. The row now stands as a sharp falsifiable prediction: a tighter up-quark measurement either agrees or kills the frozen shape here. Proven: the SG-8 dossier, with the machine checks and negative controls in full.
Data sources
- yt(MZ) = 0.9665 and |Vus| = 0.22436 — Particle Data Group (PDG) values, declared and hashed before comparison.
- Comparison values for all nineteen-plus outputs (quark mass ratios, CKM matrix elements, the CP-violating phase, charged-lepton masses, neutrino mass-squared splittings and mixing angles) — PDG world averages; the neutrino sector compared against the NuFIT global-fit bands.
- The flavor gate of record: the SG-8 dossier — the chamber construction, the two-anchor fixing, the full per-observable certificate with pulls disclosed row by row, and the up-quark resolution with its target-blind machine checks.
- The full particle ledger: /particles/ — every derived mass, mixing, and phase with its pull, alongside the quantum-number closure of the complete PDG catalog.
What this does not show
Two measured numbers are openly used as calibration inputs; the claim is over-determination given those declared anchors, not a zero-input derivation — the absolute scales are charged to measured rulers, and the genuine predictions are the dimensionless ratios and patterns. The strongest objection a reviewer can raise is exactly the right one: is the flavor chamber a real structure, or compressed curve-fitting under another name? The record answers three ways: the chamber is frozen before comparison and nothing is adjustable after the two fixes; the strict whole-construction count (~4×, stated above) is published rather than hidden; and the up-quark row behaved the way only a real structure does — a fit tunes its misses away quietly, while this one was published at ~4.4σ, then resolved by a factor the symmetry alone supplies, and now stands exposed to the next measurement. Don't take our word for any of it: open the SG-8 dossier and the particle ledger and check the pulls row by row.