Honest Assessment

σ_int = 0.086 ± 0.003 dex is not a BTFR scatter: it is independent of V (shuffling V across galaxies gives 0.0870 vs 0.0864; explorer 2026-10-23), the residual of an in-sample photometric mass model, so the CDM question below has no object (corrected 2026-10-24). The earlier reading: equal to the source session's own CDM figure (0.085; z = +0.5 in the definitive run), not below CDM. Whether it is consistent with published ΛCDM predictions is unsettled: the first external benchmark checked (Desmond 2017, abundance matching, ~0.25 dex in baryonic mass for SPARC-like samples) is about 3× larger, so the CDM-consistency verdict was suspended on 2026-10-09 and withdrawn on 2026-10-24 (see CDM Discrimination). Sample correction (2026-07-10): the 0.086 figure belongs to the source session's optimal quality cut, N = 677 (SNR > 15, e_W50 < 10, b/a < 0.65, V > 80 km/s), not to the full ALFALFA–SDSS cross-match — the full sample (N = 14,435 in the definitive session) gives σ_int = 0.118 ± 0.001. This card previously paired the headline N with the optimal-cut statistic — the same numerator/denominator splice class as the TEST-03 correction below, caught by walking the number to Session 610's own table.

Correction (2026-07-09): this card previously reported “environment-dependent RAR scatter at p = 5×10⁻⁶, R² = 0.14” as the ALFALFA-SDSS (N = 14,585) result and said its kill criterion fired. That pairing is mathematically impossible at N = 14,585 (R² = 0.14 there implies p of order 10⁻⁵⁰⁰, not 10⁻⁶) — the (R², p) pair is only self-consistent at SPARC scale (N ≈ 130–175), which is TEST-05's sample, not this one. TEST-03's actual environment-density result on the 14,585-galaxy cross-match was never computed as registered. See the corrected TEST-03 and TEST-05 cards for the full provenance trace.

Verdict correction (2026-07-04): an intermediate session had read σ_int as −6.2σ below the CDM prediction; that reading was retracted in-archive once distance-noise modeling was added — the source session's own definitive run reports CDM-consistent at z = +0.5, and the verdict is itself modeling-choice-dependent (z ranges +0.5 to +64 across choices). The retracted “below CDM” framing had propagated to this page, Galaxy Rotation, and CDM Discrimination — all three now carry the corrected verdict.

MOND's acceleration constant a₀ related to cosmological parameters via a₀ = cH₀/(2π). 13% error at H₀ = 67.4 vs observed value. Against this framework's own SPARC fit the comparison changes but gets no sharper: the scale the fit identifies is a₀′/γ (the compander's deep-regime limit; 2a₀′ at γ = ½). It stays at 1.07–1.09×10⁻¹⁰ across the whole Υ_disk = 0.4–0.6 band while a₀′ itself moves 3.6×, but only because γ floats along a valley with Υ (0.28 → 0.96 across that band). At pinned γ = ½, a₀ moves 1.44 → 1.10 → 0.87×10⁻¹⁰, so closing cH₀/2π needs H₀ ≈ 93 / 71 / 57 at Υ = 0.4 / 0.5 / 0.6. The free-γ value reads 1.16–1.18×10⁻¹⁰ under the velocity-χ² fits. cH₀/2π sits 3–12% below it at H₀ = 67.4. With γ free, closing the gap needs H₀ ≈ 69–70.5 for the Υ-sweep values (inside the Planck–SH0ES range) and ≈ 75–76 for the velocity-χ² values (above SH0ES's 73), so only the low end is absorbed by the H₀ tension. Still Milgrom's 1983 coincidence, still a reparametrization (maintainer/scripts/a0_identified_scale_across_upsilon.py). This numerical coincidence has been noted since Milgrom (1983). Other frameworks share its dimensional origin but not a prefactor: McCulloch 2007 has 0.99 cH, Verlinde 2016 cH₀/6, Smolin 2017 a_Λ/8.3 (table). The quantities c, H₀, and G are the only dimensionally relevant cosmological constants, and cH₀ naturally has units of acceleration. Best classified as dimensional analysis, not a unique derivation.

1,703 logged chemistry sessions (a counter, not a census). Sound velocity r = 0.982, electronegativity r = 0.94–0.96. The top correlations are strong because the γ proxies are textbook quantities renamed: γ_phonon = 2T/θ_D(so sound velocity against 1/γ is sound velocity against the Debye temperature, exactly) and γ_optical = 13.6 eV divided by the ionization energy (so electronegativity against 1/γ is electronegativity against ionization energy). A polynomial in atomic number does not reproduce them (run on the 10 recoverable rows, 2026-10-31).

Null model: the 2026-05-10 polynomial-in-Z null was run on synthetic monotonic curves (√Z, log Z, Z^1/3, Z = 1–50), not on the chemistry rows. Its |Δr| ≤ 0.07 shows only that two monotonic functions of one variable correlate highly. The earlier premise, that sound velocity, electronegativity and atomic volume rise or fall steadily with Z, is false across the periodic table (atomic volume against Z is the Lothar Meyer periodicity curve). The verdict, null-class / Reparametrization, stands on a different ground: each r is a post-hoc fit against a γ proxy that contains the target (γ_phonon = 2T/θ_D), and about 86% of the “89%” are restatements of the Debye model. The real-data Z-null has since been run on the 10 property rows recoverable from the session scripts and fails (|r with Z| ≤ 0.53). The same audit found that most of the site's 23-row table did not trace to its cited sessions, and that electrical conductivity, shown until 2026-11-02 as +0.955, is a documented wrong-sign failure (r = −0.414, session #81); see the Chemistry Correlation Explorer Trace column.

Additional caveats: ~11% failure rate. Era 2 chemistry (sessions 134-2660) identified as template-based. The ~1,700 chemistry sessions include statistically dependent samples (sound velocity, electronegativity, and atomic volume co-vary for well-known bonding reasons). See Chemistry Limitations.

Surface density Σ₀ = cH₀/(4π²G) ≈ 119 M☉/pc². ≈4% error vs Freeman's observed value (124 M☉/pc²) — corrected 2026-07-09 from a previously stated 110 / 12% arithmetic error (two simulated-visitor personas independently caught it the same day).

Like a₀ ~ cH₀, this is the only surface-density scale buildable from the available cosmological constants (c, H₀, G). Any framework that imports these constants will recover the same dimensional relation. It is also not independent evidence from a₀: Σ₀ = a₀/(2πG) exactly, so the ≈4% Freeman match is the same ≈13% a₀-vs-Milgrom gap propagated through a fixed linear relation, not a second derivation. Reclassified from “Validated” to Reparametrization alongside a₀ on the same grounds.

Caveats (updated 2026-07-17): the H₀ inconsistency with Parameter Derivations' former 123.3/0.5% figure is resolved — the split was an undisclosed Hubble-constant switch (119.0 at H₀ = 67.4, 123.6 at H₀ = 70; identified by a visitor physics persona). Site standard is now H₀ = 67.4 km/s/Mpc (Planck 2018), giving the ≈119 shown here. Separately, Freeman's (1970) value is a surface brightness; converting to surface density needs M/L_B ≈ 1–3, so the target is uncertain by a factor of ∼2 — neither page's percentage is meaningful standing alone.

An explicit, everywhere-analytic sigmoid has no critical point — no fixed point, no diverging susceptibility, no diverging correlation length. C(ρ) is not a critical phenomenon: ρ is an external input evaluated directly, not solved self-consistently (unlike Ising's m = tanh(βJzm), where m appears on both sides — that's what generates criticality). Comparing “predicted critical exponents” to renormalization-group exponents presupposes the very criticality C(ρ) does not have. The prior framing (“exponents 2× off”) was a fossil from the phase-transition era of the project — retained too long alongside the (correct) compander reframe. Corrected verdict: the Landau-universality argument fails not because the exponents are wrong by a factor, but because the category does not apply.C(ρ) belongs to the compander family (μ-law / Hill / Naka–Rushton class). No critical exponents exist to compare.

Average error 53%. Crystal structure dominates melting behavior, and C(ρ) has no crystal-specific parameters.

T_c = Δ/(1.76k_Bη) predicts 607K for YBCO (yttrium barium copper oxide). Actual: 93K. Off by 6.5×.

Independently derived, then found to be identical to Abrikosov-Gor'kov pair-breaking efficiency (1960). All 23 superconductor predictions are standard condensed matter in different notation.

C(ρ) was proposed to explain cross-scale hierarchy boundaries. 36/36 tests: 0/7 boundaries predicted. The tanh form is generic (Landau theory). C(ρ) is description, not explanation.

In plain terms: this card previously said the framework failed its own pre-registered environment test. It didn't — the number reported as the failing result belongs to a different test on a different, smaller sample. The environment test on the 14,585-galaxy sample was never actually run. That's a gap, not a failure, and it's a more embarrassing kind of error to have made on the site's own self-audit page than the failure it was reporting. Glossary →

R² = 0.14 with p = 5×10⁻⁶ and N = 14,585 is internally impossible: at that N, R² = 0.14 implies t ≈ 48.7 (p of order 10⁻⁵⁰⁰). The pair is self-consistent only at N ≈ 130–175 (SPARC scale) — TEST-05's sample. Archive tracing confirms it: 0.14 is a Hubble-type/morphology term from Session 377 on N ≈ 171, misattributed to the ALFALFA-SDSS cross-match. The registered TEST-03 environment-density correlation on 14,585 galaxies was never computed.

Independently found three ways: archive Session 639 (2026-04-30) first traced the metric conflation; the explorer track's 2026-07-08 citation-walk closed the provenance chain (registered threshold was r² < 0.09, which R² = 0.138 passes, and the catalog postdates the measurement by 15 days); two simulated-visitor personas independently re-derived the (N, p) inconsistency on 2026-07-09. See the corrected TEST-03 and TEST-05 cards for the full trace. TEST-05, which does own the (R²=0.14, p=5×10⁻⁶) result on its registered terms, was reclassified MOND-shared on 2026-07-09 — a verdict re-adjudicated 2026-07-15: the tie dissolves on lever magnitude (MOND+EFE's external-acceleration lever moves outer g_obs by ~0.09 dex; C(ρ)'s ambient-density lever by ≤2×10⁻³ dex — and the C(a) law predicts exactly zero). The registered environment-density run now exists (research repo, 2026-07-14: SPARC RAR offsets vs Cosmicflows-4 density, r² = 0.0001 vs the registered >20% claim — refuted by execution; weak secondaries are opposite-signed, EFE-like). See the TEST-05 card.

In plain terms: across 175 galaxies analyzed together, the framework's preferred curve-shape fits measurably worse than MOND's (ΔBIC is a fit-quality penalty; +184 holds Υ frozen and counts every point as an independent measurement, and they are not — scored by galaxy the margin is about +11, at the threshold, 2026-09-29). If you let the framework's shape parameter float freely, it simply turns into MOND's curve — so it is either worse than MOND or identical to it, never better. Glossary →

The only non-degenerate galaxy-scale discriminating test between the Synchronism compander (μ_Syn = tanh(γ ln(1+x)), γ=2) and MOND's RAR interpolating function was executed on 2807 real SPARC points. Kill criterion: ΔBIC > 10 favoring McGaugh refutes γ=2. Actual result: ΔBIC = +184 (Υ frozen; +46 with Υ free per law) (conservative intra-galaxy correlation correction: ΔBIC ≈ 33 — a guessed discount; the measured galaxy-level margin is ≈ +11, at the threshold, 2026-09-29).

Which coherence function this kills (target corrected 2026-09-08). This run used the compander as an implicit interpolating function keyed on g_obs(the script solves g_bar = g_obs·tanh(γ ln(1 + g_obs/a₀′)); verified in sparc_tanhlog_profile.py). So +184 refutes γ = 2 in the acceleration-keyed realization — MOND's own variable with a pinned return exponent q = 2γ = 4 — and not the density-keyed C(ρ) the headline equation states. Two simulated-visitor personas (2026-09-08) caught that this page, Dark Matter, the plotter and the landing scoreboard were all presenting +184 (Υ frozen) as the density-keyed kill while Coherence Function says every fit ran in acceleration. The density-keyed law has its own refutations, and they are stronger: head-to-head on the same SPARC points with γ free it loses at ΔBIC +2843 (read algebraically, g = g_bar/C; this excludes the framework's permittivity, not the Refracted Gravity field-equation class it belongs to: what “excluded” covers) (best-fit γ → 0.046 with its knee at 4.4×10⁻⁶ M☉/pc³, below every disc: a weak density response, boost ∝ ρ^−0.10, not a switched-off one; that response is g_bar's: at fixed g_bar, density adds nothing, partial slope +0.003, 95% CI −0.022 to +0.027, explorer 2026-11-04), its floored form is capped by the boost ceiling (B_max = 3.17 vs the 13.7 SPARC dwarfs require; lensing ν = 110–347, MOND's √(a₀/g) at 10⁻¹⁴–10⁻¹⁵ m/s² with a₀ = 1.2×10⁻¹⁰), and at the solar Oort limit it predicts f_DM = 0.685 against 0.13 ± 0.04 for any knee above ~0.15 M_☉/pc³. The strongest real-data constraint on it is not on SPARC at all: lunar laser ranging. A boost shared by a whole neighbourhood cancels in any same-instant ratio, but not in time. The Sun moves through the disc and past nearby stars, so G_eff = G/C(ρ) drifts. At the 1–10 pc density smoothing the galaxy fits need, the knee grid predicts Ġ/G ≈ 10⁻⁸ to 10⁻⁵ per year, against LLR's (7.1 ± 7.6)×10⁻¹⁴ (Hofmann & Müller 2018). That excludes 68 of 74 laws at 1 pc (62 at 10 pc), and the survivors are Newtonian to 10⁻⁸ at the Sun. The count holds under L2 and L3 alike (explorer 2026-09-23, pre-registered; L3 check 2026-10-19; details on the TEST-02 card). This is a second, SPARC-independent root for the same density-keyed kill, not a seventh refutation. Whether it becomes its own ledger row, or replaces the weaker environment registration (row #4) as the density-keyed entry, is the project lead's call. Its one registered per-object test is the globular-cluster fork below. Lead with those; +184 (Υ frozen) is the right number attached to the wrong model.

Audited-Negative — Form-free: local density adds nothing beyond g_barⓘ The simplest density negative needs no γ, no ρ_crit and no compander. On 2,622 SPARC points (149 galaxies), the boost each point requires scatters 0.161 dex when conditioned on local baryonic density and 0.118 dex when conditioned on g_bar: 1.37× wider, and 1.34–1.77× across nine gas and scale-height treatments, including a free scale height per galaxy. At fixed g_bar the partial correlation with density is +0.0012 (explorer 2026-08-02). The 2026-11-04 rerun on the exact 2,438 points behind the +945 and +2843 gives a partial slope of +0.003 (95% CI −0.022 to +0.027, galaxy bootstrap; injected slopes of −0.05 and −0.10 are recovered), and a model that never sees density reproduces the data's density slope to 0.2%. So the fitted density law's weak response is g_barseen through a proxy (their correlation is +0.84). This is the same root as the density-keyed rows above, not a new refutation. It sat in the explorer's findings for three months without reaching this site; a visitor never asked for it, because nothing on the site said it existed (added 2026-11-05).

The residual is a coherent S-shaped ≈0.05–0.10 dex signature at the RAR transition (g_bar ≈ a₀), significant at ~8σ per bin. Free-γ fit converges to γ≈0.49 = MOND, with RMS identical to McGaugh to four digits (Υ frozen at 0.5, unweighted). ΔBIC = +7 for free-γ is entirely the BIC parameter penalty, not a fit difference — the compander at its best-fit γ is MOND. Galaxy tests that selected Synchronism over MOND: 0, by execution (2 discriminated; both selected MOND).Script: explorer/scripts/rar_transition_shape_real_sparc.py. See also Galaxy Rotation: RAR Transition Shape.

“Converges to MOND” is not a coincidence — it is an identity, and here it is (added 2026-08-08). The free fit could not have landed anywhere else. Using this page's own Hill form, tanh(γ ln(1+x)) = [(1+x)^2γ−1]/[(1+x)^2γ+1], set γ = 1/2 exactly. Then 2γ = 1 and it collapses to C(x) = x/(x+2) = μ_simple(x/2) — MOND's simple interpolating function, with the factor of 2 absorbed entirely into the fitted ρ_crit. SPARC's 0.489 is 2.2% from that exact point. So the four-digit RMS agreement with McGaugh is the only possible outcome once the fit is free to find γ = 1/2: the functional form was never a difference from MOND at all. The only difference that ever existed is the argument (local ρ in place of g_bar) — which is exactly what For Researchers isolates and refutes. Corollary: the Hill index at the preferred fit is n = 2γ = 0.978 ≈ 1, and n = 1 is the non-cooperative limit of a Hill function. The framework's distinctive content is that density drives a collective transition; the data set the cooperativity parameter to the value meaning there is no collective transition — the same fact as the criticality retraction on The Core Idea, arriving through the fit rather than through the math.

In plain terms: galaxies never get dense enough to reach the “knee” where this framework's density switch is supposed to turn on — but globular clusters (tight balls of ~10⁵ stars orbiting our galaxy) do, from their cores outward. So they are the one place the density idea can be tested on its own terms. Result: if the switch has one sharpness everywhere, the clusters rule it out at every setting the framework has ever used. If the sharpness resets inside each cluster — which the framework had predicted, seven months earlier, and nobody had run — the clusters are marginally consistent, but only if the density switch acts in the simple algebraic way (L2). Under the framework's own action (L3), which adds a force the simple version leaves out, the registered setting is excluded too, narrowly (2026-10-08). The price of the L2 survival is that “one equation” becomes two. Glossary →

Data: Baumgardt & Hilker (2018) Galactic globular-cluster database — 167 N-body-fitted clusters, 2,025 binned velocity-dispersion measurements; 42 clusters pass the analysis cuts. Statistic: error-weighted outer slope d log σ / d log r over the outer factor 3 in radius, measured vs an isotropic Jeans solution on the catalogue mass model — a shape test, which a radially varying boost cannot hide from and mass normalization cannot enter. The registration text (Session 611): “Globular cluster internal dynamics should follow γ = 2 (member stars are resolved individually), despite the cluster acting as N_corr = 1 from the galaxy's perspective. This tests whether γ resets at each Markov blanket boundary.” It registers a γ, not a knee.

Three results that do not depend on γ. (1) Globular clusters discriminate density-keyed from acceleration-keyed gravity, and the discriminating variable is the external field effect: MOND with its EFE switched off scores like the density law. MOND survives these clusters because of the EFE, and a density-keyed law read algebraically (g = g_N/C, exact for an isolated spherical cluster) has no external field to appeal to. Scope added 2026-09-16 (simulated-visitor researcher persona): under the field equation ∇·[C∇Φ] = 4πGρ, which is linear in Φ, the Milky Way's field superposes and is refracted by the cluster's own C profile, a linear external-field dependence rather than MOND's nonlinear EFE. That is the L2 statement. Under L3 the internal residuals are 1–30 g_ext, plus a quadratic term. These rows do not contain it, so “EFE = 0 exactly” describes the algebraic reading only. (This answers, by measurement, the question a visitor researcher persona asked on 2026-09-07.) (2) The exclusion is a window on the knee, transferable to any theory of the form g = g_N/C(ρ): ρ_c ∈ 0.1–300 M_☉/pc³ is excluded at γ = 0.489, narrowing to 0.5–100 at γ = 2 — and the two knees this framework has calibrated (the A·V² value at 650, the measured 0.161) are inside or at the edge of it, while Refracted Gravity's published elliptical-galaxy knee (0.0083, Cesare et al. 2022; attribution corrected 2026-09-16, it had been cited as the disc value) and both of its disc knees (7.4×10⁻⁴, 4.3×10⁻³) lie below the excluded window, the first across its whole 1σ range — Refracted Gravity is not excluded, even with this framework's Ω_m floor substituted in (wording disambiguated 2026-09-14). Form is not what fails; the number is — the knee sits 13–500× above RG's elliptical calibration (more above its disc ones) in the theory this sector rediscovered. Correction 2026-09-09: this sentence previously also listed “the Oort-window 0.074–0.154” among the excluded knees. That was a γ mismatch — 0.074–0.154 is the solar-neighbourhood window at γ = 2, set against the exclusion band at γ = 0.489. Computed at the same γ, the solar window and the cluster verdict overlap at every γ from 0.3 to 3: the clusters call the bottom of the solar window ok and its top marginal. The joint local window is ρ_c ∈ 0.0039–0.0079 M_☉/pc³ at γ = 0.489 and 0.0735–0.078 at γ = 2, sliding as e^1/γ between them. The Sun and the clusters do not close this sector — there is a place for the knee at every γ. Scope condition (added 2026-09-15): that holds only if “density” means a smoothed density, and no archive document says over what length. Read pointwise, the Sun's own neighbourhood is the solar wind (≈0.14 M_☉/pc³ at Earth, falling as 1/r²), which crosses every knee in this window between Earth and Saturn; the field equation then makes the Sun's mass inferred from Earth's orbit and from Saturn's disagree by order unity (all 240 points of a pre-registered grid, including Refracted Gravity's own form, fail a deliberately loose 10⁻⁶ test). Refracted Gravity's authors anticipated this: they say a density-keyed law “unavoidably” needs a smoothing length “tens of astronomical units or larger” and postponed it (Matsakos & Diaferio 2016, §2.2.1). So this is a missing parameter (planets need > ~30 AU), not a new refutation. Replaced 2026-09-16 (explorer 2026-09-15): the stellar bound is not ~1 pc. Under the same field equation a compact body sits in a high-C bubble made by its own mass and feels 3C_out/(C_in + 2C_out) of the field a diffuse tracer feels (the dielectric-sphere factor). To fall like gas, stars need D ≳ 5–35 pc and globular clusters ≳ 100–900 pc. But this cluster window needs D ≲ 10 pc to say anything (smoothing over 30 pc turns every excluded knee marginal), and at that D halo clusters would orbit at ~0.63 of the field acceleration (0.34 at RG's elliptical parameters); Milky Way cluster and red-giant mass estimates agree to 1.00 (−0.17), a 2.1σ / 3.9σ tension, post-hoc. Under this field equation, the window can have its cluster leg or normal cluster orbits, not both. Adding the striction force a variational action implies cancels the bubble exactly, so the answer depends on which dynamics the framework commits to, and no archive document commits. The window is therefore tagged: γ-specific, D ≲ 10 pc, algebraic/field-equation reading. Open question: can the framework's MRH supply that length? Script: maintainer/scripts/density_keyed_law_vs_interplanetary_medium.py. (Withdrawn by the explorer 2026-09-08, having introduced it 2026-09-07. Second published instance of the program's characteristic error; it is why every density window on this page now carries its γ in the same cell.) (3)The one placement that escapes the shape test dies on the mass budget: reading ρ_crit = A·V² with the host galaxy's 220 km/s puts the knee at ~1 pc, saturating the whole cluster at the floor — a pure G → G/Ω_m rescaling, invisible to the slope but dividing every cluster's stellar M/L_V by 3.175, which pushes 96% of 167 clusters below M/L_V = 1.2, beneath any 12 Gyr metal-poor population.

What the verdict column rests on. (i) The cutoffs were set at execution, not at registration. Session 611 registers a γ and no quantitative criterion. “Passes” = outer-slope mismatch no larger than MOND+EFE's; “marginal” = under twice it; “excluded” = more. That benchmark was chosen on 2026-09-07 because these data admit MOND+EFE. So the table adjudicates a registered γ on a post-hoc scale, which is why the badge says fork, not survives. (ii) Mass and anisotropy were not refit per law. Every row uses the catalogue mass profile normalized to Baumgardt & Hilker's Newtonian, dark-matter-free N-body masses, with isotropic Jeans. The slope statistic cancels the mass normalization, but not the shape of a profile fitted under Newton, so the Newtonian row is the fit baseline and a per-law refit could move every modified row, most likely toward better agreement. Anisotropy was varied only as a robustness check: radial anisotropy cannot close the density law's gap at any β ≤ 0.8, and the tangential anisotropy measured in cluster outskirts makes every row worse. (iii) What survives this: the ordering (density-keyed at γ = 0.489 is about 2× worse than MOND+EFE under the same bias) and the shape argument. What does not: reading “marginal” as a pass. At γ = 2 the density law does no better than Newton or MOND+EFE, and its survival costs universality while buying nothing in fit.

The fork. If γ is universal — the “one equation” reading, near where the SPARC fit (0.489 ± 0.11) and the DESI fit (0.487 +0.024/−0.021; mean-density reading only, since on the local-density fluid reading P(k) pins γ to ½ within about 5×10⁻⁶, i.e. |2γ − 1| ≲ 10⁻⁵) land. Neither is a measurement of a universal constant:γ = ½ is exactly where the galaxy sector becomes MOND's simple μ and the dark-energy sector becomes Λ, so both datasets are sending γ to the point where the framework has no content of its own — the clusters exclude the density law at every placement the framework uses. If γ resets per Markov blanket (P611.2: N_corr = 1 inside a resolved-member system, so γ = 2), the clusters are marginally consistent at MOND-with-EFE's level under L2 (under the action, L3, excluded narrowly at the measured knee, 2026-10-08; see the fork card) — and the coherence function is not one function. This is not an ad-hoc rescue: it was registered seven months before the test, on independent reasoning, and it makes a further prediction — γ = 2 for every resolved-member system and γ ≈ ½ for unresolved ones. That ladder is the first time the N_corr machinery has done predictive rather than decorative work. It is not cheaply testable beyond the globular clusters: the explorer's pre-check found that every classical dwarf spheroidal except Leo I sits entirely below every knee the γ = 2 branch admits (so γ never enters; the dwarfs test only the boost floor, which they already fail, needing boosts of 4.5–300 against 3.17), and open clusters sit in a disc whose own density caps their boost at 1.0–1.27, inside binary-star noise. Stellar streams have not been checked. So the ladder is decidable only where it has been run (finding: explorer/findings/gamma-ladder-precheck-dsphs-cannot-see-the-knee-open-clusters-are-ambient-capped.md). It is also not a win if it passes: the boost ceiling still kills the floored law on SPARC independent of γ, so a passing ladder makes the framework two partial functions, not one function that works.

What SPARC says at the knees the local data admit (executed 2026-09-08, corrected 2026-09-09). The framework's own field equation div(C∇Φ) = 4πGρ, solved on 153 SPARC discs (Q ≤ 2, i > 30°) at every knee the solar and cluster constraints jointly allow, Υ_disk profiled, same likelihood for every model. Reference: MOND simple μ reaches χ²/N = 21.2 and 0.105 dex rms in g; Newton 465 and 0.410.

Read the last two rows. Lowering the floor does exactly what the boost-ceiling diagnosis says it should — the fraction of discs that cannot be lifted at all falls from ~80 % to 10–23 % — and the fit gets 3–17× worse. The amplitude problem and the shape problem are the same problem pointing opposite ways: keep the Ω_m ceiling and four discs in five cannot be lifted; raise it and the boost arrives in the inner disc where SPARC says nothing should happen. SPARC's objection is to where the transition sits, not to how big it is — and a switch keyed to local density puts the transition inside the baryonic disc by construction. That is the honest close of this sector: not “the floor is the wrong number” but ρ is the wrong argument. What would reopen it is a differently-argued C (surface density, an MRH-smoothed density, an acceleration), not another scan of ρ_c.

Two corrections carried here, both to the 2026-09-08 finding, both found by re-reading its script output. (a) That finding reported the SPARC fit as improving monotonically as the knee falls. True at γ = 0.489; false at γ = 2 with the Ω_m floor, which has an interior optimum at ρ_c ≈ 0.004–0.008 and rises on both sides. (b) It concluded “the floor is the only parameter SPARC is objecting to.” Its own freed-floor rows say otherwise (above). Both errors trace to the same cause: the finding's results table shipped with two unsubstituted template placeholders where the γ = 2 and freed-floor rows belonged, so the conclusions were written from a narrative rather than from the run. Also worth recording on the fork: at the Ω_m floor, γ = 2 is the better SPARC branch at every knee ≥ 0.0039 and holds the run's global optimum. It is still 3.2× MOND. That is a fork datum, not a rescue.

Why this is not counted as a seventh refutation (recommendation; gates on the operator): a registered prediction survives the test it was registered for. What is refuted is the conjunction{universal γ = 0.489} ∧ {any framework knee}, which the ledger never registered as one row. Near-miss recorded: the analysis was first run at γ = 0.489 — before the registration text was read — and would have published a confident refutation of a prediction nobody made. Third instance in six weeks of “which parameter does the criterion name?” Robustness: ordering invariant across King vs Plummer mass models, three radial windows, radial-velocity-only bins; jackknife moves no mean by more than 0.042; radial anisotropy rescues Newton at β ≈ 0.75 and MOND+EFE at 0.81 but the density law at no β ≤ 0.8, and real tidally-limited clusters go tangentially anisotropic, which makes it worse. Significance was quoted as the ratio to the Newtonian residual (3.7–4.4×); that ratio is not a statistic when the baseline is consistent with zero (under Plummer the Newtonian residual is +0.021 ± 0.027 and the ratio swings 5–9×, 2026-09-19), so read the density-law residual as −0.211 ± 0.027 against Newton's −0.057 ± 0.027, a 4σ-scale gap on the statistical error alone, not as sigmas — mass segregation, mass model and anisotropy are unmodelled and all push the same way. Not tested here: the S691 placement at 10⁻²³ kg/m³ (below any cluster density — out of scope, not exonerated); the sectors that never define ρ_crit at all. Finding: explorer/findings/globular-cluster-knee-test-executed-universal-gamma-excluded-registered-gamma2-survives.md + eight scripts; execution note back-annotated to Session 611 in the research archive.

In plain terms: the framework predicted the universe's large-scale structure should be about 10% “smoother” (less clumpy) than standard cosmology expects. The DESI galaxy survey measured the clumpiness (that's σ₈): the prediction misses by 2.4 standard deviations — unlikely to be luck — and it was written down only after earlier hints pointed that way, which makes it weaker still. Glossary →

The load-bearing constraint is amplitude, not direction. Session 107 (Dec 2025) predicted σ₈ ≈ 0.76 (calibrated to the then-live S₈ lensing tension). DESI DR1 full-shape (arXiv:2411.12021) combined σ₈ = 0.841 ± 0.034 (Table 10) — a 2.4σ tension. That tension is robust regardless of which single bin is inspected.

The single-bin “enhancement” reading is a qualified, not load-bearing, finding. LRG1 (z_eff=0.51) fσ₈/(fσ₈)_fid = 1.09 (+0.12/−0.14) — above the ΛCDM fiducial in that one bin. But the DESI DR1 full-shape RSD ensemble growth index γ_growth ≈ 0.58 ± 0.11, above GR's 0.545 — which leans mildly toward suppression, the framework's own predicted direction. Reading LRG1 alone as “sign-wrong regardless of which bin” over-reads a single ~0.6σ bin against an ensemble that trends the other way.

Correction (2026-07-14): the registered kill criterion and the delivered verdict use different statistics, and the substitution is not innocent. The kill criterion below is registered on fσ₈(z=0.51) > 0.46 for a >3σ ruling-out. Computed directly from this page's own numbers — LRG1 fσ₈ = (fσ₈)_fid × (1.09 +0.12/−0.14)  = 0.4733 × 1.09 = 0.516 (+0.057/−0.066) — the 0.46 threshold is exceeded by only ~0.9σ, well short of the >3σ the criterion demands for a ruling-out (the weaker >2σ “disfavors” clause at 0.45 is not met either: the prediction sits ~1.5σ from the data). Corrected 2026-10-10: until 2026-10-10 this read 0.474 × 1.16 ± 0.062 = 0.550 ± 0.062 and ~1.5σ; the 1.16 ± 0.13 was Table 9's QSO row (z = 1.49), not LRG1, on nine pages since 2026-05-26. A researcher persona asked for the table value to be quoted directly; it was, and the number moved the other way from the seventh refutation the persona thought it might become. The 2.4σ figure that formerly carried a “Kill Criterion Triggered” label on this page (last instance removed from the Verdict section 2026-07-17) is instead a comparison on σ₈, a different parameter inferred from a full-shape EFTofLSS fit whose perturbation kernels assume GR growth — using a GR-conditioned amplitude to falsify a modified-growth model risks circularity, a point this page's own EFTofLSS citation below already half-concedes (1–2σ theory systematic). DESI's own purpose-built modified-gravity analysis — Ishak et al., “Modified Gravity Constraints from the Full-Shape Modeling of Clustering Measurements from DESI 2024,” arXiv:2411.12026 (JCAP 09 (2025) 053), previously uncited on this site — constrains exactly the parameter a growth-suppression mechanism lives on: μ₀ = 0.11 (+0.45/−0.54) from DESI FS+BAO+BBN+n_s alone, tightening to μ₀ = 0.05 ± 0.22 with CMB+DES-SN. A ~12% fσ₈ suppression maps to a substantially negative μ₀, which sits inside DESI-alone's 1σ band (the exact mapping needs the assumed time-dependence worked out — not yet run on this site; seeded as an explorer topic).Honest reading: the test as registered lacks the statistical power to discriminate this framework from GR — a more useful negative result than a manufactured kill, but materially weaker than “Kill Criterion Triggered” implied. The framework's cosmology sector is not rescued by this correction: the growth suppression was never derived— it was calibrated. The σ₈ discrepancy on its own predicted value (0.76 vs. 0.841) remains a real, if GR-conditioned, 2.4σ miss. (Found and verified: 2026-07-14 simulated-visitor Pass 3 and Pass 4, cross-checked against this page's own cited numbers before editing.)

Two aspects of the honest verdict: (1) Post-hoc origin — doubly so — Session 107 was committed 2025-12-10 after DESI DR1 was published April 2024; σ₈ calibrated to the lensing S₈ tension (KiDS/DES era), then propagated to DESI. No prospective prediction registered. Additional layer: the S₈ tension itself is receding — DES Y3 6×2pt and KiDS-Legacy (2024–2025) reanalyses have pulled S₈ back toward Planck, so the calibration anchor was a transient observational state, not a stable target. This makes TEST-04a post-hoc against a moving baseline. (2) Disfavored 2.4σ on amplitude— the σ₈≈0.76 prediction sits 2.4σ below the DESI DR1 combined fit; this is the durable disfavor (a GR-conditioned statistic, not the registered kill — see the 2026-07-14 correction above). A 2026-05-25 “correction” that claimed kill not triggered was itself an error: 0.4497 ± 0.0548 belongs to arXiv:2512.03230 (DESI Peculiar Velocity Survey, z≈0.07) misattributed to the z=0.51 full-shape slot. The “mechanism-class transferable contribution” is NOT restored — it was an overstatement.Context (2026-05-23): EFTofLSS analyses (Cabass, Simonović, Zaldarriaga et al. 2024-2025) explain DESI DR1 fσ₈ within ΛCDM at 1-2σ via one-loop counterterms.Calibration note (2026-06-24): the kill fires on LRG1 alone at ~2.15σ if read as a single-bin sign test — standard practice treats that as sub-threshold. The amplitude tension (2.4σ, ensemble σ₈) is what previously carried the “Kill Criterion Triggered” label (label retired 2026-07-14; see correction above), not the single-bin direction.Currency (2026-07-02, sharpened 2026-09-10): DESI DR2 full-shape growth (fσ₈) constraints are not yet published (expected ~Spring 2027); this verdict is frozen at DR1. The re-open policy below is unaffected — no DR2 growth datum exists yet to trigger it. DR2 power (2026-09-21): as adopted, the DR2 kill (branch B) fires with under 1% probability if ΛCDM is true, so the near-certain DR2 outcome is “underpowered” again; re-registering before the data gates on the project lead (arithmetic on Tier 1 — TEST-04a). Do not confuse three DESI products: DR2 BAO (arXiv:2503.14738) published March 2025 and this program has already fitted it (the 2026-08-12 dark-energy likelihood: DR2 BAO + Planck priors + Dovekie SN) — but BAO constrains expansion, not growth, so it cannot adjudicate an fσ₈ criterion; DR2 full-shape is the registered statistic and is unpublished; and there is no DESI DR3 (DR3 appears here only as the ~2027–2028 venue proposed for TEST-26). A 2026-09-10 researcher-persona reader concluded from this site that the registration was two releases overdue. It is not — the trigger has not fired — but the conflation was the site's fault, and on a site whose method is pre-registration that is the costlier error.

Session #581 (2026-02-08) conducted an 8-test quantum audit. Overall verdict: zero confirmed predictions, 4 reparametrizations, 1 refutation, 1 post-hoc fit, 1 not-preferred.

Say it in velocities, not χ² (explorer 2026-09-09) — and note it is radius-graded. The ceiling refutation is easier to check and harder to wave away in the units the data are measured in: a ceiling of B_max = 3.17 leaves a +14.1% median velocity deficit at the last measured point of a SPARC disc, against a 5.2% median measurement error there. The cost is not uniform across the disc — 1.62× in χ² overall, 1.53× on the inner half, 3.08× on the last two points. The failure is an outer-disc failure, which is exactly where a bounded boost must fail. This strengthens the badge; it does not soften it.

Provenance worth stating, because it is the reason to trust the number: a 51-disc validity control initially said the ceiling costs only ~5% in χ² — which would have made this refutation invisible to the statistic the model is fitted with. That draft was retracted insidethe session when the full 153-disc decomposition reversed it to 1.62×. A control run on a subsample is a control on the subsample.Key refutation — boost ceiling B_max = 3.17: The framework predicts a maximum gravitational boost ratio B = g_obs/g_bar of B_max ≈ 3.17 (from SPARC — Spitzer Photometry & Accurate Rotation Curves — calibration). The deepest SPARC bin shows ⟨B⟩ = 10.82, with 579 individual SPARC data points exceeding B_max (106/153 discs) (corrected 2026-08-07: this page previously read “579 galaxies,” but SPARC contains 175 galaxies — the 579 counts rotation-curve points, as key claims states correctly). This is the strongest direct refutation in the framework's own internal audit. It was not previously visible on this page. Its corollaries were executed 2026-07-14/15 and both fail: a bounded boost has no deep-MOND regime, forcing the wrong BTFR slope (TEST-09 — kill fired at B_max = 1/Ω_m, 3.3σ; convention-dependent, and convention-free B_max ≲ 5.3 is excluded), and caps the apparent DM fraction at 1 − Ω_m = 68.5% under one convention — the “69% of SPARC exceeds it” headline doesn't survive the alternative baryon-budget convention (see Verdict, above); the class exclusion via SPARC's max observed fraction (TEST-10) does. Everything gravitational on this page that differs from MOND is downstream of this one bound. (Notation note 2026-06-12: the source audit called this quantity “γ”, colliding with the transition-sharpness parameter γ = 2/√N_corr used everywhere else on this page — including the separate “γ = 2 refuted at ΔBIC = +184” result (points as independent, Υ frozen; +46 with Υ free per law; ≈ +11 with the galaxy as the unit). They are different quantities; we use B here to keep them apart.)

The two “literature-consistent” quantum results (Γ = γ²(1−c) and Bell-freezing c(d)) are both reparametrizations: the decoherence formula is the textbook correlated-dephasing variance (Palma–Suominen–Ekert 1996); the Bell-freezing functional form was imported from waveguide QED (Session #235 admission). Audit finding propagated to site 2026-05-16.

In plain terms: Bell tests give quantum mechanics a distinctive score: real experiments reach about 2.83, while any classical, local mechanism can score at most 2. Simulations built from this framework's own machinery score 1.85–2.00 — classical territory — and only beat 2 when allowed to cheat with faster-than-light signaling. So the framework doesn't sidestep the famous quantum weirdness; it fails the same test classical physics fails. Glossary →

The framework's single-observer substrate was tested directly against Bell: CHSH simulations with freely chosen settings, measured only through observer-pattern phase-lock. Local construction: S = 1.98. Nonlocal-grid construction: S ≡ 2.00 at every coupling strength (gauge-equivalent to relabeling the measurement angles — a local model in disguise). Global-clock construction: S up to 2.67, but only by introducing signaling. None of the three reaches the Tsirelson bound (2√2 ≈ 2.83) without signaling. Scope: that is a statement about three constructions, not about the substrate class. Toner & Bacon (PRL 91, 187904, 2003) reproduce the singlet correlations exactly with local hidden variables plus one hidden bit of communication per trial and no observable signaling, and no such construction was built here. Stated accurately: purely local substrates are excluded by Bell's theorem (these runs illustrate it, they do not add to it); the two nonlocal constructions that were built failed; hidden-communication substrates are untested. It stays in the ledger as a construction check, not as an executed refutation of the class. The registered bet (research ledger B1, verbatim): “a substrate measured only via an observer-pattern's phase-lock to a target (free CHSH setting choice) reproduces quantum correlations without superluminal signaling”; “refuted if observer-relative statistics obey CHSH S ≤ 2 OR if violation only appears with signaling / superdeterminism.” Run sizes: 400,000 trials per setting pair (local, nonlocal-grid and density-substrate scripts), 200,000 (global clock); the statistical error on S is about ±0.005 (binomial error on each correlator; the scripts do not report it, this is our estimate). The nonlocal-grid sweep reads 2.000–2.004 over ten couplings, i.e. 2 within error: a deterministic local strategy at the standard angles reaches exactly 2, which is what an angle relabelling leaves you with. The global-clock 2.67 is not a Bell test: once a model signals, any S up to 4 is available, so that arm shows only that this construction signals (setting-dependence of the far marginal 0.43). (Added 2026-10-13, from a simulated-visitor researcher persona who asked for the bet and the error bars.)

Substrate-independence (run 2026-07-06): the same cap holds on the framework's own saturation-gated Intent-density substrate (S = 1.85 ≤ 2, no signaling) — the S ≤ 2 bound is Bell's structure theorem for any real-valued local-realist model, not an artifact of the borrowed phase substrate. The Born-rule cos² projection law reaches 2√2 exactly, but only by importing Hilbert-space structure wholesale. This is the framework's cleanest self-executed negative result; full construction detail on Two Reframes. The substrate's Bell behavior is not an untested protocol: it was tested, by execution, and capped at the classical bound.

In plain terms: the heavier a galaxy, the faster its outer edge spins — and the exact mathematical relationship between mass and speed (the “slope”) is measured precisely. MOND's prediction matches it. This framework's gravity boost has a built-in maximum, and a maxed-out boost behaves like ordinary Newtonian gravity — which forces the wrong slope. Run on 123 real galaxies, the framework misses by more than its own pre-registered failure threshold at the cap this site uses; at two other candidate caps it lands on or just inside the threshold. What holds at every cap: caps below about 5.3 are ruled out. Glossary →

Executed 2026-07-14 on real SPARC (same V_flat estimator applied to observation, MOND, and Synchronism): observed n = 3.75 ± 0.10 (reproduces Lelli 2019's 3.85 ± 0.09); MOND n = 3.81 ± 0.04 (passes, 0.6σ); Synchronism n = 3.35 ± 0.07 (fails, 3.3σ; registered kill criterion > 0.3 fires at 0.41, at B_max = 1/Ω_m). At that ceiling no choice of φ or a₀ rescues it — reaching 3.75 requires Ω_m → 0.001 and φ → 2, where the law degenerates algebraically into MOND.

Convention-dependent, like TEST-10 (ceiling sweep executed 2026-09-18). The ceiling B_max = 1/Ω_m is asserted, not derived. Re-running TEST-09's own pipeline with only the ceiling changed: B_max = 1/Ω_m = 3.175 gives n = 3.35, |Δn| = 0.41, fires; (Ω_m−Ω_b)/Ω_b = 5.389 gives n = 3.456, |Δn| = 0.297, does not fire (the registered threshold is a strict > 0.3); Ω_m/Ω_b = 6.389 gives n = 3.49, |Δn| = 0.26, does not fire. Under the pre-fixed rule (the kill stands iff it fires under every candidate), the kill does not survive its own sweep. Convention-free result: a bounded-boost law with B_max ≲ 5.3 is excluded by the SPARC BTFR slope — the slope analogue of TEST-10's “B_max ≲ 14 excluded by SPARC dwarfs”. The Failed badge stands; whether TEST-09 stays in the count of 6 is pending the project lead. The three ceilings are not equally consistent with the framework's own premises: 1/Ω_m is the floor the framework writes itself (the cosmic C₀ = Ω_m calibration, with cold dark matter inside that Ω_m), so the kill under it is the kill of the framework as written; Ω_m/Ω_b is what a galaxy sector that replaceddark matter would need, and the dark-energy sector cannot run without dark matter (2026-09-22). So the branch on which TEST-09 does not fire belongs to a variant whose cosmology fails, and the fork has a loser on each side. This ranks the conventions; it does not settle the counting rule.

Why the old “MOND-shared / cannot discriminate” badge was wrong in structure: the framework's boost is bounded (B ≤ 1/Ω_m = 3.17 — the ceiling in the boost-ceiling entry above), so it has no deep-MOND regime: its deep limit is Newtonian-times-constant, slope n → 2, not MOND's n → 4 (which requires the divergent √(a₀/g) boost). The previously listed “n → 4 deep-MOND” limb was asserted (S193, synthetic 9-galaxy ladder), never derivable from the bounded formula; archive S58 had recorded the discrepancy honestly and was overwritten. The BTFR is exactly an asymptotic-boost observable — this failure is a corollary of the boost ceiling, as is TEST-10's (dwarf DM fractions — the class exclusion via SPARC's max observed fraction, not the convention-dependent 69%/68.5% headline; see Verdict above; executed 2026-07-15). Full trace on Tier 1 — TEST-09.

In plain terms: the framework computes its gravity boost from how dense matter is right at each point. But real galaxy data organize by the total pull of everything enclosed within your orbit — a fundamentally different quantity. No single density threshold can bridge the two across systems (it misses clusters by factors of thousands), which is why the galaxy fits can't extend anywhere else. Glossary →

C(ρ) is a function of local density ρ. The RAR/MOND relation it mimics in galaxies is a function of g_bar — the enclosed-mass acceleration, a non-local, geometry-dependent quantity. A pointwise intensive variable cannot reproduce an acceleration-space relation across systems with different mass geometries except by per-system calibration.

The failure surfaces exactly where the two variables decouple: clusters require ρ_crit,cluster 10⁴–10⁶× smaller than the galaxy value — destroying universality. Four natural ansätze tested on Coma: A1/A4 overshoot by 10⁴; A2 collapses to Newtonian; A3 is structurally impossible (C ∈ [0,1) bounds velocity at ≤2, observed dispersion requires ~4.6).

The no-go on the velocity axis (2026-07-02): the same obstruction as a sign statement — for the density knee to track MOND's acceleration threshold, the BTFR forces ρ_crit ∝ a₀²/(GV²) ∝ V⁻² (profile-independent); the framework asserts ρ_crit = A·V⁺². As first stated (restated below), the exponent is inverted, the magnitude 240×–300,000× too high, and with the framework's own values the entire luminous disk sits at C ≲ 0.28 — the knee is never crossed inside a galaxy. No recalibration of A repairs a sign. Full derivation on Parameter Derivations. Restated 2026-08-28 (propagated here 2026-09-24): measured on SPARC (N = 129), the Jeans knee carries no velocity exponent at all, ρ_crit ∝ V^{−0.16 ± 0.19} (median 0.161 M⊙/pc³). The framework's V⁺² is excluded at ~11σ (one point per galaxy, N = 129, galaxy bootstrap). A V⁻² law for this Jeans-construction knee is excluded at ~10σ by the same exponent, so the construction cannot be the a₀ knee. That is not a test of MOND: MOND's transition density (the mean enclosed density where g_bar = a₀) is a different quantity, and it scales as V⁻² by construction on any sample that obeys the BTFR. The refutation stands as a wrong exponent, not a sign inversion. The 240×–300,000× magnitude is withdrawn: it used ρ_crit = 0.029·V², but A = 0.028 was derived (Session 53) for ρ_crit = A·V^0.5. Whether a disc crosses the knee is therefore reopened and estimator-dependent.

CFD mapping: C = 1/μ_eff. Dark matter (low C) should mean high viscosity = more sticky. But the Bullet Cluster shows dark matter passes through itself — LESS sticky than baryons. The viscosity interpretation predicts the wrong direction. The deeper structural failure (local ρ vs non-local g_bar variable mismatch) is described in the box above.

In plain terms: the framework needs a universal “absolute clock,” which physics-wise means a preferred frame of reference. Coupling that to known particle physics generically produces effects that experiments have already excluded at 15–27 orders of magnitude below the expected size. Other theories have known escape mechanisms; this framework implements none of them. Not a data refutation — a severe “why don't we see it?” problem. Glossary →

The framework's defining commitment — a discrete substrate with absolute time — generates a preferred rest frame. Two computed consequences: (1) no discrete 3D lattice has continuous rotational symmetry SO(3); “no preferred scan-axis” does not imply “full Lorentz invariance,” and grid geometry is unspecified. (2) The sharper result (2026-06-26 computation): the dim-4 SME Lorentz-violating coefficient c_μν is not Planck-suppressed here — tree-level c_μν = 0 by single-substrate universality, but the one-loop radiative correction is UV-dominated and Planck-cutoff-independent: c_μν ~ α/π ≈ 2×10⁻³. Existing cavity and nucleon-comagnetometer bounds reach 10⁻¹⁸–10⁻³⁰ — a 15–27 order-of-magnitude fine-tuning gap (a CPSU 2004 naturalness problem, not a data-driven refutation). Two standard perturbative escapes exist in the literature (SUSY — Groot Nibbelink & Pospelov 2005; anisotropic scale-hierarchy — Pospelov & Shang 2012, demonstrated for Hořava-class absolute-time gravity) but neither is exhibited in this framework; single-substrate universality is itself the obstacle to adopting the scale-hierarchy route. Status: open custodial-mechanism gap, not a closed refutation — this is the framework's most severe naturalness constraint and its only non-MOND, non-Zurek falsification channel.

A separate, structurally protected LIV channel (time-of-flight dispersion, dim≥5) is notaffected by this gap — even-k lattice symmetry forbids the linear term, and the surviving quadratic term sits ~10⁷ below current LHAASO reach and is non-unique to this framework (shared with LQG/causal-set models). Full three-lock argument and both LIV faces: For Researchers →.

The only genuine novel prediction path (R(I) viscosity correction to quantum pressure) gives corrections of ~10⁻⁸⁰ at the densest accessible physics. Lives at Planck-scale densities. Not accessible to any foreseeable experiment.

Previously labeled the sole surviving novel prediction. The condition Γ < m is the standard narrow-resonance / narrow-width condition from QFT — required for a Breit-Wigner pole to be well-defined. Synchronism's contribution is an ontological interpretation (“coherence cycle completion”), not the condition itself. QFT already classifies broad resonances as poor quasiparticles and narrow ones as well-defined particles. Novel-survivor count after audit: 0 of 9 (audit by LLM agents, not an outside domain expert; instrument uncalibrated, positive control and post-cutoff arm not run; 3,308 total sessions — see the A2ACW methodology for the full chain).

34 predictions, contingent on a C-axis calibration protocol that does not yet exist. The EEG (electroencephalography) experiment sketch ($150K, 12 months) cannot run as stated: per Key Claims, no procedure maps any measurement (EEG phase coherence, fMRI, Φ) onto the C-axis — the claim is currently unrunnable, which is the stronger verdict. Calling these “falsifiable” without the calibration step would contradict that verdict. Consistent with that: the one internal test ever cited against the 0.50 value (gnosis-research Session 63) measured SNARC salience — a different variable with no mapping to C — so no threshold value has actually been tested.

6 testable protocols for MRH-based (Markov Relevancy Horizon) measurement theory. Requires dedicated experiments.

Formerly listed here as “testable with existing survey data.” That was stale: TEST-04 (BAO peak-shift modulation) was withdrawn 2026-05-04 by internal contradiction — the framework's own Session 107 forecasts BAO matching ΛCDM at 0.0% in all five DESI redshift bins (the sound horizon is set at z~1100 when C ≈ 1 everywhere), so no modulation is predicted. The growth sector moved to RSD fσ₈ (TEST-04a, adjudicated above). See test catalog.

The Hubble tension — a ~5σ discrepancy between early-universe (CMB) and late-universe (distance-ladder) measurements of H₀ — is the dominant open problem in cosmology (2018–2026). Synchronism's a₀ = cH₀/(2π) ties a₀ to H₀ (see the correction below), but the framework offers no resolution of the tension. If C(ρ) couples to expansion-rate physics via ρ_crit ↔ Λ, there should be a prediction about how coherence modifies recombination (early-time fix) or late-time acceleration (late-time fix). Neither has been worked out. A framework claiming cosmological scope that does not address H₀ tension is leaving the most-cited empirical opening in cosmology off the table.

Correction (2026-07-27): “makes no statement on H₀” is too strong, and this page was the one over-claiming its own silence. The framework carries a₀ = cH₀/(2π) as Claim 3's key equation. That relation is invertible: with the measured a₀ = 1.2 × 10⁻¹⁰ m/s² it returns H₀ = 2πa₀/c ≈ 77.6 ± 15.5 km/s/Mpc once a₀'s ±0.24 systematic is carried (0.7σ from Planck's 67.4, 0.3σ from SH0ES' 73.0 — consistent with both). So there is an H₀ consequence; it is currently non-discriminating rather than failing, and it was being booked as no statement at all. What remains true is the substantive point above: no mechanism for the tension has been worked out, neither an early-time nor a late-time fix. The correct label is a non-discriminating soft constraint, not silence — and it is one more reading on which a₀ ≈ cH₀/2π looks like numerology rather than a derivation. Flagged by the graduate-physics simulated-visitor pass; the same inversion had been noted internally 2026-07-26 without reaching this page, which is the propagation break rather than the physics.