The Registry

Falsification ledger

Every test we armed against ourselves

Each claim in the registry ships the test that would kill it. 82 are on record. 10 have fired — taking 7 claims to the graveyard and putting 2 in quarantine. Every one of those kills was our own. 27 have never been run, which is the honest weak spot in this table.

10 fired 1 failed, fixed same day 3 closed by enumeration 1 searched, nothing found 40 survived 27 never run

key↑	claim	domain	status▪	re-run↓
USR-2026-0073	One collapse-noise window is still open, and testable The substrate-noise window is OPEN and experimentally live: GRW's collapse rate λ = 10⁻¹⁶ s⁻¹ at r_c = 10⁻⁷ m survives every published mass-proportional CSL bound — the tightest (Majorana Ge X-ray, coherent-nuclear channel) leaves ×49 headroom — while Adler's λ ≈ 10⁻⁸·⁵ is excluded by ×7 to ×6×10⁵ and GRW's original 1986 non-mass-proportional coupling is dead. 36–57% of the historically proposed GRW-to-Adler parameter box survives. MAQRO-class space interferometry reaches 1900× past GRW's point on a ~10–15 year horizon. GRW alive by ×49 at the tightest bound; Adler dead; MAQRO closes the window in ~10–15 years	Quantum foundations	PROVISIONAL	2026-07-31
USR-2026-0072	Causal sets get Λ's size right and its motion wrong Everpresent Λ — causal-set theory's fluctuating cosmological 'constant' — splits cleanly. Its magnitude prediction is real: Λ_rms = 1/√V₄ evaluates to 1.39× the observed Λ with this registry's own fitted H₀ (the only advance prediction of Λ's size in physics), and it tracks H² at every epoch, dissolving the why-now coincidence. But its dynamics are excluded by this registry's own pipeline: 0 of 2000 wandering-Λ realizations fit DESI DR2 + Pantheon+ + CMB better than ΛCDM (ensemble-typical Δχ² ≈ +3000; best single draw +260), 14% recollapse before reaching us, and 97% flip the sign of ρ_Λ inside z ≤ 2.5 — so w(z) is typically not even definable, the wrong shape for the registry's smooth 2.16σ descent. magnitude reproduced (1.39× observed); dynamics excluded — 0/2000 realizations beat ΛCDM	Spacetime discreteness	PROVISIONAL	2026-07-31
USR-2026-0071	Discreteness closes by weak prediction, not weak data The discreteness map is complete, and the frame-independent flank closes by weak PREDICTION, not weak data: FIRAS and cold cosmic hydrogen exclude unsuppressed Planckian momentum diffusion ('swerves') by 65–111 orders of magnitude, but causal-set theory derives no value for the diffusion constant κ — candidate scalings span >100 orders, the massive and photon sectors cross their bounds at different suppression powers (n* = 3.8 vs 2.0), and the next rung down is unreachable by 10³–10³⁰. No feasible swerve measurement distinguishes a Lorentz-invariant-discrete universe from a continuum. no discriminator in practice — bounds crush candidates by 15–111 orders, theory predicts no rung reproduced ≠ discovered	Spacetime discreteness	PROVISIONAL	2026-07-26
USR-2026-0070	Emergent gravity has no observational discriminator Discriminator audit of the emergent-gravity programme, applied to our own work: the thermodynamic-derivation wing has NO observational discriminator. Non-equilibrium Jacobson yields f(R), whose unscreened form is dead by 2.2×10⁴ in PPN γ (Cassini) and whose screened survivor reproduces ΛCDM for any f_R0 including zero. Logarithmic horizon-entropy corrections are 56–120 orders of magnitude too small at every accessible scale. Every zero-parameter holographic dark-energy member is dead. Ġ/G from G ∝ 1/N_horizon is dead by 293×. NO discriminator in the wing we built; the one live prediction is systematics-limited reproduced ≠ discovered	Quantum information	PROVISIONAL	2026-07-26
USR-2026-0065	Our own measurement drives gravity's thermodynamic derivation Our own measured modular temperature drives the thermodynamic derivation of gravity: reading the boost from the full modular dispersion gives β(x) slope = 3.141533 with R² = 0.9999999995, matching 2π/v_F to 2×10⁻⁵, and the Clausius relation δQ/T = δS closes to 6.6×10⁻⁵ using that temperature — while the naive infinite-space formula is 150% wrong. Λ enters the derivation as an integration constant permitted by the Bianchi identity, not as a sum of zero-point energies. β(x) matches 2π/v_F to 2e-5; Clausius closes to 6.6e-5 on the measured temperature reproduced ≠ discovered	Quantum information	PROVISIONAL	2026-07-26
USR-2026-0061	Entanglement transmits nothing Entanglement transmits nothing: across 18 interventions on one half of a critical chain (Haar unitaries, projective measurements, non-unital channels, a hard reset) the other half's state is unchanged to 9×10⁻¹⁶, and its full eigenvalue spectrum moves by 4×10⁻¹⁶ — so every observable is unaffected, not merely the tabulated ones. Holevo information about which intervention was chosen is exactly zero against 4.25 bits of choice entropy at the source. Holevo χ about the intervener's choice = 0 exactly (vs 4.25 bits at source) reproduced ≠ discovered	Quantum information	PROVISIONAL	2026-07-26
USR-2026-0060	A small patch cannot steer the whole system NEGATIVE RESULT bounding a popular claim: in a finite spin chain, operations confined to k sites reach exactly 4^k dimensions — the full local-algebra bound, versus only 2^k on an unentangled product state — but that is merely 0.1–1.6% of the full 2^N space. A small patch does NOT reach the whole space in finite dimensions; full cyclicity needs k = N/2, where float64 fails outright (3264 of 4096). reach = 4^k exactly — only 0.1–1.6% of the full space (not the whole) reproduced ≠ discovered	Quantum information	PROVISIONAL	2026-07-26
USR-2026-0059	A region of empty space has no state of its own A region of empty space has no state of its own: an 8-site block of vacuum has nonzero entropy and is mixed (1 − Tr ρ² = 0.988) while the whole ring is pure, and its mutual information with the complement equals exactly twice its entropy — 100% of an empty region's entropy is correlation with the rest. Proving correlations never vanish at maximum separation required 120-digit arithmetic (3.68×10⁻⁵⁷). I(A:Aᶜ) = 2·S(A) exactly — an empty region's entropy is 100% correlation reproduced ≠ discovered	Quantum information	PROVISIONAL	2026-07-26
USR-2026-0058	A two-sided black hole built from entanglement The thermofield double behaves exactly as ER=EPR describes a two-sided black hole: one side alone is perfectly thermal (trace distance 2.7×10⁻¹⁵ from the Gibbs state), the two-sided correlations equal the imaginary-time bridge correlator to 3.3×10⁻¹⁶, entanglement falls monotonically with β (hotter = more bridge), and the cross-side commutator is exactly zero. one side exactly thermal (2.7e-15); bridge identity to 3.3e-16; commutator exactly 0 reproduced ≠ discovered	Quantum information	PROVISIONAL	2026-07-26
USR-2026-0057	Entanglement counts a theory's degrees of freedom Entanglement counts degrees of freedom: fitting the Calabrese–Cardy law to block entropies alone recovers the central charge c = 0.500062 for the critical Ising chain (CFT value ½, converging to 0.500013 on inner windows) and c → 1 for the free boson as the regulator mass goes to zero. Two universality classes, both counted correctly from entanglement with no other input. c = 0.500062 (Ising, CFT ½) and c → 1 (boson) — read off entanglement alone reproduced ≠ discovered	Quantum information	PROVISIONAL	2026-07-26
USR-2026-0056	The vacuum's entanglement Hamiltonian is a boost The entanglement Hamiltonian of half the vacuum is a boost: at 200-digit precision the nearest-neighbour couplings ramp linearly with distance from the entangling cut (R² = 0.998–0.999), with the exact near-cut coupling t₁ = 3.141539 matching 2π/v_F to 1.7×10⁻⁵. The implied local temperature T(x) = 1/(2πx) is the Unruh effect on a lattice — and the T in Jacobson's Clausius relation. boost linearity R² = 0.998–0.999 (200-digit); near-cut t₁ = π to 1.7e-5 reproduced ≠ discovered	Quantum information	PROVISIONAL	2026-07-26
USR-2026-0055	Space rebuilt from entanglement alone Space can be reconstructed from entanglement alone: fed only the mutual-information matrix of a 64-site oscillator ring (blindness proven by a permutation test), classical MDS recovers the ring at r = 0.995 against true geodesics with 93% of the spectrum in a single Fourier pair. Removing entanglement tears space by DISCONNECTION rather than distortion — surviving pairs keep r ≈ 0.99 while the graph shatters. blind reconstruction r = 0.995; connectivity collapses 1.00 → 0 as entanglement is removed reproduced ≠ discovered	Quantum information	PROVISIONAL	2026-07-26
USR-2026-0054	Entropy rises in both directions of time In an exactly reversible HPP lattice gas (Loschmidt echo returns the initial state bit-for-bit), coarse-grained entropy rises in BOTH time directions away from a low-entropy initial state. The thermodynamic arrow of time lives in the boundary condition, not the dynamical laws (Boltzmann, demonstrated exactly). arrow in the state, not the laws — echo bit-exact, entropy rises both ways reproduced ≠ discovered	Quantum information	PROVISIONAL	2026-07-26
USR-2026-0053	A polymer's pull comes from counting states alone A polymer with zero energy in every configuration exerts a real tension matching T·dS/dx from pure state counting (Monte Carlo vs exact enumeration, max deviation 4.5×10⁻⁴). Verlinde's holographic-screen composition (Unruh temperature + N = Ac³/Għ bits + equipartition) returns Newton's F = GMm/r² to 2×10⁻¹⁶ — an algebraic identity given its assumptions, recorded as a consistency check, not evidence that nature's gravity is entropic. entropic force real (4.5e-4 match) · Newton composition = algebraic identity, not evidence reproduced ≠ discovered	Quantum information	PROVISIONAL	2026-07-26
USR-2026-0052	Entanglement obeys a first law of thermodynamics The first law of entanglement δS = δ⟨K_A⟩ holds in an interacting transverse-field Ising ground state: the ratio reaches 0.9945 at ε=10⁻³ and relative entropy is positive and quadratic (log-log slope 2.05). This is the identity from which linearized Einstein equations emerge in holographic derivations (Faulkner et al. 2013), verified exactly in a spin chain. δS/δ⟨K⟩ → 0.9945 at ε=1e-3 · rel. entropy positive, quadratic (slope 2.05) reproduced ≠ discovered	Quantum information	PROVISIONAL	2026-07-26
USR-2026-0051	Time emerges inside a frozen quantum universe A globally static quantum state contains ordinary time evolution for internal observers: the solver-found null state of a clock+qubit constraint is stationary to machine precision (⟨H²⟩ ≈ 6×10⁻³⁰, global qubit maximally mixed), yet conditioned on the clock the qubit Rabi-oscillates at full amplitude, matching Schrödinger evolution to 4×10⁻¹⁶ (Page–Wootters 1983, recomputed). Time as clock–system entanglement in a toy model — not a claim about nature's time. ⟨H²⟩ ≈ 6e-30 global · conditional evolution matches Schrödinger to 4e-16 reproduced ≠ discovered	Quantum information	PROVISIONAL	2026-07-26
USR-2026-0050	Empty space stores its entropy on the boundary The vacuum of a free lattice scalar field carries entanglement entropy that scales with a region's boundary, not its volume: fitted 3D exponent 2.09±0.01 (area predicts 2, volume 3), 2D 1.05 (perimeter), 1D flat to 9 digits. Bekenstein–Hawking-like area scaling emerges from a bare field theory with no gravity input (Srednicki 1993, independently recomputed). 3D exponent 2.09 (area = 2, volume = 3) — fit s.e. ±0.01, not an evidence σ reproduced ≠ discovered	Quantum information	PROVISIONAL	2026-07-26
USR-2026-0041	One gamma-ray photon pushes discreteness past Planck energy Assuming the 31 GeV photon of GRB 090510 (z = 0.903) was emitted within Δt ≤ 1 s of the low-energy photons, linear-in-energy vacuum dispersion requires E_QG > 1.4×10¹⁹ GeV — of order the Planck energy. The limit scales exactly as 1/Δt and carries no quoted uncertainty on that association window. E_QG > 1.4e19 GeV, of order Planck — scales as 1/Δt on an unquantified 1 s window	Spacetime discreteness	PROVISIONAL	2026-07-09
USR-2026-0040	No cubic lattice imprint in the quasar sky No cubic lattice anisotropy in the quasar sky: cubic ℓ = 4 signal at p = 0.2298 (battery NULL, p = 0.230); any cube-aligned distortion is below 1.20% rms (95% CL). null result — cube distortion < 1.20% rms (95%)	Spacetime discreteness	PROVISIONAL	2026-07-09
USR-2026-0039	Three quasar isotropy checks, stated separately Three separate results, stated separately. The quasar multipole battery (ℓ = 1–8, p = 0.226–0.495) draws its null from the data's own C_ℓ, so it is a self-consistency check that constrains nothing about isotropy. The eROSITA L_X–T residual dipole (amplitude 0.0996, p = 0.002) correlates with galactic-absorption proxies and is NOT claimed as cosmological pending an absorption-corrected refit. No shared axis was found among the measured axes at the sensitivity of a five-axis random-axis test (p = 0.20). three separate results — one is not an isotropy test at all	Spacetime discreteness	PROVISIONAL	2026-07-09
USR-2026-0038	The universe holds far fewer bits than Planck volumes Recomputation of a known bound, not a measurement: evaluating the Bekenstein–Hawking bound S = A/(4ℓ_p²) for a sphere of radius 4.4×10²⁶ m gives 3.4×10¹²³ bits, about 10⁶² times fewer than one bit per Planck volume. a recomputed bound (3.4e123 bits) — nothing was measured against it	Spacetime discreteness	PROVISIONAL	2026-07-09
USR-2026-0037	A cutoff divergence is framing, not evidence Framing, not evidence: a hard-cutoff one-loop integral diverges logarithmically, and renormalization absorbs that cutoff-dependence into the bare coupling so physical amplitudes are cutoff-independent. This is standard effective-field-theory bookkeeping with a complete non-simulation reading — the cutoff encodes ignorance of physics above it. no discriminator — recorded as framing, not evidence	Spacetime discreteness	PROVISIONAL	2026-07-09
USR-2026-0036	Photons show no sign of a fundamental tick No discrete-clock signature: photon arrival times show no energy-dependent dispersion attributable to a fundamental tick. null result — no dispersion	Spacetime discreteness	PROVISIONAL	2026-07-09
USR-2026-0035	Nature is not doing floating-point arithmetic No floating-point artifacts in nature: naive machine-precision arithmetic as the substrate of physics is ruled out. null result — naive floats ruled out	Spacetime discreteness	PROVISIONAL	2026-07-09
USR-2026-0034	No hidden pattern in delivered quantum random bits No lazy-PRNG structure found in the delivered quantum bits examined: 32,768 post-processed ANU vacuum-fluctuation bits (bias 0.0003, compression 1.003, next-bit prediction 0.501) and 56,832 SHA-512-whitened NIST beacon bits (bias 0.0006, compression 1.002, prediction 0.506), with weak-LCG and LFSR-32 controls correctly caught. null on delivered quantum bits — bounds the delivered stream, not the raw device	Quantum foundations	PROVISIONAL	2026-07-09
USR-2026-0033	Macroscopic definiteness needs no observer Macroscopic definiteness needs no observer: the standard GRW/CSL objective-collapse law (Λ = λN², λ = 1e-16 s⁻¹) makes a 1 µm dust grain definite in 2.8×10⁻⁸ s and a 6×10⁹ kg mass in 7.7×10⁻⁵⁸ s — far below any resolvable timescale, with no role for an observer. definite in ≲10⁻⁸ s (1 µm grain) under GRW/CSL objective collapse — no observer required reproduced ≠ discovered	Quantum foundations	PROVISIONAL	2026-07-09
USR-2026-0032	Which-path information erases the double-slit fringes A time-dependent Schrödinger double-slit simulation reproduces two effects: fringe visibility falls from 0.883 to 0.000 as the which-path marker overlap goes to zero, and complementary-basis sub-ensembles recover anti-phased fringes that sum back to the flat pattern, with closure verified to 1×10⁻⁶. visibility 0.883 → 0.000 with the marker; eraser closure to 1e-6 reproduced ≠ discovered	Quantum foundations	PROVISIONAL	2026-07-09
USR-2026-0031	How much a stronger-than-quantum box would leak Accessible information under single-box information causality is 0.798 bits at S = 2√2 and 2.000 bits at S = 4, and van Dam's one-bit protocol succeeds at every n with a PR box while decaying to 0.498 by n = 16 with quantum correlations. But the 1-bit budget is crossed at S = 3.12, NOT at 2√2 — the quantum ceiling does not coincide with this bound. 0.798 bits at 2√2; the 1-bit budget crosses at S = 3.12, not at 2√2 reproduced ≠ discovered	Quantum foundations	PROVISIONAL	2026-07-09
USR-2026-0030	A simulated Bell test lands on Tsirelson's bound A from-scratch CHSH Monte Carlo reproduces Tsirelson's bound: simulated \|S\| = 2.829 matches the quantum analytic value 2√2 = 2.828, exceeds the local-realist bound 2.000, and sits below the PR-box ceiling 4.000. This is a textbook result recomputed — a check that the code obeys quantum mechanics, not a measurement of nature. simulation reproduces 2√2 = 2.828 (textbook; not a significance) reproduced ≠ discovered	Quantum foundations	PROVISIONAL	2026-07-09

Open to the field This registry is not just ours — register any published result you want checked and it enters like everything else: provenance bundle, kill test, status set from a public audit, labeled reproduced ≠ discovered. How submission works →

The graveyard

Claims we killed, and what killed them

USR-2026-0044 2026-07-26

The growth index γ evolves from 0.41 (z < 0.3) to 0.10 (z > 1) — gravity weakens with redshift.

Died by its own kill test, on our own data, at our own hand. The 4.84σ evolution reproduces at 1.03σ with the sign reversed, no jackknife sample recovers it, and the claimed Δχ² is arithmetically unreachable given how well the no-evolution model already fits. The autopsy is the useful part: fσ₈ loses 93% of its sensitivity to γ by z ≈ 2, and γ is 91% correlated with σ₈ — so a high-z γ 'measurement' is mostly the amplitude degeneracy talking. Gravity does not weaken with redshift in this data; it looks like General Relativity (γ = 0.525 ± 0.077 vs 0.55). Post-mortem addendum: fixing a wrong fσ₈ point uncovered during the autopsy makes the evidence against this claim stronger, not weaker — γ_a moves from +0.63 to +1.19 while the claim requires about −0.68.

USR-2026-0007 2026-07-26

Sign-switching Λ (Λs-CDM) is NOT preferred by this data: ΔAIC = +1.50 versus ΛCDM with SH0ES included, with the vacuum flip at z = 3.11. The...

Published as weakly preferred at ΔAIC = −2.5; re-runs at +1.50, which is disfavoured. The whole reversal came from one transposed covariance block — the same Lyman-α pairing error that cost USR-2026-0001 a quarter of a sigma. Killed not because the idea is bad but because our own number was wrong, and the corrected number points the other way. The lesson is propagation: this claim sat unchanged for hours after the fix that invalidated it.

USR-2026-0047 2026-07-12

A Gaia wide-binary velocity analysis shows an apparent MOND-direction boost, but only under a clip-biased median estimator (residual 1.167). The...

Killed by its own estimator audit: the MOND boost lived only in the clip-biased median. Under the unbiased mean the wide binaries are Newtonian (0.948). The 28.7% clip rate in the a₀ regime was doing the work.

USR-2026-0023 2026-07-09

The Quaia quasar-dipole excess (raw D = 0.0208, 9.9σ over shot noise) is not cosmological: adversarial magnitude and galactic-latitude cuts...

The adversarial battery did its job: 9.9σ of raw excess collapsed to a kinematic D = 0.0066 under magnitude and latitude cuts. The dipole belonged to the catalog, not the cosmos.