Cookie Consent by Free Privacy Policy Generator Scientific peer review — Forbidden Directions research program, blog series, and article | Igor Moiseev

Scientific peer review — Forbidden Directions research program, blog series, and article

Scientific peer review — Forbidden Directions research program, blog series, and article

Review date: 2026-07-13
Material reviewed: current working-tree versions of the Preferred/Forbidden Directions research program; Parts 1–8 of The Geometry of Forbidden Directions; Appendices D1–D5; the companion article Growth Vectors and Caustics of Magnetic Flux Lifts; supporting reports, source code, and result artifacts under research/preferred-directions/.
Perspective: sub-Riemannian geometry and mathematical/observational astrophysics.
Recommendation: major revision for both the article and the public blog series.

1. Executive assessment

This is an unusually transparent independent research program with a real mathematical core, executable experiments, explicit negative results, and repeated attempts to falsify its own preferred interpretation. The strongest result in the current version is the exact elliptic-integral reduction for the launch-angle-averaged orbital period in the planar exponential magnetic profile,

\[\frac{\langle T\rangle}{2\pi}=\frac{2}{\pi}K(2\varepsilon), \qquad 0\leq\varepsilon<\frac12.\]

I independently reran the symbolic/numerical reduction; it agrees to machine precision. The flag calculation giving $Q=d+k+2$ at a finite-order zero is also correct under its stated local hypotheses.

The present package is nevertheless not ready to be presented as a settled research article or as authoritative scientific exposition. Four issues are decisive:

  1. The claimed caustic law remains conditional on the unproved identification of the first conjugate time with the orbital period for the exponential profile. The period theorem is exact; the caustic theorem is not yet proved.
  2. The article and blog are internally out of sync. Most importantly, Part 7 still says a generic magnetic fold has $Q=7$, whereas the article correctly shows that a generic rank-2 fold has $Q=6$. The obsolete “149 nulls, 79 spiral in real physics” conclusion also survives in summary documents despite the later finding that all 149 lie outside the T96 model magnetopause.
  3. The astrophysical evidence is mostly evidence about constructed or extrapolated fields, not direct evidence of physical solar, magnetospheric, or Jovian events. The real-data inputs are genuine, but the claimed nulls and folds are outputs of fragile potential-field/model continuations. Several census methods do not establish completeness.
  4. The novelty discussion omits the closest prior art. Montgomery’s 1995 magnetic lift already identifies the lifted sub-Riemannian structure, its extra bracket at a nondegenerate magnetic zero, and abnormal minimizers on the zero locus. A 2025/2026 paper by Barilari, Bossio, and Franceschi explicitly studies lifted magnetic sub-Riemannian structures, their step, zero sets, and abnormal trajectories. Until the article compares theorem-by-theorem with this literature, its contribution cannot be evaluated or advertised accurately.

My overall judgment is therefore:

2. Scoring rubric

Scores use a five-point scale:

Score Meaning
5 Publication-ready; only editorial changes
4 Strong; limited, non-load-bearing revisions
3 Substantive value, but major revision required
2 Important conceptual/evidentiary problems
1 Fundamentally invalid or unsupported

2.1 Companion article

Criterion Score Assessment
Scientific contribution level 3.0/5 The period-average elliptic reduction and profile-dependence calculation are potentially publishable. The magnetic lift and $Q=d+k+2$ flag count are mostly classical/elementary once placed beside Montgomery, Martinet, Bellaïche, Jean, and the recent magnetic-lift literature. Novelty of the caustic statistic remains unverified.
Logical correctness 3.4/5 The theorem/measurement/conjecture ledger is excellent and resolves several earlier overclaims. However, the article still draws unsupported conclusions from the profile expansion, calls two-profile evidence a general caustic law, and occasionally switches from period statements to caustic statements without preserving the condition.
Mathematical correctness 3.5/5 Lemma 2.3 and Theorem B are correct. Conjecture A is explicitly open. Several secondary statements need correction: the two-radius identifiability claim, “exact” square-root inversion, the $O(\varepsilon^4)$ remainder as currently proved, the solenoidal codimension count, and the universal-function notation for per-angle conjugate time.
Clarity of content 4.0/5 The status ledger, definitions, and separation of proof from measurement are very good. Clarity is reduced by an abstract that sounds more complete than the paper is, by result/history interleaving, and by terms such as “critical gradient” and “caustic law” being used beyond their proved scope.
Reproducibility 4.1/5 Core scripts run and reproduce the reported numerical relationships. Statistical uncertainty, null-census completeness, data provenance, and environment pinning need improvement.
Astrophysical grounding 2.6/5 The Lorentz-force reduction is physically standard, but the paper does not yet connect its observable to a feasible measurement, a plasma regime, dimensional units, electric fields, guiding-centre validity, or observational uncertainty.

2.2 Blog series and appendices

Criterion Score Assessment
Scientific contribution level 2.7/5 Strong research diary and methods narrative; weaker as a source of established new physics. The negative results and robustness audits are more defensible than several positive “world” claims.
Logical correctness 2.5/5 Later posts correct earlier claims, but the corrections are not propagated consistently. Contradictory fold, magnetosphere, detector, and exact-law statements coexist.
Mathematical correctness 2.7/5 Most local formulas are sound, but the series repeatedly presents a conditional caustic identity as exact, retains the obsolete generic-fold $Q=7$ claim, and describes singular-point $Q$ too casually as a ball-volume exponent.
Clarity of content 3.5/5 The figures, glossaries, and narrative arc are excellent. Precision suffers from superlatives, shifting meanings of “detected,” “real,” “certified,” and “pre-registered,” and an exceptionally dense Part 6 that contains several partially retracted stories.
Reproducibility 4.0/5 Scripts and artifacts are unusually visible for a blog. Some figure claims are not tied to uncertainty estimates or exhaustive algorithms, and some top-level documents are stale.
Astrophysical grounding 2.4/5 Solar and planetary inputs are real measurements/models, but the inferred topologies are model-dependent. The text sometimes equates magnetic nulls with reconnection and model interpolation with physical evolution.

3. Major findings

M1. The closest prior art must be addressed before any novelty claim

The article currently positions the flux lift through Montgomery’s book and Kostant–Souriau language, but omits the most directly relevant paper: Richard Montgomery’s Hearing the Zero Locus of a Magnetic Field (1995). That paper already:

This overlaps directly with the article’s magnetic lift, the $k=1$ instance of the flag law, and the abnormal-geodesic question left open in Remark 1.3/2.6. The $k=1$ statement is not merely analogous to known work; it is already part of that work’s mechanism.

The article must also compare with Barilari, Bossio, and Franceschi, Magnetic flows on 3D contact sub-Riemannian manifolds via the Rumin complex (arXiv:2504.09274, revised 2026, to appear in Transactions of the AMS). That paper explicitly characterizes the step and abnormal trajectories of a lifted structure when the magnetic field may vanish. Its base geometry differs from the Euclidean flux lift here, but it is close enough that omission is not acceptable in a 2026 article.

Required revision: add a “relation to prior work” section containing a theorem-level comparison. State precisely which of the following is new: the general finite-order Euclidean flag formula; the period-average elliptic identity for the exponential profile; the conjectured conjugate-time identity; the measured profile coefficient; and/or the astrophysical applications. Do not claim novelty for the magnetic lift, Heisenberg reduction, nondegenerate-zero extra bracket, or existence of abnormal curves at the zero locus.

M2. The period theorem and the caustic conjecture are still conflated

The current article makes a substantial improvement by labeling $t_c(\theta_0)=T(\theta_0)$ as Conjecture A. That distinction must now be propagated to every headline, summary, appendix, and post.

What is proved:

  1. the reduced exponential-profile orbit admits the displayed first integral;
  2. its heading period has the displayed elementary formula;
  3. the launch-angle average of that period equals $2K(2\varepsilon)/\pi$; and
  4. the period average diverges at $\varepsilon=1/2$.

What is not proved:

  1. that the first conjugate time equals that period for every launch angle and every $0\le\varepsilon<1/2$;
  2. that no earlier even-multiplicity Jacobian zero is missed by the sign-change detector;
  3. that the absence of a detected zero within a finite time window above $\varepsilon=1/2$ means no conjugate point exists; and
  4. that the physical “beam caustic” has the same operational meaning as the sub-Riemannian exponential-map caustic in a realizable plasma measurement.

The blog still calls the relation an “exact law” and says a charged-particle beam stops refocusing beyond one-half. Even under Conjecture A, only the slowest launch loses its period at $1/2$; many other launch angles continue to satisfy the period condition for $\varepsilon>1/2$. The mean over all angles becomes singular because a subset approaches the turning condition. “The beam stops refocusing” is therefore too broad.

Required revision: use one of these formulations consistently:

M3. Several statements in article §4 do not follow from the displayed mathematics

Four corrections are required.

  1. Two Larmor radii do not separate gradient and curvature at leading order. From the article’s measured law,

    \[\delta(r_L)=-\left[(L')^2-\frac34L''\right]r_L^2+O(r_L^4).\]

    Dividing by $r_L^2$ gives the same single combination at every sufficiently small radius. Two radii do not identify $L’$ and $L’’$ without a derived and independently parameterized higher-order term. The current claim of separation is unsupported.

  2. The square-root inversion is not exact even for the exponential profile. Since $-\delta=\varepsilon^2+\tfrac94\varepsilon^4+\cdots$, $\sqrt{|\delta|}/r_L$ is a leading-order estimator. Exact inversion would require the inverse of $K(2\varepsilon)$, not a square root.
  3. The $\beta=2$ example is a blind spot of the period average, not of the caustic. The article itself proposes the caustic coefficient $1-3\beta/4$, which equals $-1/2$ at $\beta=2$, not zero. The sentence saying the “caustic statistic cannot see” this field contradicts the subsequent measured law.
  4. The caustic coefficient is not yet a general law. It is supported by one control and two nonzero $\beta$ values ($-1$ and $-1/2$). Two points can select a line against one tested alternative, but they do not establish universality over general profiles. Moreover, the quoted “$32\sigma$” is not an inferential significance: the script’s sigma is a deterministic blend of fit-form spread and resolution bias, with no stochastic sampling model.

There is also a proof-writing issue in Proposition 4.1: the displayed expansion is carried only to an $O(\varepsilon^3)$ remainder, yet the result states $O(\varepsilon^4)$. The parity/reflection argument needed to eliminate the averaged cubic term should be written for a general profile, including the transformation of all profile jets, rather than imported from the special exponential family.

M4. The singular-point interpretation of $Q$ is improved but not fully repaired

Lemma 2.3 correctly computes the bracket flag and weights at a finite-order zero. The article also correctly retracts the automatic identification of that weight sum with a small-ball volume exponent at a non-equiregular point. That is a major improvement.

The blog, however, continues to define $Q$ globally as the exponent in $\operatorname{vol}B(r)\sim r^Q$ and then applies it directly at magnetic nulls. Work by Ghezzi and Jean shows that Hausdorff dimension and volume near singular strata depend on the singular locus and nonholonomic orders, not only the pointwise flag. The article’s “open problem O1” is the appropriate status; the blog must match it.

The criterion in Appendix D2 that “$Q>n$ on a set of full measure” distinguishes genuine sub-Riemannian geometry is not a standard equivalence and should not be presented as one. A Lagrangian caustic does not acquire a sub-Riemannian homogeneous dimension at its fold; an estimator may return a similar scaling exponent, but that is an estimator-confounding problem, not equality of the geometric invariants.

Required revision: reserve “pointwise weight sum” or “homogeneous dimension of the nilpotent approximation” for the null; reserve Hausdorff/ball-volume assertions for cases where the non-equiregular theorem has actually been applied.

M5. The article’s smaller mathematical issues should be corrected

M6. “Detection” is often a theorem-assisted consistency check

On the solar gallery, the reported $Q=6$ values are obtained by:

  1. finding a null with Newton’s method;
  2. measuring its Jacobian;
  3. building the exact linear tangent-cone field from that Jacobian; and
  4. measuring a weight that is already fixed by the nonzero linear jet.

This does not independently detect the null. It verifies that the code realizes Lemma 2.3 after the standard method has supplied the null and its Jacobian. The later raw-grid $w_4\approx3$ calculation is more independent, but it is scale-window selected and needs a frozen, externally validated window-selection rule before it can be called a detector.

The Jupiter raw-field calculation is stronger because it uses an analytic vector potential without injecting the Jacobian. Even there, the null locations come from the standard root finder first, so the result is local confirmation, not blind discovery.

Required revision: use a three-level vocabulary throughout:

M7. The solar validation is not yet adequate for astrophysical claims

The solar pipeline is useful as a software demonstration, but it lacks several controls expected for a solar-physics result:

  1. It treats HMI line-of-sight magnetograms as the normal boundary field without an explicit radial conversion/deprojection and without propagating HMI uncertainties. The appropriate HMI vector/SHARP products and remapped coordinate systems exist.
  2. It uses 45-second line-of-sight products in some runs, downsampled from 4096 to 1024, rather than a 720-second definitive radial/vector product designed for field modeling.
  3. It subtracts the mean flux and applies a periodic FFT continuation to cropped windows. The repository’s own window test shows low nulls disappear when the crop changes. Such nulls are properties of the chosen boundary-value model, not established coronal structures.
  4. The Newton seed search is not an exhaustive cell-by-cell trilinear/Poincaré-index census. run_r3_real_gallery.py also limits output to two nulls per region despite the prose saying “every interior null kept.” Completeness and recall therefore cannot be inferred.
  5. The eight-frame AR11158 sequence is too sparse and too model-sensitive to establish null count as a topological complexity index or a relationship to flare capability.

The standard trilinear method of Haynes and Parnell is specifically designed to be consistent with field-line interpolation and to find candidate cells before subgrid positioning. It should be the external baseline, supplemented by domain-size, resolution, noise, and boundary-condition ensembles.

M8. A magnetic null or fold is not itself a reconnection event

The blog repeatedly calls a null a “reconnection site” and a null collision “the transition state of reconnection.” In three dimensions, reconnection is defined by non-ideal evolution and a nonzero integrated parallel electric field; it can occur at a null, at separators or quasi-separators, or without any null. Conversely, a static null or a topological fold in a family of potential fields does not demonstrate reconnection.

The “certified solar fold” is a mathematically well-resolved saddle-node in a linear interpolation between two cropped magnetogram boundary arrays. It is not an observed temporal event, and machine precision in the spectral continuation does not reduce model or measurement error. In addition, a changing interior null count can be mediated by a null crossing the finite model boundary, so “every continuous path between the endpoints crosses a fold wall” requires a fixed domain and an explicit no-boundary-crossing proof.

Required revision: call this a “fold in a real-data-anchored boundary interpolation family.” Reserve “solar fold event” and “reconnection transition” for a temporally tracked, robust topology change accompanied by appropriate physical diagnostics.

M9. The magnetosphere result has been retracted locally but not globally

The T96 audit found that all 149 modeled nulls lie outside the model’s own magnetopause, while the valid interior was null-free over the tested parameter range. This invalidates the earlier claim that the census demonstrates 79 spiral nulls “in real physics.” The correct output is a warning about continuation beyond the empirical model boundary and a negative result about the tested smooth T96 interior.

Part 6 contains this correction, but README.md, the earlier S3 report, and Part 7’s summary retain the positive conclusion. A reader entering through the summary therefore receives the opposite of the final audit.

Required revision: propagate the retraction everywhere. The existence of spiral magnetic nulls in non-force-free plasma is not in doubt, but this T96 run does not provide valid interior examples. A Cluster/MMS catalogue comparison or event-conditioned field reconstruction would be the appropriate astrophysical test.

M10. The Jupiter result is an interesting but fragile model prediction

The Jovian analysis is potentially the most interesting astrophysical component because it states a precise, falsifiable location for a null and a fan dome. Its current wording is still too strong.

  1. JRM33 is a degree-30 model. Its coefficients are reasonably determined through degree and order 13, with useful information extending through degree 18. The calculation uses an $l\le18$ truncation, so “full JRM33 (degree 18)” is internally contradictory. Call it “the $l\le18$ JRM33 continuation.”
  2. Connerney et al. infer a Lowes radius near $0.81R_J$, not $0.85R_J$. If $0.85R_J$ is a chosen evaluation surface, it should not be called the measured dynamo surface or the established metallic-hydrogen boundary.
  3. “Exactly two nulls” is not established by approximately 3,000 Newton seeds. An exhaustive spherical-cell degree test, convergence with seed density, or a global topological-degree accounting is needed.
  4. No covariance/ensemble propagation of the Gauss coefficients is performed. This is especially important because the claimed polar null vanishes at $l\le13$ and therefore depends on the less-certain $l=14$–18 tail amplified by downward continuation.
  5. Seventy-two fan lines show a dome-like footprint but do not establish that the full patch boundary is the fan footprint. The reported boundary-to-footprint 90th percentile of $15^\circ$ is large. The absence of bald-patch segments does not by itself prove that all remaining boundary segments belong to the null separatrix.

The defensible conclusion is:

In the $l\le18$ JRM33 potential continuation evaluated down to $0.85R_J$, the numerical pipeline finds a polar radial null near $(r,\mathrm{lat},\mathrm{lon})=(0.873R_J, 69^\circ,282^\circ)$ whose sampled fan surface forms a dome associated with much of the reversed-flux patch boundary. The structure is absent at $l\le13$ and must be treated as a high-degree, model-dependent prediction pending coefficient-uncertainty propagation and a future field model.

M11. Statistical language is stronger than the uncertainty model

The reported uncertainty bars are predominantly numerical convergence and fit-model spread. They are not repeated-sample standard errors, posterior intervals, or calibrated confidence intervals. Consequently:

The exact elliptic derivation already settles $c_4=9/4$ for the period average; a dramatic sigma claim is unnecessary.

M12. The research record and the publication layer must be separated

The repository usefully preserves failed hypotheses and superseded interpretations. The public article and blog, however, should present the current scientific state, with the history moved into dated correction notes. At present:

Create one machine-readable/current status ledger and generate or manually synchronize all summary tables from it. Preserve the historical reports, but label them superseded.

4. Minor and editorial findings

  1. Use “sub-Riemannian” consistently; avoid “SR” in the article abstract and theorem statements unless defined.
  2. State dimensions/units in the Lorentz-force reduction. Explain how the fibre momentum $w$ maps to charge-to-mass ratio after nondimensionalization.
  3. Restrict “uniform rotation is Heisenberg” to the planar reduction. A 3D base produces a different rank/corank structure.
  4. Replace “contact where $\mathbf B\ne0$” by “contact in the planar lift; quasi-/even- contact in the 3D-base lift,” with conventions cited.
  5. A magnetic null is a topological feature associated with possible current focusing, not automatically a reconnection site.
  6. Give NOAA active-region identifiers, exact UTC timestamps, HMI series names, map projection, pixel scale, disk position, and uncertainty product for every solar figure.
  7. State the seed density, convergence tolerance, deduplication radius, and completeness test beside every null count.
  8. “Pre-registered” should remain qualified as repository-internal and timestamped. Git history can document ordering but is not an external registry.
  9. Distinguish “real data,” “data-fitted model,” “field extrapolation,” “analytic test field,” and “direct observation” in figure captions.
  10. Reduce Part 6 into at least two posts: fold methodology/solar interpolation, and magnetospheric model audit. The current density makes logical updates difficult to track.
  11. The article’s title promises caustics, but its only exact new closed form is presently a period formula conditional on Conjecture A for caustics. Either prove the conjecture or reflect the conditional status in the title/abstract.
  12. The repository should add regression tests that ensure stale, retracted headline strings such as “149 nulls, 79 spiral — real physics” and “generic fold $Q=7$” do not reappear in current summaries.

5. What is strong and should be retained

  1. The exact factorization in Theorem B. The tangent-half-angle/Gauss-integral chain is compact, correct, and well suited to a short mathematical-physics note.
  2. The status ledger. Explicit labels for theorem, measurement, and conjecture are a major improvement and should become the series-wide standard.
  3. The negative races. Reporting that the least-squares field fit beats both the flow classifier and the crossover estimator is scientifically valuable.
  4. The boundary and robustness audits. The T96 magnetopause audit, solar window tests, JRM truncation test, bald-patch alternative, and dipole control are good habits even where they weaken the original claim.
  5. Gauge awareness. The treatment of gauge equivalence and adapted coordinates is careful and relevant.
  6. Executable artifacts. The core numerical results are reproducible locally and the scripts are readable enough to audit.
  7. The corrected generic-fold result. Showing that the scalar order-of-zero statistic is blind to a rank-2 fold is an important clarification of what the method cannot see.

6. Contribution hierarchy after review

Tier A — plausible publishable units after major revision

  1. Exponential-profile period average. Exact elliptic-integral expression, coefficient series, and period divergence. To become a caustic theorem, Conjecture A must be proved.
  2. Actual caustic profile expansion. The numerical evidence for $c_2^{\mathrm{caustic}}=1-3\beta/4$ is interesting, but publication requires an analytic derivation and tests on more independent profile jets.
  3. Model-robustness methodology. A methods note centered on how window choice, interpolation, model boundaries, harmonic truncation, and null-finder completeness can manufacture or destroy modeled magnetic nulls. The negative findings are the result.
  4. Jovian conditional prediction. Potentially a short planetary-magnetism note after coefficient-uncertainty ensembles and exhaustive topology tests.

Tier B — correct but incremental/classical

  1. The Euclidean magnetic flux lift and Lorentz-force reduction.
  2. The Heisenberg uniform-field model.
  3. The $k=1$ Martinet/nondegenerate-zero growth step.
  4. The general finite-order flag count $Q=d+k+2$ as a concise extension/dictionary.
  5. The fact that a scalar order-of-zero statistic cannot distinguish radial from spiral linear nulls or identify a generic rank-2 fold.

Tier C — not established by the present evidence

  1. A practically superior SR magnetic-null detector or classifier.
  2. A physical solar fold/reconnection event.
  3. Interior spiral nulls demonstrated by the T96 census.
  4. A general two-scale inversion separating field gradient and profile curvature.
  5. A universal caustic critical gradient at $\varepsilon=1/2$.
  6. A physically established Jovian polar separatrix skeleton independent of uncertain high-degree coefficients.

7. Required revision plan, in priority order

Before public release of either article or blog

  1. Synchronize every current summary with the generic-fold $Q=6$ correction and T96 boundary retraction.
  2. Replace “exact caustic law” by conditional language everywhere, or prove Conjecture A.
  3. Correct the §4 inversion, two-radius identifiability, $\beta=2$, and “sigma” claims.
  4. Add Montgomery (1995) and Barilari–Bossio–Franceschi (2025/2026), with explicit novelty comparison and abnormal-geodesic discussion.
  5. Separate model-derived topology from observed astrophysical events in titles, abstracts, figure captions, and status tables.

Before submitting the mathematical article

  1. Prove first-conjugate-time equality, or refocus the article on the exact period result.
  2. Complete the general-profile parity/remainder proof and derive the actual caustic coefficient analytically.
  3. Correct the smaller notation, codimension, global-bundle, and dimensional-analysis issues.
  4. Provide a formal proposition describing precisely what the reach estimator converges to and under what scale/noise assumptions.

Before making astrophysical discovery claims

  1. Rebuild the solar analysis with definitive remapped radial/vector HMI products, exhaustive cell-based null finding, domain/resolution ensembles, and propagated uncertainties.
  2. Treat temporal topology with tracked regions at adequate cadence and explicitly rule out boundary crossings.
  3. Reanalyze Jupiter with JRM coefficient ensembles, several truncations/models, and an exhaustive spherical topology census.
  4. Compare any magnetospheric null claim against event-conditioned Cluster/MMS data or a validated reconstruction inside the model domain.

8. Verification performed for this review

The following repository checks were run on the current working tree:

Check Result
scripts/smoke_test.py Passed: uniform-field conjugate time, gauge invariance, dilation collapse, leading $-\varepsilon^2$, 2D/3D weight laws, analytic-field null check
scripts/run_t2_reduction_check.py Passed: direct integral, Gauss chain, and $2K(2\varepsilon)/\pi$ agree to machine precision; symbolic factorization exact
scripts/run_v1_linear_profile.py Reproduced: $c_2\approx1.7466$ for $\beta=-1$ and $1.3741$ for $\beta=-1/2$; also reproduced $t_c\ne T$ away from the exponential profile
Artifact audit Confirmed tangent-cone caveat for solar $Q$, T96 population outside model boundary as reported in later text, crossover race loss, and JRM truncation sensitivity

These checks validate the implementation against its own analytic targets. They do not replace an independent proof of Conjecture A, an external statistical error model, or an astrophysical validation dataset.

9. Primary literature that must anchor the revision

10. Final recommendation

Article: major revision. A publishable short mathematical-physics paper is plausible if it is narrowed to proved content, positioned against the closest prior art, and either proves Conjecture A or clearly makes the caustic formula conditional. The current article should not be submitted with “complete statements, hypotheses, and proofs” in the subtitle while its central caustic identification remains open.

Blog series: major revision before publication. Retain the experimental narrative, figures, negative races, and honesty boxes, but synchronize all corrections and adopt stable evidence labels. The blog can be excellent open research communication; it should not ask polished prose to carry more certainty than the mathematics or astrophysical data provide.

Program: continue, but concentrate effort on the two decisive frontiers rather than adding more “worlds”: (1) prove or refute the exponential-profile conjugate-time identity; and (2) perform one astrophysical test with externally standard data products, uncertainty propagation, exhaustive null finding, and a frozen detection protocol.