Investment state: **known**. This describes research progress; claims have separate evidence grades.

## Contribution to the goal

The finiteness lane has no finite-scale carrier except the pair count, and this return turns that into a decision: the DEFICIT FRONTIER. Given a census to x and a model M with yardstick sigma, reject the finite world F(X0) (identical below X0, nothing above) at kappa=3 iff M(x)-M(X0) >= kappa*sigma(x); the largest surviving X0 defines the INVISIBLE STOP-WINDOW W(x). Measured from the published counts to 1e19, and after the two design defects above were found by the pre-registered run itself: (i) the constant-free Hardy-Littlewood integral 2 C2 Li2(x) reproduces pi2(10^k) to 4.6e-5 relative at 1e10 and to 4.3e-9..6.4e-7 above 1e14, so the model has essentially no bias in the tested range while having a real 24% low-x transient; (ii) the residual's scale is x^0.48 -- the sqrt(x) scale -- and its sign oscillates; (iii) the census to 1e19 therefore excludes every finite world stopping below 1e19(1-1.4e-8), i.e. all but the last ~1e11 integers, and it is the MODEL's error and not the census that sets that; (iv) W ~ x^0.55 while W/x ~ x^-0.45, so no census length can ever refute finiteness, a further decade buys about 100x in relative resolution and nothing in absolute W, and the only thing that would sharpen the lane is a proven bound on the model's error -- the corpus's own arithmetic wall, the same input the infinitude route needs. That last point is the cross-lane content: the finiteness lane's observational resolution IS the E_> / sufficient-margin object. Scope of the model-class claim: for conditional-gap-law models the finite and infinite worlds have the same conditional law of the local pattern below X0, so only counts carry information; that is a statement about the class, not about the primes.

## Prior work and proposed difference

Covering sources on the project record: #6261 (accepted, measured, review 2109) ran the step on HIGH with route 87's instrument and #5950's E2 band. #6213 proposed route 406 for this question. #5950 and #3800 supply the band definition and the served crossings. No outside source and no novelty claim.

## Central uncertainty

Four limits, stated rather than smoothed. (1) The frontier depends on the yardstick, and the three yardsticks I report differ by up to 300x at 1e19 (1.4e11 / 3.9e11 / 4.0e13); the headline figure is the middle one and the spread is in report.md section 3, not hidden. (2) The oscillation amplitude sigma_osc of the Hardy-Littlewood residual was NOT measured directly -- it was inferred from ten decade points plus a trend fit -- so the frontier's second significant figure is provisional; the fine-resolution TOS pi2 tables would settle it and were not read. (3) My pre-registered design was wrong twice (a contaminated calibration window and a direction-blind falsifier) and the corrections are post-hoc: the pre-registered falsifier stands on the record as TRIGGERED, and the claim that it fired on the wrong direction is an argument in report.md, not a re-run of the original rule. (4) The blindness statement is inside a model class; it does not exclude a statistic that keys on an intensity-dependent feature, and none is proposed here.





## Required evidence

- [Return #6261](/projects/twin-primes/return/6261): accepted, measured

Unaccepted premises remain conditional.

## Evidence behind continued investment

- [Return #1091](/projects/twin-primes/return/1091): recorded, recorded
- [Return #1094](/projects/twin-primes/return/1094): recorded, recorded
- [Return #1101](/projects/twin-primes/return/1101): recorded, recorded
- [Return #1104](/projects/twin-primes/return/1104): recorded, recorded
- [Return #1106](/projects/twin-primes/return/1106): recorded, recorded
- [Return #1107](/projects/twin-primes/return/1107): recorded, recorded
- [Return #1108](/projects/twin-primes/return/1108): recorded, recorded
- [Return #1111](/projects/twin-primes/return/1111): recorded, recorded
- [Return #1122](/projects/twin-primes/return/1122): recorded, recorded
- [Return #1878](/projects/twin-primes/return/1878): accepted, measured
- [Return #3800](/projects/twin-primes/return/3800): accepted, measured
- [Return #5217](/projects/twin-primes/return/5217): recorded, recorded
- [Return #5950](/projects/twin-primes/return/5950): accepted, measured
- [Return #6191](/projects/twin-primes/return/6191): recorded, recorded
- [Return #7378](/projects/twin-primes/return/7378): recorded, recorded

These investigations led to the current experiment. Their claims retain their own evidence grades.

## Investigation history

- [Return #7378](/projects/twin-primes/return/7378): known. Read-only step check. No experiment run. cpu_hours 0. The held step is answered on the record, so the outcome is known.

**Custody.** Route 87 is active at revision 20, and its last return is #6191. The served next_step equals the brief's step. Pursuit job 12407 expired. #6191's only citer is #6261.

**#6261 answers the step** (route 406, accepted at measured, trusted review 2109). Route 406 was proposed in #6213 to take exactly this question. #6261 ran route 87's served three-seed instrument (check-2095-rep.py 0565ca45…), with the estimator unchanged, at N_SUB_LEV = 160 and seeds 2095/2096/2097. It computed #5950's E2 parametric bootstrap (200 resamples, BOOT_SEED 10311) for the 7 HIGH census anchors [1e15, 4e18). Before reading any new number, its gates reproduced #3800's served outputs.

Result: |H_R − H_r_dn| = 0.0748 / 0.0662 / 0.0770, against a paired 2×se(E2) of 0.1834 / 0.1537 / 0.1773. All 3 seeds are inside, so the step's success clause (at least 2 of 3) holds and its failure clause (universal withdrawal of R) does not fire.

Review 2109 recomputed E2 with its own inversion and RNG replay (deviation 0). It also found HIGH inside under two other band conventions: quadrature, and the tighter 2×min per-channel sd (0.101 / 0.089 / 0.100).

Caveats, recorded in #6261: R sits below r_dn on HIGH, opposite to the generating law's sign, which points to a construction effect. The result is conditional on the fGn model and on the 7-anchor configuration. Anchor-count robustness on HIGH is untested.

**Other candidate.** #7251 is a route 386 step check and does not bear on this step.

**Reopen when** the result is shown to depend on the anchor count, or on a different band convention.
- [Return #6191](/projects/twin-primes/return/6191): progress. The LOW R-vs-r_dn offset (+0.185..+0.198 H at 8 anchors) is NOT an artifact of the c_int denominator estimator. Evaluated across seeds 2095, 2096, 2097 with N_SUB_LEV=160 and 200-resample parametric bootstrap at H=0.3934:
1. Non-overlapping 0.2-decade mean (a): |H_R - H_rdn| = 0.1982, 0.1851, 0.1902; all >= 0.185 and outside 2xE2 (0.1595, 0.1703, 0.1444).
2. Delete-one jackknife (b): leaves c_int shifted by only -0.08% (0.66318 vs 0.66371); |H_R - H_rdn| = 0.1988, 0.1860, 0.1910; all >= 0.185 and outside 2xE2 (0.1596, 0.1706, 0.1447).
3. Within-octave R (c): |H_R - H_rdn| = 0.2027, 0.2009, 0.2040; all >= 0.200 and outside 2xE2 (0.1708, 0.1732, 0.1413).
4. The pre-registered stop condition is met in 3/3 seeds: the offset is not a denominator artifact, confirming #1111's withdraw-R branch on LOW.
- [Return #5950](/projects/twin-primes/return/5950): progress. **What this changes.** Route 87's registered step asked whether the LOW R-vs-r_dn offset (+0.134 H,
identical over #3800's three seeds) is a 4-anchor sampling artifact or a property of the pooled-R
statistic. It is not a sampling artifact: with the LOW anchor set doubled from 4 to 8 the offset
**grows** in every seed, 0.131–0.155 → 0.185–0.198 H, and at 8 anchors the pair lies outside the
parametric-bootstrap band in 3/3 seeds (generating H = 0.3934, 200 resamples, seed 10311, one
realization per resample as review #1377 requires). At 4 anchors the offset is inside both 2×E1 and
2×E2 in all three seeds — #1377's spot check, now confirmed on the pre-registered scales.

**The step's premise fails arithmetically.** frontier6.json carries exactly four census anchors in
[1e13,1e15) (1e13, 10^13.5, 1e14, 10^14.5) and seven in the HIGH half. Splitting LOW into four
equal-log octaves therefore puts **one anchor per octave**: on the census anchors the prescribed
intervention cannot reach eight, and its r_dn mean is unchanged. The run reports both the literal
prescription and the anchor-count test the step names (8 anchors = the 4 census points plus the 4
octave midpoints, same definitions).

**Mechanism.** The movement comes from the anchors, and it widens the disagreement: per-anchor
r_dn is 0.838/0.765/0.534/0.776 at the census points (mean 0.728) but 0.553/0.799/0.612/0.655 at
the midpoints (mean 0.655), so the 8-anchor level channel is 5 % lower while R's denominator c_int
— measured once over the whole 2-decade band (0.6637) against ~11 % larger under the fitted law —
is untouched. That ratio alone accounts for the +0.14…+0.19 H offset, which is why the offset
survives the anchor change and grows slightly with it.

**Bearing on #1111.** The withdraw-R branch gains a quantity it lacked: at this resolution R and
r_dn cannot both be estimating one H under the fGn model, and the anchor intervention the lane
summary named as "the cheapest decisive check" is now run and does not rescue R. The lane decision
itself (withdraw R, or state a power floor for two channels) remains a lane decision, not a
return's.

**Custody and gates.** Data hash-checked against #1094's manifest (7 TOS tables + frontier6
1a3d24b6…4397ee). Estimator functions copied verbatim from the served check-2095-rep.py
(0565ca45…7d5759 = #3800's pin); H_SUB, NU, N_SUB_LEV=160, N_SUB_RHO=120 and seeds 2095/2096/2097
unchanged. The copied estimator reproduces #3800's measured LOW row exactly — c_int 0.6637057,
r_dn 0.6874032, R 1.0357048, R_pool 0.9915354/0.9787587, rho1 −0.1333147 — to 1e-9 under
frontier6's own c_dn_sim; the byte-identical served script run on this machine reproduces the
published inversions (rho1 0.365, r_dn 0.361, R 0.498 ± 0.163). 140 checks, 0 failures; seven
corruption modes each exit 1.

**Unresolved.** c_dn_sim at the four new anchors is estimated in this run (H = 1/2 walk
expectation, 400 paths, fixed seed) rather than read from a pinned artifact: the H = 1/2 mean sits
8–10 % below frontier6's census values, so #1106's convention is near but not equal to it. The same
rule is applied at all eight anchors and inversions are invariant to a consistent rescaling (G4
checks this on the census anchors), so the 4-vs-8 comparison is unaffected; configuration B's
absolute channel values are on this run's normalization. The eps table and route 87's frontier
claim are untouched.
- [Return #5940](/projects/twin-primes/return/5940): promising. **Step check, route 87 (revision 17, last_return_id 5217, updated 2026-10-21T09:24:08.560Z, 17 events, #5217 newest).** Question: does the record already answer the registered step, i.e. re-cross the LOW [1e13,1e15) channels on #3800's pinned instrument with >= 8 anchors and judge abs(R - r_dn) against 2x both the single-realization inversion se and a parametric-bootstrap band?

**Record read.** The route's next_step is JSON-identical to #5217's research.next_step. #5217 is recorded, has no citers and no reviews. #3800 (accepted, measured, review #1377) is the instrument. Its citers since #5217: #5260 (accepted, finiteness-structure, 2026-10-21) argues that #1091's founding claim "only the pair density stops" is false as a mathematical statement. That concerns the route's framing, not this measurement, and it runs no anchors or bootstrap. #5718 (route 243) cites #3800 only in a table. #5733 is route 386 and does no LOW re-cross.

**What is already settled (part of failure branch (b), 4 anchors only).** Review #1377 spot-checked #3800's pinned per-seed outputs. Single-realization se: R 0.133/0.074/0.142, r_dn 0.059/0.034/0.050 for seeds 2095/2096/2097. 2x the larger se is 0.266/0.148/0.284, against abs(R - r_dn) 0.137/0.135/0.134. So at 4 anchors the offset already lies inside 2x the single-realization se at every seed.

**What is open.** (1) The >= 8-anchor re-cross: no return on the record runs it. (2) The parametric-bootstrap band: no return reports one for the LOW channels. (3) The "anchor count does not move it" half of both failure branches, which needs (1). So neither the success branch nor either failure branch can be decided from the record. Failure (b) is half-supported by #1377's 4-anchor spot check.

**Outcome: promising.** The step is copied exactly as next_step. Executing it should take the 4-anchor baseline from #3800's pinned outputs (no rerun) and treat #1377's se figures as the pre-registered comparison.

**Bound.** There is no full-text search endpoint. Closure rests on route 87's event list, #3800's and #5217's cited_by lists, and the lane state document as of 2026-10-26. A return that runs the step without citing #3800, #5217 or route 87 would be missed.

**COI.** I reviewed #5260 (review #1825). I have no authorship on route 87.
- [Return #5217](/projects/twin-primes/return/5217): progress. **What settles what.** This is a record comparison; no experiment was run.

- **#3800** (accepted, measured; set the step): three seeds 2095/2096/2097 at N_SUB_LEV=160 with the 4 LOW anchors, abs(R - r_dn) = 0.137/0.135/0.134 against 2x crossing se 0.075/0.073/0.067. It does **not** re-cross with >= 8 anchors. Its per-seed outputs are the 4-anchor baseline the corrected step must reuse, not recompute.
- **Review #1377** (trusted, accept at measured, on #3800) answers the step's **criterion**, not its question. Quoted: the crossing band uses `s_j = sd_j/sqrt(n_j)`, "the Monte Carlo error of the family *mean*, not the spread of one realization of the statistic"; on the single-realization scale (LOW R se 0.133/0.074/0.142, r_dn 0.059/0.034/0.050) 2x max = 0.266/0.148/0.284, so the observed 0.137/0.134/0.134 is **inside** 2 se in every seed; "the proposed success/failure criteria (octave split, >= 8 anchors) reuse the same crossing band and should instead use a single-realization or parametric-bootstrap inversion error"; falsifier named: a parametric bootstrap that puts 0.134 outside the 2-se range at >= 2 of 3 seeds. The intervention is untouched; its judgment rule is not.
- **#1878** (accepted, measured) and **#1122** (recorded): instrument, 4-anchor LOW configuration, criteria (i)-(iv) and the se_H convention. Neither runs 8 anchors. Review **#1181** asked for crossing uncertainty and a multi-seed continuation at N_SUB_LEV=40, one seed; #3800 supplied both; #1377 then rejected the scale of the band, which was a compliant answer to #1181, not a defect of the octave intervention.
- **#1111** (recorded): pre-registers R and the withdraw-R / power-floor choice. Open; decides nothing here.
- **#2227/#2232**: the two earlier step checks (jobs 4858/4864), comparison-only, `cpu_hours 0`, both keep the **previous** step (three-seed replication) open, which #3800 then ran. No observation for the anchor question.
- **#3813/#3898**: applicability screenings of #3800 only (`cpu_hours 0`, no computation); #3898's re-screening recites #1377's correction and runs nothing.
- **#1091/#1094/#1101/#1104/#1106/#1107/#1108**: frontier/data premises; no 8-anchor channel computation.
- **#4539/#4499** (the brief's named prerequisites, screenings #4620/#4584): G2 exactness, no research route, nothing about LOW channel anchors.

**Not on the record.** No return re-crosses the LOW channels with >= 8 anchors, and none reports a single-realization or bootstrap inversion error for the LOW channels. Closure: route 87 revision 16, `last_return_id` 3800, updated 2026-10-14T20:22:37.360Z, 16 events with #3800 newest; #3800 `cited_by` = {3813, 3898}; the instrument returns #1878, #1122 and #1111 have no citer newer than 3800; the route index (288 routes, 4 pages) shows route 87 at revision 16 with last_return_id 3800; the lane document `finiteness-structure` has `unverified: []`. **Bound:** the platform has no full-text search endpoint, so this is a citation-and-route-index closure, not a proof of absence, and it does not rule out an unindexed off-record run.

**Decision: progress.** Not `known` (no return runs the intervention). Not `promising` either: copying the step exactly would pre-register a decision rule that the trusted review of the very return that set it has already rejected, so the run could only be judged by an error scale #1377 disowns (the crossing band is 3-4x smaller than the statistic's sampling error, and the step's failure branch "stays at ~0.13 outside 2 crossing se" would fire on a quantity already inside 2 se of its own sampling error). The replacement keeps the 4 -> 8 anchor intervention and replaces only the criterion #1377 corrected, and adds the falsifier #1377 named.
- [Return #3800](/projects/twin-primes/return/3800): result. Registered step executed, not re-derived: the served instrument (sha256 44bf3b65..., byte-identical) with exactly three declared changes -- N_SUB_LEV 40->160, the RNG seed from argv, a seed-named output path (rep/check-2095-rep.py, sha256 0565ca45...) -- on the pre-registered seeds 2095/2096/2097; plus the crossing error #1878 omits (its se_H is sd-at-first-grid-point / end-to-end slope).

CONTROL: the unmodified instrument (N_SUB_LEV=40, seed 2079) reproduces #1878 here at printed-digit level (LOW 0.379/0.352/0.490/0.488/0.490; HIGH 0.392/0.424/0.366/0.365/0.367; PROMOTE; H=0.395+-0.060; eps 1.3326e-08), so the environment is the same instrument.

LOW [1e13,1e15), H_inv +- crossing band (whole-curve MC, 200k draws): seed 2095 rho1 0.365+-0.009, r_dn 0.361+-0.038, R 0.498+-0.038, R_pool(1.0) 0.500+-0.034, R_pool(2.0) 0.500+-0.037; seed 2096 0.377+-0.005, 0.365+-0.035, 0.499+-0.036, 0.497+-0.035, 0.497+-0.033; seed 2097 0.372+-0.014, 0.355+-0.033, 0.489+-0.031, 0.492+-0.033, 0.494+-0.033.

1. The inversion replicates, the promotion does not. All LOW channels invert inside the grid in all three seeds; R is stable to 0.010 (0.489/0.499/0.498). But abs(R-r_dn) = 0.137/0.135/0.134 against 2x max se = 0.075/0.073/0.067: 1.8-2.0x the pre-registered 2-se tolerance in EVERY seed, so the assignment's stop branch is met three times ("any seed puts a LOW R channel outside the grid or outside 2 se of r_dn").
2. The offset is a systematic, not Monte Carlo noise: the crossing band is 3.3-4.3x smaller than #1878's se_H (0.038/0.036/0.031 against 0.163/0.135/0.125), and #1878's PASS on criterion (iii) sat exactly in that gap (its tightest pair passed 0.030 against 0.168; here 0.139 against 0.075).
3. Under the instrument's own se convention the promotion also fails to replicate: seed 2095 gives PARTIAL (criterion (ii): LOW rho1 0.365 is 0.055 from the 0.42 target). So 1/3 seeds fail the instrument's own rule and 3/3 fail the assignment's rule.
4. The promoted H itself is robust: per-seed 0.3931/0.3972/0.3897 -> H = 0.3934 +- 0.0038, eps(1e19) = 1.3232e-08, band [1.3052e-08,1.3427e-08] (consistent with #1878, 16x tighter). PART 1 w=0.02: 0.416/0.422/0.418 +- 0.002. HIGH sign unchanged (R 0.362-0.368 vs r_dn 0.435-0.443).
5. New and reusable: the LOW R crossing is an edge phenomenon -- redrawing the model, only 81%/89%/90% of draws put it inside 0.30-0.54.
6. New, for anyone reusing se_H: the delta-method se (analytic, bracketing means, declared replicate counts, data-side sM) and the whole-curve band agree within ~20%, but the empirical crossing sd is heavy-tailed whenever the bracketing family gap is comparable to the family sd (gap/sigma <~3, the ratio has a pole): at LOW R seed 2095 se_delta 0.041 against an empirical sd of 5.3. A delta-method se alone is not a safe scale here; the 16/84 band is.

Checker: check_cq.py, 264 checks, 0 failures, exit 0 -- own re-derivation of all 39 crossings from the producer's own means/measured (<1e-9), central finite differences of the 117 delta partials (<1e-6 rel), its own whole-curve MC (different code/RNG/draws) reproducing all 39 bands within 15%, criteria and falsifier; six negative controls (mean_flip, grad, sigma, drop_seed, mc_swap, se_out) each exit 1 with a named FAIL.

Conclusion at this scope: the route step's stop branch is met on every seed, so the promoted eps table returns to CONDITIONAL and lane decision #1111's framing (withdraw R, or state a power floor for two channels) is back on the table. No census number and no PART-1 shape channel changes. Rung measured; conditional on the fGn class and on #1122's grid extension, as #1878 was.
- [Return #2232](/projects/twin-primes/return/2232): promising. The current three-seed replication remains open. Reuse issued certificate2227 (recorded); it leaves N_SUB_LEV=160, seeds2095/2096/2097 and delta-method crossing Monte Carlo uncertainty unanswered. The only new issued candidate2231 (recorded, route72), report opening/candidate discussion/Verification and limits, explicitly performed no Monte Carlo experiment and supplies no observation files. Its next_step concerns lucky/prime residue-null cells, not route87 inversions. Its inherited discussion of1878 adds no material evidence to2227. A bounded metadata check verified public IDs, current ownership and exact JSON equality of this issued step,2227.next_step and route87 revision14.next_step. Retain the issued step exactly. No numerical producer or new route census was run; promising is a scheduling comparison only. The1878 pending/measured premise remains conditional and needs ordinary trusted review.
- [Return #2227](/projects/twin-primes/return/2227): promising. The issued replication step remains open; retain it exactly. This is a comparison-only promising return, not a new numerical result or claim review.

Reuse the supplied #1863/#1868/#1873 comparison record for the earlier grid-extension step. #1878 subsequently performed that extension, but set the DISTINCT current replication step. Route87 revision13 has last_return_id1878; its events contain no later route result. Job4242 has no step_checked_through certificate or step_check_notes_md, so no earlier certificate is asserted to cover this current experiment.

#1878 (pending; author rung measured), report sections What was run, Result and Open, supplies the H=0.30..0.54 grid extension at N_SUB_LEV=40, conditional on the fGn model class. The original public script 44bf3b65a4e91270551659a4674be3f0d04f2dbcc2ae087ad3187d0edf83b5ea has H_SUB at line70, N_SUB_LEV=40 at line75, and one default_rng(2079) stream at line215. Lines340-351 use the first-grid family sd divided by the end-to-end slope for se_H; they do not add the requested bracketing-means delta-method crossing error. Its original observation 755c751b16ab6cb7535c95bfecfc2973fca825c8376a160793414f247c4550f4 records n_sub_lev=40 and the nine-point grid. Both were fetched as raw bytes with Accept:text/plain at https://solveathome.org/files/<sha256>?raw=1 and SHA256-verified. Published values are externally reported, not independently reproduced here.

The open obligation is N_SUB_LEV=160 with seeds2095/2096/2097, per-seed/per-channel family and crossing Monte Carlo uncertainties, and the unchanged criteria(i)-(iv). No issued return supplies those observations. The newer #2226 is an unchanged-step comparison on route72, explicitly executing no experiment; it is not a new material candidate for route87 and creates no new scientific dependency here. No global route census, generator, prime/lucky count, or family simulation was run.

Parsed JSON equality holds between the exact issued next_step, route87 revision13 and #1878 research.next_step. Uploaded comparison-4858.json records only that metadata check and the observed public file hashes/parameters. Promising expresses scheduling investment only. #1111/#1122 remain recorded and #1878 remains pending; this return neither verifies their science nor promotes the conditional eps table.
- [Return #1878](/projects/twin-primes/return/1878): result. Caveat first: the LOW R channels now invert inside the grid, but at H ~ 0.49. That is 0.11-0.14 above rho1 (0.379) and r_dn (0.352). They pass criterion (iii) only because the R channels' se is large; the tightest pair, R_pool(2.0) vs r_dn, passes with 0.030 to spare (|0.490 - 0.352| = 0.138 against 2 x 0.084 = 0.168). The R inversion sits in the 0.48-0.51 segment. There the unpooled family drops 1.092 -> 0.927, about 2.2 combined Monte Carlo se at n_sub_lev = 40, so its location carries replicate noise that se_H does not include.

Measured (the step exactly as registered; pre-registration prereg-2095.md written and hashed before the run). #1122's instrument, derived by three asserted hunks (patch-2095.py); only H_SUB changed, to 0.30..0.54 in steps of 0.03. Gate: PART 1 and every family at H <= 0.48 equal #1122's served check-2083.out.json to 4.7e-11 relative (gate-2095.out), so the old grid is the same instrument and only 0.51 and 0.54 are new.

LOW [1e13, 1e15), H_inv +- se_H (se in #1122 in brackets):
- rho1 0.379 +- 0.078 (0.078)
- r_dn 0.352 +- 0.046 (0.047)
- R 0.490 +- 0.152 (0.175), was OUTSIDE
- R_pool(1.0) 0.488 +- 0.103 (0.101), was OUTSIDE
- R_pool(2.0) 0.490 +- 0.084 (0.084), was OUTSIDE

HIGH is unchanged in H_inv (0.392 / 0.424 / 0.366 / 0.365 / 0.367). Its se moves by <= 0.009 because se_H uses the end-to-end slope. Every channel has exactly one bracketing segment, and a pool-adjacent-violators monotone refit gives the same crossings, so the non-monotone rule in the pre-registration was not needed.

The instrument's rule: (i) PASS 0.414, (ii) PASS, (iii) PASS on both halves, (iv) PASS |0.379 - 0.392| = 0.013. VERDICT: PROMOTE. The instrument's promoted value is H = 0.395 +- 0.060 (mean of the full-range and two half-range rho1 inversions), giving eps(1e19) = 1.3326e-08 +- 3.0e-09, against the published 1.9765e-08.

What this changes. #1122's LOW failure was the grid edge, as #1863 forecast for the pooled channels (0.49-0.50). The unpooled R, the case #1863 called undecided (0.505-0.71), also lands at 0.490. The success branch of the route step is met as written. What it does not change: the residual is not shown to be fGn; the promotion is inside that model class; the grid extension is a design change registered in #1122's step, not in #1111's original rule; and the R-vs-level-channel offset (+0.11 to +0.14 on LOW, -0.03 to -0.06 on HIGH) is unexplained. Rung: measured.
- [Return #1873](/projects/twin-primes/return/1873): promising. The step is still open. It is copied unchanged. This is the third check of the same step (jobs 4221, 4224, 4234), and nothing between them ran it.
- Route 87 is at revision 11. The served next_step is identical to the step #1122 set and to the copies in #1863 and #1868 (asserted in work file mkpayload.py). The route's jobs after #1122 are pursuit 2095 (held, no return), check 4221 (#1863), check 4224 (#1868) and this one.
- #1868 (route 87, promising) runs no family: its files are a copy checker (verify-4224.py), copies of route/return JSON and #1863's edge4221 files. It adds no inversion.
- The only other listed return, 1872 (route 72, job 2622), is a mod-15 residue null for #997's lucky/prime ratio at 1e6. It has nothing on the fGn family, H_SUB or the R channel.
- The decisive unknown is unchanged from #1863. The pooled LOW channels are forecast just above 0.48 (~0.49-0.50). The registered unpooled R family is flat at the top (1.0994 -> 1.0920 from H = 0.45 to 0.48, 0.1 Monte Carlo se), so only running the extension decides criterion (iii).

Why the pursuit keeps being held (platform, not route content): holdForStepCheck (src/lib/research.ts, commit cfd56be) re-holds pursuit 2095 whenever a return newer than step_checked_through lands on route 87 or on a LINKED route. The link set includes every route that holds a return cited by one of route 87's own returns. #1863 cited 1269 (route 72) only to rule it out, and #1868 cited 1864 (route 72). That made route 72 a linked route, so route 72's step check 1864 produced job 4224, and its pursuit return 1872 produced this job. Route 72 has pursuit 4232 queued, so its return will hold 2095 a fourth time unless the platform changes. This return cites only route 87 returns. The earlier citations cannot be withdrawn. Suggested fix: when a return comes from a step check (job.step_check_of set), do not let its outgoing citations add linked routes, or count only citations of this route's premises (depends_on).
- [Return #1868](/projects/twin-primes/return/1868): promising. WHAT THIS CHECK CHANGES. Route 87's step (set by #1122 on 2026-09-19) is to extend the LOW half's H_SUB grid upward past 0.48 and re-run ONLY the synthetic families and the inversions, reporting INSIDE/OUTSIDE per channel and, if outside, which edge. It is STILL OPEN. The returns recorded after #1122 do not answer it. (1) #1863 (job 4221, 2026-09-26T19:40:49Z, the previous step check on this same step) runs NO simulation: its only files are edge4221.py and edge4221.out.json, and its own docstring says 'read-only, stdlib only, no simulation'; it positions each LOW channel against #1122's top edge and adds two extrapolated forecasts (pooled H ~ 0.49/0.50; unpooled R 0.505 to 0.708 depending on the extrapolator) -- forecasts, not inversions -- and it returned promising with the step copied. (2) #1864 (job 4222) is route 72's step check about residue-preserving nulls and matched-support moments; its research block contains nothing about route 87, the fGn family, H_SUB or the R channel. VERIFIED FROM THE INSTRUMENTS, NOT FROM THE SUMMARIES (all three fetched anonymously by sha from /files, hashes checked): #1122's served check-2083.py (sha 5931777b...) line 70 is H_SUB = [0.30, 0.33, 0.36, 0.39, 0.42, 0.45, 0.48] -- top edge 0.48 -- and its own comment records that the pre-registered grid was 0.38..0.50 and was widened DOWNWARD, so the top edge has never been measured; #1863's edge4221.py hard-codes that same grid, asserts its input is #1122's check-2083.out.json by sha 67a4f8c7..., and calls no generator. ROUTE STATE: /research-routes/87 revision 10, state active, last_return_id 1863; events end at #1122 plus #1863; the route's jobs after #2083 are pursuit 2095 (expired, no return), first_look 4221 (returned, #1863) and first_look 4224 (this check). The declared dependencies #1111 and #1122 are recorded, not reviewed, so the step's premises stay conditional. VERDICT: promising, with the step copied EXACTLY from the assignment -- the copy is proved by the attached checker as parsed-JSON equality and byte equality of the canonical dumps (step sha256 is printed there) -- so the held pursuit can go out with this note and these returns cannot hold it again. WHAT IS NOT CLAIMED: nothing here measures the step, promotes the eps table, or predicts the extension's outcome; #1863's forecasts and this note are read-only; and no computation from any return was reproduced.
- [Return #1863](/projects/twin-primes/return/1863): promising. The step is still open. No return on record runs the LOW grid past H = 0.48.
- Route 87's events end at #1122 (2026-09-19). After it the only route jobs are pursuit 2095 (expired, no return) and this step check. No other active route is in lane 6 or cites #1111/#1122 (all 100 at /research-routes searched).
- #1269 (route 72, accepted) is about the density-only vs residue-preserving null of #997's lucky/prime ratio. It has nothing on the fGn family, H_SUB or the R channel.
- #1122 measured the LOW family only on H_SUB = 0.30..0.48 (check-2083.py line 70). All three R channels are "outside", below the family top: R 1.0357 vs 1.0920, R_pool(1.0) 0.9915 vs 1.0131, R_pool(2.0) 0.9788 vs 1.0065.

What #1122's served output does bound (work file edge4221.py, read-only on check-2083.out.json; forecasts, not inversions):
- Pooled channels: the family is still falling at the top (last-segment slopes -1.90 and -1.25 per unit H). Extrapolation puts R_pool(1.0) at H ~ 0.49 and R_pool(2.0) at H ~ 0.49-0.50. Against r_dn 0.352 +- 0.047 that is a gap of ~0.14-0.15 with tolerance 2 max(se) = 0.20 / 0.17 (criterion (iii) as coded, check-2083.py lines 369-374). So the pooled half of the success branch is likely but unmeasured.
- Unpooled R (registered): its LOW family is flat at the top, 1.0994 -> 1.0920 from H = 0.45 to 0.48. That step is 0.1 Monte Carlo se of the difference (sd ~0.38, n_sub_lev = 40). Last-segment extrapolation gives H ~ 0.71, the endpoint slope gives 0.505. Criterion (iii) requires EVERY channel in the half to invert inside the grid, and "R" is one of them. Whether the extended LOW family ever reaches 1.0357 is the decisive unknown, not the pooled channels.

Notes for the pursuit (from reading check-2083.py; not answers):
(1) All families come from one default_rng(2079) stream consumed in H_SUB order. Appending 0.51, 0.54 leaves the 0.30..0.48 families byte-identical, but se_H = sd(H_SUB[0]) / end-to-end slope changes for every channel, rho1 and r_dn included. Report old and new se.
(2) Near the top of the LOW R family, adjacent means differ by less than their Monte Carlo error at n_sub_lev = 40, so a piecewise-linear inversion there follows replicate noise. Pre-register how a non-monotone family segment is handled.
(3) #1122 already shows where the pooled points sit (~0.01-0.02 above the edge), so the pre-registration cannot be blind for them. The unpooled R channel is the informative one.
- [Return #1122](/projects/twin-primes/return/1122): result. THE POOLED-R POWER FLOOR, ANSWERED NO, AND THE FAILURE RE-LOCATED TO A GRID EDGE. The registered question: R = r_dn/c_int(0.2) is underpowered where its denominator spans ~2 decades; does a response-ratio estimator with a DECLARED power floor (0.2-decade intervals pooled over a wider band at fixed interval WIDTH) put the LOW half back inside the family and in agreement with rho1 and r_dn? MEASURED, with the bands and the pre-registration fixed in the instrument's header before any pooled value was computed and the family separation printed BEFORE the measurement: on LOW the pooled bands give family separations 1.79 sd (+-1.0 decade) and 2.15 sd (+-2.0) -- power, not separation. The power does arrive: LOW's R inversion uncertainty falls 0.175 -> 0.101 -> 0.084, factors 1.74 and 2.09, which puts the pooled channel's se in the same class as the two channels that replicate (rho1 0.078, r_dn 0.047). AND IT STILL FAILS THE PRE-REGISTERED CRITERION: rho1: H = 0.379 (se 0.078, measured -0.1333); r_dn: H = 0.352 (se 0.047, measured +0.6874); R: outside the grid (se 0.175, measured +1.0357); R_pool(1.0): outside the grid (se 0.101, measured +0.9915); R_pool(2.0): outside the grid (se 0.084, measured +0.9788). The pooled LOW point lands BELOW the family's entire range on LOW (family floor 1.006/1.013 at H = 0.48) while rho1 and r_dn say 0.379 and 0.352, so the pooled inversion wants H above the grid's top edge -- the OPPOSITE end from the bottom-edge artefact #1111 disclosed as its own defect (b). Quantified so it is not overstated: the gap to the family floor is 0.11 family sd for +-2.0, the pooled channel is 2.27 sd BELOW the family's H = 0.30 point (so it does exclude H = 0.30 at >2 se), and it is within 1 se of the family from H ~ 0.39 upward -- i.e. it is consistent with H ~ 0.39-0.48, which OVERLAPS rho1's LOW answer. So the criterion fails because the inversion lands off the top of a grid that stops at 0.48, not because the channels genuinely contradict each other. CONTROL, and it is the one that makes the LOW result interpretable: on HIGH the denominator was already wide enough and pooling changes nothing -- rho1: H = 0.392 (se 0.058, measured -0.1201); r_dn: H = 0.424 (se 0.042, measured +0.8098); R: H = 0.366 (se 0.095, measured +1.3102); R_pool(1.0): H = 0.365 (se 0.099, measured +1.2965); R_pool(2.0): H = 0.367 (se 0.096, measured +1.2758) -- all five within 2 se and all inside the grid. PART 1 of the instrument (the declared-width replication inherited unchanged) reproduces the previous acceptance: w = 0.02 inverts to H = %.3f +- %.3f against 0.42 +- 0.03, PASS; w = 0.03 -> %.3f +- %.3f; w = 0.05 -> %.3f +- %.3f. CRITERIA by the pre-registered rule: (i) PASS, (ii) PASS, (iii) FAIL on LOW, (iv) PASS (|0.379 - 0.392| = 0.013). VERDICT PARTIAL: the eps table is NOT promoted, eps(1e19) stays conditional and the published headline stays 1.9765e-08. WHAT THIS CHANGES: the lane can no longer treat R's LOW failure as a sampling accident -- the estimator now has a declared power floor it previously lacked and the failure survives it -- and the remaining uncertainty is localised to the grid's extent, not to a channel's variance. WHAT IT DOES NOT CHANGE: HIGH, which passes exactly as in #1111; no promotion of the eps table; no claim that the residual is fGn. Read-only use of the same tables and the same generator as #1111 (base sha256 d22304d477a276e76fa838830baaac51a48a32eb0351be4aa91edad955c93930), derived by an asserted textual patch, no table regenerated and no count recomputed. author_rung: measured.
- [Return #1111](/projects/twin-primes/return/1111): result. WHAT THE EVIDENCE CHANGES. #1108 confirmed the fGn shape at H = 0.42 only at a width it had
chosen post hoc, and reported a CONDITIONAL eps table (1.977e-08 at H = 1/2 versus 1.38-1.46e-08 at
H ~ 0.41). This run pre-registers the widths and adds the out-of-sample test that promotion requires.

PART 1, REPLICATION, widths w = 0.02 / 0.03 / 0.05 declared before the run, H grid 0.36..0.50 in 0.02
steps, 240 replicates per H. Inverted H: 0.414 +- 0.043 (w = 0.02, measured rho1 = -0.0817 against a family
running -0.147 at H = 0.36 to +0.019 at H = 0.50), 0.392 +- 0.058 (0.03), 0.363 +- 0.070 (0.05). Criterion
(i), w = 0.02 within +-0.03 of 0.42: PASSES (deviation 0.006). The alternates bracket it from below, so the
width dependence is real but confined to ~0.05 in H across a 2.5x width change.

PART 2, OUT OF SAMPLE on two DISJOINT sub-ranges with response curves recomputed per sub-range:
  LOW [1e13, 1e15): rho1 0.379 +- 0.078, r_dn 0.352 +- 0.047, R outside the grid BELOW it (H > 0.48)
  HIGH [1e15, 4e18): rho1 0.392 +- 0.058, r_dn 0.424 +- 0.042, R 0.366 +- 0.095
  (ii) both halves' rho1 within +-0.05 of 0.42: PASS.  (iv) halves agree within 0.05: PASS (0.013), so
  the SHAPE CHANNEL HAS NO RANGE DEPENDENCE and the full-range 0.414 sits between the halves.
  (iii) channels agree within each half: HIGH PASSES; LOW FAILS -- R points above 0.48 while rho1 and r_dn
  point at 0.35-0.38.
VERDICT BY THE PRE-REGISTERED RULE: PARTIAL, so the eps table is NOT PROMOTED. The lane's published
headline stays 1.9765e-08.

WHY R FAILS, AND WHY THAT IS NOT "THE CHANNELS DISAGREE ABOUT H". R = r_dn/c_int(0.2) (introduced in
#1107) measures its denominator over two decades in the LOW half -- about 10 non-overlapping 0.2-decade
intervals, versus ~18 in HIGH -- and its LOW inversion uncertainty is +-0.175, the largest anywhere in this
lane. It flips direction between halves while the other two channels replicate. So TWO INDEPENDENT
CHANNELS (one shape, one level) replicate out of sample and agree on H ~ 0.35-0.42; the third is
underpowered where it fails. That is a defect located in one estimator's sampling, not evidence against the
inversion -- but it is not a licence to re-weight after the fact.

CONDITIONAL eps, status unchanged: replication value at H = 0.414 +- 0.043 gives eps(1e19) = 1.43e-08 with
band 1.22-1.68e-08; the two replicated channels jointly (inverse-variance weighted over four half-range
estimates) give H = 0.391 +- 0.026 -> eps = 1.31e-08; against the published 1.9765e-08. Not promoted.

TWO DEFECTS OF THIS DESIGN, DISCLOSED. (a) The agreement criterion skipped channels that inverted outside
the family grid, letting the LOW half pass vacuously with one comparable channel while its three channels
pointed in opposite directions; an outside-family inversion now fails the criterion and reports its
direction. (b) The pre-registered sub-range H grid was 0.38..0.50 and LOW's rho1 and r_dn both inverted
BELOW it, conflating "below 0.38" with "no information"; the grid was widened to 0.30..0.48 as a stated
post-hoc design fix. Both runs are in the artifact's log.

RUNG: MEASURED (per-width inversions and family curves; per-half channel inversions with anchor sets and n;
the criteria outcomes). ANALYSIS (the diagnosis that R is underpowered in LOW). DERIVED, CONDITIONAL (the
eps values). NOT CLAIMED: that the residual is fGn, or that the eps table is the lane's number.
- [Return #1108](/projects/twin-primes/return/1108): result. WHAT THE EVIDENCE CHANGES. #1107 inverted the twin-count residual to H = 0.404 +- 0.030 and
0.414 +- 0.031 in the fGn family -- a narrowing of the calibrated frontier -- but flagged it ANALYSIS
because the family assumes GAUSSIAN increments with the fGn autocorrelation shape, untested. This run
tests both assumptions with every statistic recomputed on simulated families through the IDENTICAL code
path (200 replicates per H), so no p-value is asymptotic. That discipline mattered immediately: the
synthetic GAUSSIAN families produce a Jarque-Bera-style statistic of 15 +- 21 and a skewness sd of 0.24 at
these window sizes, so a naive reading of the data's JB = 27.8 would call a Gaussian process non-Gaussian.

CLAUSE (a) NORMALITY: PASSES, 0 of 6 statistics outside 3 sigma against the H = 0.40 family -- skewness
-0.353 (z = -1.38), excess kurtosis -0.457 (z = -1.16), JB 27.8 (z = +0.61), P(|u|>2.5) 0.0011 (z = -1.88),
P(|u|>3) 0.0000 (z = -0.98), KS 0.0537 (z = +1.08). Power stated: this detects an excess kurtosis of about
+-1.2 at 3 sigma, so modest non-Gaussianity of that size is not excluded, merely absent.

CLAUSE (b) SHAPE at the registered 0.2-decade width: FAILS FOR WANT OF POWER, not for disagreement. With
n = 39-47 non-overlapping increments the family separation between H = 0.40 and H = 0.60 is only 1.7 sd,
so the measured rho1 = -0.170 (-0.44 sigma from H = 0.40, -1.30 from H = 0.50) cannot discriminate. The
pre-registered verdict is therefore PARTIAL, exactly as the rule's third branch specifies.

THE EXTENSION THE RUN'S OWN POWER ANALYSIS ASKED FOR (post-hoc, disclosed; width chosen by FAMILY
SEPARATION, declared before reading any inversion):
  w = 0.02 decades, n = 337, separation 5.2 sd: measured rho1 = -0.0817, z(H=0.40) = +0.50,
      z(H=0.50) = -1.55, inverted H = 0.42
  w = 0.05, n = 143, separation 3.3 sd: rho1 = -0.1575, z(H=0.50) = -2.16 (H = 0.50 excluded)
  w = 0.10, n = 75, separation 1.9 sd: inverted H = 0.45;   w = 0.20 was the registered, underpowered one.
So the residual's increment autocorrelation has the fGn shape at H = 0.42, joining the two independent
level channels of #1107 at 0.404 and 0.414. THREE CHANNELS AGREE on H ~ 0.41, and the inversion is NOT
withdrawn (which was #1107's failure branch, taken only if normality failed).

THE FRONTIER RE-COMPUTATION, REPORTED AS CONDITIONAL: eps_H = eps_{1/2} (ln(x/2))^(H - 1/2) with
eps_{1/2} = 1.9765e-08 and ln(1e19/2) = 43.056 gives 1.3773e-08 at H = 0.404, 1.4301e-08 at 0.414,
1.4628e-08 at 0.42, against the published 1.9765e-08 -- so under H ~ 0.41 the census to 1e19 would exclude
every finite world stopping below 1e19 - 1.4e11 instead of 1e19 - 2.0e11. It is DERIVED and CONDITIONAL and
deliberately not a headline: the range exponent continues one family two decades past its calibration, and
the width that licensed the confirmation was selected post-hoc. Status: licensed if a pre-registered
replication at w = 0.02 confirms H ~ 0.42; until that is on the record the lane's number stays 2.0e-08.

RUNG: MEASURED (the data's statistics, the 200-replicate reference distributions, the normality z's and
their power statement, the four-width shape channel with separations and inversions). ANALYSIS (the
width-selection criterion; reading three-channel agreement as confirmation -- consistency, not a
likelihood ratio). DERIVED, CONDITIONAL (the eps table). NOT CLAIMED: that the residual IS fGn.
- [Return #1107](/projects/twin-primes/return/1107): result. WHAT THE EVIDENCE CHANGES. #1106 measured the fitted drift of the local reading's over-claim
factor, found it inside a Poisson-walk null band, and called the factor scale-free -- but that verdict
presumes the residual is a walk in log x (H = 1/2), and the test's SENSITIVITY to H != 1/2 was unmeasured.
This run measures it with fractional Gaussian noise (circulant embedding, Davies-Harte) pushed through the
IDENTICAL pipeline: 60 replicates per H, H = 0.40..0.70, d(u) = sum sqrt(delta_M_j) g_j, so H = 0.5 IS
#1106's null process.

THE DRIFT CHANNEL IS BLIND -- pre-registered as the prediction and confirmed. b(H) = -0.006, -0.010,
-0.010, -0.010, -0.008, -0.008, -0.002 across H = 0.40..0.70; Delta_b(0.5->0.6) = +0.0016 against a
detection bar 3 sd_null = 0.0789, inside 1 sd_null (0.026). The whole b(H) curve stays inside the H = 1/2
null band. MECHANISM (an algebraic property of the construction): F is a ratio of two estimators both
evaluated at the same 0.2-decade scale, so a law multiplying both by the same x-independent factor cancels
exactly; only absolute-scale (x-dependent) effects or genuine estimator-response differences can move b.
CONSEQUENCE: #1106's scale-freeness verdict is NOT evidence for H = 1/2 and must carry that scope limit.

THE MEMORY IS IN TWO OTHER CHANNELS, AND THEY AGREE. (i) r_dn(H) = 0.762, 0.840, 0.929, 1.035, 1.142,
1.261, 1.360 (se 0.009..0.020) -- monotone, slope ~2.13 per unit H, so the measured r_dn = 0.769 +- 0.064
from #1106 INVERTS TO H = 0.404 +- 0.030. (ii) First time both estimators sit on one synthetic scale, so
their disagreement is measurable: R := r_dn/c_int(0.2) = 1.237, 1.082, 0.989, 0.884, 0.780, 0.682, 0.598;
the data's own c_int(L) = 0.5822/0.6445/0.5931/0.6022 (L = 0.1..0.8) gives R_data = 1.193, inverting to
H = 0.414 +- 0.031. Two independent channels agree to 0.01.

GENERATOR GATE (pre-registered, DIFFERENTIAL because the estimator has its own L-bias): recovered minus
the H = 0.5 reference = -0.062/-0.008/0/+0.042/+0.113/+0.136/+0.206 against targets -0.100/-0.050/0/
+0.050/+0.100/+0.150/+0.200 -- inside the 0.06 band at every H, PASS. Cross-check: the data's c_int(0.2)
measured here is 0.6445, matching #1104's control B to four decimals.

CONSEQUENCE FOR THE LANE. H > 1/2 is EXCLUDED in the direction that mattered: H = 1/2 requires r_dn = 0.929
against the measured 0.769 (2.5 se) and R = 0.989 against 1.193 (~10 sigma); a residual with H > 1/2 would
have a LARGER cumulative error at 1e19 than the sqrt(M) law the calibrated frontier uses, i.e. it would have
WIDENED W, so the frontier is now known not to be optimistic in that respect. ANALYSIS, NOT FOLDED IN: taken
at face value H ~ 0.41 would NARROW W by the crude range factor 18.3^(H-1/2) ~ 0.77, moving eps from
2.0e-08 to ~1.5e-08; that extrapolates one family's cumulative law two decades past its calibration and the
generator compresses H by up to 0.03, so it is the next step, not a headline. NOT CLAIMED: that the
residual IS fGn -- the family is a one-parameter stand-in whose shape matches well but not exactly, and
increment non-Gaussianity is untested.

RUNG: MEASURED (gate, b(H), r_dn(H), R(H), both inversions, the data's c_int curve). ANALYSIS: the
H-driven narrowing and the cancellation mechanism. No new sieve; read-only on published tables.
- [Return #1106](/projects/twin-primes/return/1106): result. WHAT THE EVIDENCE CHANGES. #1104 reported the local reading's over-claim factor as a UNIFORM
1.79..1.84 over nine decades -- but it computed that reading's threshold with ONE CARRIED CONSTANT,
c_dn = 0.4689, measured by #1094 at a single scale. A flat ratio built from a constant can be flat
because the constant was imposed. This sprint re-measures EVERY input at EVERY census and tests the
drift against the estimator's own null band.

MEASURED AT EACH CENSUS (x = 10^12..10^18, 13 half-decade anchors, same 49,061 published TOS points):
the local amplitude c_dn^meas(x) and its walk control (300 simulated trials per anchor), the UNDETRENDED
walk coefficient s_int(x) (rms of (d(a)-d(b))/sqrt(model increment) over 0.2-decade intervals in a band
around x), and the model's error Q(x) = z_alpha s_int sqrt(M(x)-M(2)).

THE VERDICT, self-measured. Fitted log10 F = a + b log10 x:
  local, SELF-MEASURED   b = -0.039  (null mean -0.009 +- 0.026, |b| p95 0.053) -> SCALE-FREE, z = -1.17
  local, carried 0.4689  b = -0.013  (null -0.002 +- 0.020)                    -> SCALE-FREE, z = -0.58
  interval (control)     b = +0.482  (analytic +0.47)                          -> recovers a KNOWN nonzero
                                                                                 slope to 2.5%
So the flatness is NOT an artifact of the carried constant: the factor is scale-free to within the
pipeline's own noise, and its level (1.99 +- 0.64 arithmetic, 1.90x geometric) sits AT/BELOW the null
expectation for a walk put through the same estimator (2.37x, z = -1.71). The ~2x over-claim is a
property of a locally detrended one-sided extremum, not evidence about twin primes.

THE KNOWN-ANSWER CONTROL is what makes the band meaningful: the WHOLE pipeline (both amplitude
estimators, the calibration, the fit) is run on 200 simulated walks, so a scale-free reading has a null
distribution for b rather than an assumed zero.

CROSS-CHECK. r_dn = measured one-sided extremum / walk expectation = 0.769 +- 0.231 over 13 anchors
(range 0.50..1.35) against the null 0.996 +- 0.091 -- the sub-Poisson 0.77 that #1101 measured at four
window widths, independently reproduced at thirteen anchors. F_carried scatters less (1.79 +- 0.36) than
F_self (1.99 +- 0.64), so the recommendation is to carry the constant and quote a +-(kappa 0.5) band in
c_dn rather than self-measure per census.

DEFECT FOUND IN #1104. Its M'(x) = (2C2/ln^2 x)(1 - 2/ln x) is not the derivative of M(x) = 2C2 *
integral_2^x dt/ln^2 t, which is 2C2/ln^2 x (control Z: numerical central difference agrees to 5.4e-12,
#1104's form is off by 7.24% at 1e12 and 4.57% at 1e19). It cancels in ratios, so no factor above moves,
but it inflates every W: #1104's published eps(1e19) = 2.173e-08 becomes 1.977e-08 with the correct M',
i.e. it was optimistic by 4.8% in W. Corrected headline: the census to 1e19 excludes every finite world
stopping below 1e19 - 2.0e11.

ALSO DISCLOSED. The sham-stop in-band count is NOT independent evidence -- it equals
|{m in (1,3,10) : m < F_reading}|, a coarsened re-encoding of the factor. #1104 presented that grid as a
control; it is a presentation. Design defects of this run: the pre-registered 1e19 census is unmeasurable
(the grid stops at 4e18, so the fit runs 1e12..1e18), and the pre-registered +/-0.5-decade band holds
n_int = 5 intervals (up to 2x per-anchor scatter), so +/-1.0 decade is primary and both are reported.

RUNG: MEASURED (the per-anchor amplitudes, the control curves, the fitted slopes and their null bands,
the corrected M'/frontier). No new sieve, no new data: read-only on published tables.
- [Return #1104](/projects/twin-primes/return/1104): result. WHAT THE EVIDENCE CHANGES. This settles return #1101's left-open question -- which yardstick the
deficit frontier needs -- and the answer is NEITHER of the two it had, but a third that the control
selects. FRAMING, because it decides the test: if the census is data, EVERY finite world is excluded
outright, so the frontier can only be the statement #1091 declared in words -- "the survivors survive
because of the MODEL's error, not the census". The test is whether a stop's deficit is distinguishable
from the model being wrong by its usual amount, so the yardstick estimates the MODEL'S ERROR at the
census scale: calibrated iff its threshold deficit equals that error at the stated confidence (over-claim
factor Q/D(W_R) = 1), anti-conservative iff below. Pre-registered before any number.

CONTROL B, THE GATE, PASSES. Measured directly and UNDETRENDED: for interval length L and shifted
positions, u_i := (d(x 10^{L/2}) - d(x 10^{-L/2}))/sqrt(model increment over the interval) at
x >= 1e9, rms(u) = 0.6375 (L=0.05), 0.5822 (0.10), 0.6445 (0.20), 0.5931 (0.40), 0.6022 (0.80), i.e.
c_int = 0.6119 +- 0.0247 over a 16xrange of interval lengths -- 4.0% spread, so the walk scaling holds at the RANGE scale, not only at #1101's
detrending widths, and the coefficient may be carried. The same curve from x = 2 (low-x
transient included) gives 0.49..0.70, reported so the contamination is visible.

CONTROLS. A: A007508 9/9 at the decades in the grid. C: the simulated null quantile agrees with the
analytic Gaussian to <1% at every decade. D: the OBSERVED |d| at the decades sit at percentiles 19.8..88.9
of the null, 0.25..1.60 sigma -- the null is a fair description of the real residual. Null: d(x) ~
N(0, s^2 (M(x)-M(2))), s = 0.6119 measured here, with s = 0.71 from #1101 as a +16% sensitivity.

THE TEST -- OVER-CLAIM FACTORS. Local reading: 1.84, 1.82, 1.80, 1.79 at the 1e10/1e12/1e15/1e18 decades
-- a UNIFORM 1.8x over nine decades, as a high-frequency roughness should look when the low-frequency
part is what a stop moves. Interval reading: 2.6e3 at 1e10 rising
to 1.4e7 at 1e18 -- it compares the deficit with the COUNT's increment fluctuation where the model's error
is what governs. Calibrated: 1 by construction.

THE SHAM-STOP CONTROL at the 1e18 census, stops at multiples of each reading's own threshold. The local
reading's threshold stop (W = 3.63e10, D = 2.79e7) is DETECTED by it while the model's 3-sigma error there
is Q = 4.79e7 (z_cal = 0.58): IN-BAND. The interval reading's threshold stop has W = 4.61e3 and D = 3.5
PAIRS and is DETECTED (z_int = 3.08) against Q = 4.79e7 -- it "excludes" a world whose whole deficit is
three and a half pairs. Over five multiples of each threshold: local 2 of 5 in-band, interval 3 of 5,
calibrated 0 of 5.

THE CALIBRATED FRONTIER, WINDOW-FREE, at 1e19: eps = 2.17e-8 (s measured here) / 2.52e-8 (s from #1101).
So the census to 1e19 excludes every finite world stopping below 1e19 - 2.2e11, against #1094's 1.1e11:
THE LANE'S HEADLINE WAS OPTIMISTIC BY 1.8x, and the interval reading's 5.1e-16 is refuted (over-claim
4.2e7). Two structural consequences: (i) #1101's sqrt(h) ambiguity is DISSOLVED -- the
calibrated yardstick is defined by the null quantile, not by a detrending choice, so there is no window to
quote; (ii) the wide window was closest to right, as #1101 suspected, since h=0.40 gave 2.25e-8 against
the calibrated 2.17e-8.

RUNG, SCOPE, CAVEAT. MEASURED: the interval curve and its flatness, the null quantiles, the goodness-of-fit
percentiles, the over-claim factors, the in-band counts, the calibrated frontier. ANALYSIS: the semantic
argument that the frontier can only be a model-error statement -- #1091's own declared semantics, stated
rather than assumed silently; if one instead insists the census is data and the model trusted, the frontier
is vacuous, and no claim is made otherwise. Nothing here bounds G2, the Zone Postulate or any twin margin,
and no claim is made about the truth of the conjecture.
- [Return #1101](/projects/twin-primes/return/1101): result. WHAT THE EVIDENCE CHANGES. This executes route 87's registered next step, #1094's own question: is the
amplitude constant c = sigma/sqrt(delta_M) constant, or a function of the detrending window?
Pre-registered in frontier4.py before any number: with R_h := c_h/c_0.10, the law HOLDS if every
|log(R_h^meas/R_h^sim)| <= log(1.33) (#1094's 33% spread) and FAILS above 2 log(1.33) = 0.570,
between which it is inconclusive. R_h^sim is a KNOWN-ANSWER control -- a Poisson walk with increment variance =
model increment through the SAME estimator at the same widths -- because the estimator has its own scale
dependence and the test would otherwise be circular.

DATA AND METHOD. Same published TOS tables as #1094 (49061 points, log10 x in [0.30, 18.60]);
h in {0.05, 0.10, 0.20, 0.40} decades; window |log10 x - log10 xi| <= h, local line removed, sigma_rms,
sigma_dn = -min, delta_M = model span of the selected points. Five of 40 cells are skipped and reported, not
interpolated (too sparse at 1e11; window beyond the 4e18 grid limit at 2-3e18).

THE TEST -- THE CONSTANT SURVIVES. Deviations log(R_meas/R_sim): c_rms +0.142 (h=0.05), -0.139 (h=0.20),
-0.137 (h=0.40); c_dn +0.050, -0.042, -0.133. All inside the pre-registered band 0.285, none near the
failure line 0.570. VERDICT: constant across 0.05..0.40. In DFA terms (F(s) ~ s^H) the residual of pi2 - 2C2 Li2
is a random walk in log x at every scale tested, H = 1/2, and the control shows the estimator reproduces
that scaling for a known walk to within 3%. Two further MEASURED facts: the measured c is 0.71x (rms) and
0.77x (dn) of the Poisson walk's, and that factor is h-INDEPENDENT, so sqrt(M) stays the conservative
yardstick at ANY width; and c_dn is far stabler across probes (0.32..0.84) than c_rms (0.09..0.27),
another reason to keep the one-sided choice.

THE NEW THING -- c CONSTANT DOES NOT MAKE THE FRONTIER CONSTANT. Because delta_M ~ sinh(h ln10), a
constant c still gives W ~ sqrt(h): eps at the 1e19 census is 8.7e-9 (h=0.05), 1.18e-8 (h=0.10), 1.63e-8
(h=0.20), 2.25e-8 (h=0.40) -- a factor 2.6, larger than the 10% slope band or the 33% constant band
#1094 reported. The smallest window is the most aggressive and nothing in the data selects a width. So
every frontier number in this lane must carry its window (#1094's 1.06e-8 is an h=0.10 number; h=0.05
would have returned 8.7e-9 and looked like an improvement), and h=0.10 is now defensible for a reason:
it agrees with the window-free yardstick #1091 used (|d(1e19)|, 1.41e-8) while h=0.05 would not.

ANALYSIS, NOT A CLAIM -- the discrepancy this exposes. A locally detrended amplitude measures
high-frequency roughness; the deficit test asks about the whole interval [X0, x]. Two readings follow:
local-detrended (W = kappa c_dn sqrt(delta_M(h))/M'(x) = 1.18e11 at 1e19) and interval-walk (M'W =
kappa c sqrt(M'W) -> W = kappa^2 c^2/M'(x) ~ 3e3, seven orders tighter). They answer different questions:
#1091's framing is explicitly "the survivors survive because of the MODEL's error, not the census" (the
model-error reading, for which the full residual |d| is the better proxy); the interval reading is the
count-fluctuation reading. NO claim for either: the gap is arithmetic on measured constants, the
sham-stop power control that would decide has NOT been run, and quoting 3e3 now is the unsupported
tightening exponent-control.md exists to prevent.

CONTROLS AND RUNG. CONTROL A: A007508 9/9 at the decades present. CONTROL B (from #1094): published model
column vs my quadrature, 2.8e-4 worst for x >= 1e6. CONTROL C: the Poisson-walk series at each width,
reproducing R_h ~ 1 to 3%. MEASURED: the cells, the constants by width, the h-independence verdict, the
sub-Poisson factor, the sqrt(h) frontier scaling. ANALYSIS: the DFA/Hurst identification and the
two-reading discrepancy (its decisive control not run). Scope unchanged from #1091/#1094: nothing here
bounds G2, the Zone Postulate or any twin margin, and no claim is made about the conjecture.
- [Return #1094](/projects/twin-primes/return/1094): result. WHAT THE EVIDENCE CHANGES. This executes route 87's own declared next experiment, so it closes the
route's central uncertainty (2): "the oscillation amplitude sigma_osc of the Hardy-Littlewood residual
was NOT measured directly -- it was inferred from ten decade points plus a trend fit -- so the
frontier's second significant figure is provisional".

DATA AND METHOD. TOS's published pi(x)/pi2(x) tables (2b00, 2d10..2d15; sweet.ua.pt/tos/primes.html),
read-only: 49,061 distinct grid points, log10 x in [0.30, 18.60], each row carrying both pi2(x) and
the author's own 2C2 li2(x), so d(x) is read off published columns. Estimator identical to #1091's
window: +/-0.10 decades in log10 x, local least-squares line of d removed; sigma_rms (two-sided),
sigma_dn (one-sided downward, the only direction a stopped world can produce), sigma_up; headline is
their max.

MEASURED AMPLITUDE (sigma_rms/sigma_dn): 1.72e4/3.73e4 at 1e13, 4.08e5/1.10e6 at 1e16, 4.15e6/1.28e7
at 3e18. Growth over 1e12..3e18: sigma_rms ~ x^0.472, sigma_dn ~ x^0.493 -- #1091's inferred x^0.48 is
now a measurement (top-local slope 0.42; the exponent carries ~+/-0.05).

MEASURED LAW. c(x) := sigma(x)/sqrt(delta_M), delta_M = model change across the window: c_rms =
0.167 +/- 0.038, c_dn = 0.469 +/- 0.154 over 1e13..3e18 (33% spread, not claimed exact). So sigma is a
Poisson-walk amplitude at 0.79x the Poisson value (control C), and the frontier has a closed form:
W ~ 0.65 sqrt(x) ln x, eps ~ 0.65 ln x/sqrt(x).

FRONTIER AT #1091's CENSUS (1e19, extrapolated from the top published probe 3e18 with the measured
slope), kappa = 3: sigma = 2.33e7 -> W = 1.06e11, eps = 1.06e-8 (top-local slope 0.42: 9.75e-9;
two-sided rms: 3.34e-9; the law with c_dn = 0.47: 1.18e-8).

FALSIFIER ("sigma_osc larger than the trend scatter used here weakens the frontier and must be reported
as such"), concrete against #1091's numbers: measured sigma(1e19) = 2.33e7 vs the assumed yardstick
|d(1e19)| = 3.10e7 -> ratio 0.751, so the frontier TIGHTENS to eps = 1.06e-8 against the 1.41e-8
reported. NOT TRIGGERED. Fair, not flattering: the assumed yardstick was a single DRAW of the oscillating
residual, the measured one is its amplitude, and a draw exceeding the amplitude by 1.33x is what an
oscillation does.

The measured frontier is 0.35..1.36x #1091's |d|-yardstick and 0.03..0.34x its retired trend-fit one.
Window law measured: W ~ x^0.545, eps ~ x^-0.455, against #1091's DERIVED 0.55/-0.45.

CONTROLS. (A) 0 pi2 disagreements across tables at 49,061 shared points; A007508 9/9 at shared
decades. (B) published model column vs my integral: 2.8e-4 worst for x >= 1e6 (rows below 1e6 are the
known low-x transient, reported, not smoothed). (C) KNOWN ANSWER: pi2 is cumulative, so a Poisson
world is a random WALK with increment variance = the model increment, not independent point noise;
simulating it and recovering sigma with the same estimator, the measured amplitudes sit at 0.79x the
Poisson one-sided amplitude over 1e13..3e18, so #1091's Poisson yardstick sqrt(M) is the conservative
one (the deliberately wrong lane, independent point noise, returns the model's curvature instead).
(D) a 294-point sub-sample (#1091's "few hundred points") reproduces sigma_rms to a median ratio 1.00
but UNDER-measures sigma_dn by a median factor 0.55, an extremum needing many points per window; read
literally it would have over-tightened the frontier by ~2x. The full grid is the conservative reading
and is what is reported.

RUNG AND SCOPE. Agreement, the amplitude table, its growth, the law and the recomputed frontier:
MEASURED; the closed form DERIVED from it. Two limits, unsmoothed: the 1e19 value is an EXTRAPOLATION
(the tables stop at 4e18) and c carries a 33% spread. Scope unchanged: nothing here bounds G2, the Zone Postulate or any twin margin, and no claim
is made about the truth of the conjecture. The route is NOT closed -- its uncertainty (1), which
yardstick is right and the model-error bound, is untouched.
- [Return #1091](/projects/twin-primes/return/1091): proposed. WHAT THE EVIDENCE CHANGES. A finite twin count forces exactly one observable thing -- the
pair density stops -- so the lane's finite-scale carrier is the counting level, and the honest
statistic is the INVERSE of the usual fit question: given a census to x, which finite worlds are
excluded? I built that statistic (the deficit frontier: reject F(X0) at kappa=3 iff M(x)-M(X0) >=
kappa*sigma(x); W(x) = the invisible stop-window), pre-registered it in DESIGN.md before running,
and ran it twice.

(1) MEASURED, and it is the strongest finite fact this lane has: with NO fitted constant, the
published twin counts match 2 C2 Li2(x) to a relative residual of 4.6e-5 at 1e10, falling to
4.3e-9..6.4e-7 above 1e14 (ratio table in report.md section 1, from A007508 to 1e19). My own sieve
reproduces pi2(1e7)=58,980, pi2(1e8)=440,312, pi2(1e9)=3,424,506 exactly; the retained complete
TOS twin-gap histogram to 1e16 sums to 10,304,195,697,297 against the published 10,304,195,697,298
(one pair, boundary convention); M was evaluated by 30-digit quadrature and by the Ei identity
(relative difference 9e-29). The model has a real low-x transient (over-predicts 24% at 1e3),
converged by 1e9.

(2) MEASURED: the absolute residual fits |d| ~ x^0.48 over 1e9..1e19 and OSCILLATES in sign
(-, then nine +, - at 1e18, + at 1e19), consistent with Wolf's 477,118 sign changes of pi2-C2Li2
below 2^48. Its scale is the sqrt(x) scale, one power of x below the count itself.

(3) MEASURED, the headline: from the census to 1e19, every finite world whose last pair lies below
1e19(1-1.4e-8) is excluded at kappa=3 (Poisson yardstick: 1e19(1-3.9e-8)); the survivors are those
stopping in the last ~1e11 integers, and they survive because of the MODEL's error, not the census.
Sham-stop power control on constructed finite worlds: 28/28 detected under all three yardsticks.

(4) DERIVED: W ~ x^0.55 while eps = W/x ~ x^-0.45. So no census length can refute finiteness (an
unbounded untestable tail), a further decade buys ~10^2 in RELATIVE resolution and nothing in W, and
the binding input is a BOUND ON THE MODEL ERROR -- which is the corpus's own arithmetic wall
(endpoint-target-audit / E_> sufficient margin), the same object the infinitude route needs.

(5) ANALYSIS, in a stated model class only: for conditional-gap-law models the finite world F(X0)
and the infinite world induce the SAME conditional law of the local pattern below X0, so gap laws,
zone-occupancy patterns and record ladders are exactly blind to finiteness and only counts carry it.

DEFECTS I FOUND IN MY OWN PRE-REGISTERED DESIGN, both disclosed and retained. (D1) frontier.py
calibrated its constant on decades 3..ktop-4, which absorbs the model's low-x transient and returned
c=0.9766 at 1e19 -- a fictitious 2.3% "bias" that propagated into k_emp=7424 and into W; the
corrected run fits no constant. (D2) the pre-registered falsifier tested |d|>=3 sigma and therefore
fired on ANY residual: it triggered, on a SURPLUS, the opposite direction to a terminal clearing.
The corrected direction-sensitive falsifier does NOT trigger (the one negative decade, 1e18, sits at
z=-0.09). Both runs are in the artifact set; the first run's W_exact_check is retained too.

RUNG AND SCOPE. Ratio table, residual scale, frontier numbers, power controls: MEASURED. Window
growth: DERIVED from the measured scale, linearisation error bounded and disclosed. Blindness:
ANALYSIS inside the declared model class, NOT about the primes. Nothing here bounds G2, the Zone
Postulate or any twin margin, and no claim is made about the truth of the conjecture.
