Antihydra is a six-state Turing machine whose fate is equivalent to one exact
question about odd and even terms in a growing integer sequence. Step the recurrence from 8,
watch its balance move toward or away from halting, then run the machine's transition table
and check that its tape decodes to the same arithmetic frames. The browser computes both
layers live. Finite checks do not decide whether Antihydra ever halts.
Will some finite prefix contain strictly more than twice as many odd terms
as even terms? One additional odd term at balance zero would end the machine.
I. The orbitGive the sequence another term
Begin at x = 8. Replace x by
floor(3x / 2). Before moving, an even term adds two to the safety balance.
An odd term removes one. The halt line is exactly b = -1.
H(x) = floor(3x / 2) | b = 2E - O | halt when O > 2E
Live Antihydra parity recurrence
Current x
computing
Terms used
computing
Even E
computing
Odd O
computing
Balance 2E - O
computing
Computing the seed.
halt: -1live scale
even adds 2odd removes 1
The values grow, so the page uses exact BigInt arithmetic. The colored row keeps
only the latest terms on screen, but the counts and balance retain the whole prefix since the
canonical seed. Running longer is evidence about that finite prefix only.
II. The machine underneathMake the tape confess the same arithmetic
A recurrence could have been pasted onto any machine story. Here the browser
parses the published six-state code and executes its transitions on a blank tape. Whenever
the tape has Ligocki's recognizable frame, the page decodes the two runs of ones into
(x, b) and compares them with the recurrence.
Raw transition table
state
read 0
read 1
Each instruction writes a bit, moves left or right, then changes state.
An undefined instruction halts.
Tape window
Parsing the machine.
frame
raw step
decoded x
decoded b
recurrence expects
result
The finite audit has not run yet.
For the audit range, every transition is executed. Nothing is interpolated. Beyond the
displayed range, the recurrence instrument applies the closed rule directly and says so.
That division is a free engineering choice: it keeps the page responsive while leaving a
finite, inspectable bridge between tape and arithmetic.
The check
The browser reruns these checks from empty state when the page opens. Published inputs are
identified separately from values derived here.
Recurrence prefix
Computing.
Balance identity
Computing.
Raw tape frames
Computing.
First possible crossing
Computing.
HEURISTIC, NOT A PROBABILITY THAT THE THEOREM IS TRUE
Computing the published random-walk expression.
The arXiv paper prints approximately 4.84 x 10^-224394395.
Direct high-precision recomputation of its displayed expression gives a mantissa near
9.05 at that exponent. We could not reconcile the printed approximation, so
the live result above is left visible rather than adjusted to match it.
Published comparison. The peer-reviewed STOC paper proves
S(5) = 47,176,870 after a Tree Normal Form enumeration of
181,385,789 five-state machines. Those two values are cited inputs, not results
of this page. They provide context only. Antihydra has six states.
Reported finite run. Ligocki's original analysis says mxdys simulated
2^31 = 2,147,483,648 recurrence iterations and reached
b = 1,073,720,884. This page cites that report. It does not pretend to rerun
it, and the report does not prove nonhalting. The paper's heuristic exponent is one more
than that reported balance.
Dated frontier. The BusyBeaverWiki holdout table listed
1,064 BB(6) holdouts on 28 July 2026. The table explicitly allows
community submissions without independent verification. That count is mutable, informal,
convention-dependent, and not fetched live here.
Free choices, conventions, approximations, and limits
The canonical experiment starts from 8. Results for any other seed would not answer the Antihydra question.
A term is counted before applying H. Thus the seed 8 is the first counted term. E and O are exact prefix counts under that indexing convention.
The tape is bi-infinite, initially all zero, with state A and the head at position zero. A raw step writes, moves, then changes state. An undefined transition halts.
The visible tape window is cropped, but the simulator stores every nonzero cell and frame decoding checks the full stored tape.
The raw macro audit is finite. It does not import the complete formal proof of the equivalence, so it cannot establish the rule for every possible frame.
Any finite recurrence run is compatible with a later halt. It cannot prove nonhalting.
The random-walk model treats a deterministic parity process as if it followed a probabilistic model. No independence theorem justifies that move. Its tiny output is not a posterior probability, p-value, confidence level, or proof.
Settling Antihydra is necessary for determining S(6), but not sufficient. Other holdouts remain, and a different halting machine could exceed the current lower bound.
Frame decoding convention, and it is not a free choice. In a recognizable frame the tape carries two runs of ones: the right run of length a encodes the orbit value as x = a + 4, and the left run of length b encodes the parity balance. A frame is recognized only in state E, reading a 0, with a at least 4 and no ones outside the two runs. The offset of 4 is forced: any other offset fails to reproduce floor(3x/2) across the nine audited frames, so the tape and the recurrence would disagree.
The raw machine simulation stops at a finite safety cap of 1,000,000 transitions, which is reached at frame 13 (x = 1021). Beyond that the frame walker cannot advance, and that is a limit of this page, not of the machine.
The page uses no mutable remote data and makes no third-party request. It does load same-origin assets (the site fonts, the favicons and the shared top bar), and none of them carry data into the computation. Source links are ordinary navigation links.
Still open. No source checked on 29 July 2026 establishes whether the
seed-8 orbit ever reaches O > 2E, equivalently whether this machine halts.
The exact value of S(6) is also open. Antihydra is Collatz-like, not equivalent
to the classical Collatz conjecture.
Shawn Ligocki,
BB(6) is Hard (Antihydra), 6 July 2024. Original public analysis, not peer-reviewed.
It records mxdys's 28 June 2024 discovery, Racheline's high-level rules, the transition
table, the recurrence, and the reported finite trajectory.
Busy Beaver Challenge, Antihydra.
Community project page, not peer-reviewed. It publishes the short program with
a = 8, b = 0, even update b += 2, odd update
b -= 1, and a += floor(a/2).
BusyBeaverWiki, Holdouts lists.
Mutable community table, not independently verified. The dated count is context, not a theorem.
The offline differential verifier is at
research/one-more-odd-number/verify-one-more-odd-number.mjs.
Corrected 2026-09-28 (found by a scan of the site's pages for verdicts printed whatever the page computed): after the raw audit, the status line said the last decoded frame matches the recurrence without reading the comparison the audit had just made. The claim was true, but the page did not check it; the status line now reads that comparison.
The claims record48 claims re-read against their sources, 18 October 2026
Written 2026-07-29. Claims re-read against their sources on 2026-10-18: 48 checked, 39 confirmed, 5 wrong, 0 unverifiable, 4 first-hand observations checked against their record. By the assay line (research/assay-line/claude-1018b/), Claude fixer (claude-opus-5-5) directed by claude-assaying-boyle-1014; source record by a Claude agent (claude-sonnet-5-5), checked by a Claude agent (claude-opus-5-5).
This was the assay line's source-first re-read of The Machine Waiting for One More Odd Number (batch claude-1018b, tier B). A Claude agent wrote the source record (research/assay-line/claude-1018b/one-more-odd-number/N1-sources.json); a second Claude agent took 48 claims to their sources, recreated the machine and the heuristic independently, and ran the reader check from /checks/ (N2-claude.md); the Claude fixer then re-read the two WRONG findings, decided all six in triage.json, and fixed what it decided. The 38 CONFIRMED and 4 OBSERVED lines below are the checker's, carried forward as it wrote them. One WRONG claim (a line the shared placard generator writes) is held under board note 81816d, and four MINOR lines were listed without edits, as tier B asks. A second claim, that the page makes no third-party request, was fixed with a dated Corrected line and landed, then reverted by the directing instance and filed CONFIRMED, under the human's ruling of 2026-09-28 (board note 706d1d) that the host is not a third party. MINOR lines are counted in wrong.
Claims
CONFIRMED Antihydra is a six-state Turing machine whose fate is equivalent to one exact question about odd and even terms in a growing integer sequence. https://arxiv.org/html/2509.12337v2 : [dek-equivalent] Section 1.2: "is a machine that does not halt from the all-zero tape if and only if the following Collatz-reminiscent conjecture (which was discovered from analysing the machine) holds"; Conjecture 1.1: "Iterating H from x = 8, there are never (strictly) more than twice as many odd numbers as even numbers."
CONFIRMED Will some finite prefix contain strictly more than twice as many odd terms as even terms? https://bbchallenge.org/antihydra : [question-prefix] "will there eventually be a point where you have seen (strictly) more than twice as many odd numbers as even numbers (if so, halt)?"
CONFIRMED Begin at x = 8. https://bbchallenge.org/antihydra : [seed-8] "Starting with the number 8 ... (8->12->18->27->40->60->90->135->202...)"; also arXiv Conjecture 1.1 "from x = 8"
CONFIRMED Replace x by floor(3x / 2). https://bbchallenge.org/antihydra : [rule-floor] "repeatedly adding half of the number to itself, rounding down"; program "a += a // 2"; Ligocki "f(n) = n + floor(n/2)"; arXiv Appendix C "H(n) = floor(3n/2)"
CONFIRMED Before moving, an even term adds two to the safety balance. An odd term removes one. https://bbchallenge.org/antihydra : [balance-rule] "if a % 2 == 0 : b += 2 else : b -= 1" before "a += a // 2"
CONFIRMED One additional odd term at balance zero would end the machine. https://www.sligocki.com/2024/07/06/bb-6-2-is-hard.html : [one-odd-ends] "A(2n+1, 0) -> Halt(3n-3)"; arXiv Appendix C "A(0, 2k+1) -> Halt" (balance run first in its notation)
CONFIRMED 1RB1RA_0LC1LE_1LD1LC_1LA0LB_1LF1RE_---0RA https://bbchallenge.org/antihydra : [code-string] "Machine: 1RB1RA_0LC1LE_1LD1LC_1LA0LB_1LF1RE_---0RA"; identical in Ligocki, the arXiv paper and the page's MACHINE_CODE (index.html:615)
CONFIRMED the right run of length a encodes the orbit value as x = a + 4, and the left run of length b encodes the parity balance. https://www.sligocki.com/2024/07/06/bb-6-2-is-hard.html : [frame-runs] "E(a, b) = 0^inf 1^b 0 1^a E> 0^inf" and "A(a, b) = E(a-4, b)", "Start ->11 A(8, 0)"; recreated: my simulator decodes x = a + 4 at all 14 frames
CONFIRMED The arXiv paper prints approximately 4.84 x 10^-224394395. https://arxiv.org/html/2509.12337v2 : [heuristic-printed] Section 1.2: "minuscule, namely, ((sqrt5 - 1)/2)^1073720885 ≈ 4.84 × 10^-224394395"
CONFIRMED Direct high-precision recomputation of its displayed expression gives a mantissa near 9.05 at that exponent. research/one-more-odd-number/verify-one-more-odd-number.mjs : [heuristic-recomputed] recreated: Python Decimal at 80 digits gives log10 = -224394394.0432687872..., i.e. 9.0517221142 x 10^-224394395; verifier asserts mantissa 9.05 and passes on the repository bytes
CONFIRMED The peer-reviewed STOC paper proves S(5) = 47,176,870 after a Tree Normal Form enumeration of 181,385,789 five-state machines. https://arxiv.org/abs/2509.12337 : [s5-enumeration] "We prove that S(5) = 47,176,870 using the Coq proof assistant. The proof enumerates 181,385,789 Turing machines with 5 states". Read in the arXiv version; Semantic Scholar joins the DOI and arXiv:2509.12337 as one paper; the STOC PDF itself returned 403
CONFIRMED after a Tree Normal Form enumeration of https://arxiv.org/html/2509.12337v2 : [tnf] "machines are enumerated arborescently in Tree Normal Form (TNF) ... from 16,679,880,978,201 5-state machines to "only" 181,385,789"
MINOR The bbchallenge Collaboration et al., Determination of the Fifth Busy Beaver Value, STOC 2026, pages 565 to 574, DOI 10.1145/3798129.3806376. https://api.crossref.org/works/10.1145/3798129.3806376 : [stoc-cite] Title, "Proceedings of the 58th Annual ACM Symposium on Theory of Computing", page "565-574" and DOI all match. But the record for this DOI lists 16 named authors, first "Justin Blanchard", and no collaboration author (OpenAlex the same); "The bbchallenge Collaboration" is the first author of the arXiv byline. The STOC byline could not be read (403). A citation headed by an author the DOI record does not carry. Tier B: listed, not edited.
CONFIRMED Peer-reviewed conference paper, published 9 June 2026. https://api.crossref.org/works/10.1145/3798129.3806376 : [stoc-date] "published":{"date-parts":[[2026,6,9]]}; type proceedings-article, event "STOC '26"
CONFIRMED arXiv:2509.12337v2, revised 23 March 2026, https://arxiv.org/abs/2509.12337 : [arxiv-v2] "[Submitted on 15 Sep 2025 (v1), last revised 23 Mar 2026 (this version, v2)]"
MINOR Appendix C states the Antihydra equivalence. https://arxiv.org/html/2509.12337v2 : [appendix-c] The excerpt is real and supports the substance, but it is in Appendix C of the arXiv version ("halts if and only if it reaches a config A(0,2k+1). This can be reformulated as whether repeatedly applying the Collatz-like Hydra map"). The page puts this sentence in the STOC citation, immediately before "The longer arXiv:2509.12337v2 ... gives". The STOC paper is 10 pages (565 to 574) and could not be read, and the layer's own README says Appendix C was used from the arXiv preprint. Tier B: listed, not edited.
CONFIRMED It records mxdys's 28 June 2024 discovery, Racheline's high-level rules, the transition table, the recurrence, and the reported finite trajectory. https://www.sligocki.com/2024/07/06/bb-6-2-is-hard.html : [ligocki-records] "This TM was discovered and shared by @mxdys on 28 June 2024 ... Racheline (@rae) discovered and shared the precise rules described below"; the wiki uses the page's word: "Racheline discovered the high-level rules soon after"; table, recurrence and Trajectory section all present
CONFIRMED Ligocki's original analysis says mxdys simulated 2^31 = 2,147,483,648 recurrence iterations and reached b = 1,073,720,884. https://www.sligocki.com/2024/07/06/bb-6-2-is-hard.html : [ligocki-2-31] "@mxdys has simulated this iteration out to 2^31 = 2 147 483 648 steps at which point b = 1 073 720 884"; Ligocki's b is the page's b (A(2n+1, 0) halts)
CONFIRMED The paper's heuristic exponent is one more than that reported balance. https://arxiv.org/html/2509.12337v2 : [exponent-plus-one] 1,073,720,885 - 1,073,720,884 = 1; the wiki's model gives the reason: "P(n) = ((sqrt5 - 1)/2)^(n+1)" (https://wiki.bbchallenge.org/wiki/Antihydra)
CONFIRMED a = 8, b = 0, even update b += 2, odd update b -= 1, and a += floor(a/2). https://bbchallenge.org/antihydra : [project-page-program] "a = 8 b = 0 while b != - 1 : if a % 2 == 0 : b += 2 else : b -= 1 a += a // 2"
CONFIRMED The BusyBeaverWiki holdout table listed 1,064 BB(6) holdouts on 28 July 2026. https://wiki.bbchallenge.org/wiki/Holdouts : [holdouts-1064] row "July 28th, 2026 @mxdys 1064 BB6_holdouts_1064.txt"; the revision of 2026-07-29T00:16:23Z (wiki API) shows 1064 in the summary table. Dated on the page; the newest row is now "October 16th, 2026 @mxdys 663"
CONFIRMED The table explicitly allows community submissions without independent verification. https://wiki.bbchallenge.org/wiki/Holdouts : [holdouts-unverified] "This is a table where people have added lists with no restriction or independent verification."
CONFIRMED convention-dependent https://wiki.bbchallenge.org/wiki/Holdouts : [holdouts-convention] "Some lists exclude known equivalent machines, i.e., machines whose halting status is implied by the halting status of another machine."
CONFIRMED Settling Antihydra is necessary for determining S(6) https://bbchallenge.org/antihydra : [necessary-for-s6] "So it is necessary to understand its behavior in order to conclusively prove the value of BB(6)."
CONFIRMED Other holdouts remain, and a different halting machine could exceed the current lower bound. https://arxiv.org/html/2509.12337v2 : [other-holdouts] "about 1,214 holdouts remain to date ... The possible existence of a halting machine that exceeds the current 2↑↑↑5 champion"
CONFIRMED The exact value of S(6) is also open. https://wiki.bbchallenge.org/wiki/Holdouts : [s6-open] 6-state 2-symbol holdouts: 663 at fetch; arXiv "we predict that S(6) will never be proved"
CONFIRMED No source checked on 29 July 2026 establishes whether the seed-8 orbit ever reaches O > 2E https://wiki.bbchallenge.org/wiki/Antihydra : [still-open] wiki at fetch (2026-10-18): "Unsolved problem: Does Antihydra run forever?"; bbchallenge: "even attempts at a proof have been elusive"
CONFIRMED Antihydra is Collatz-like https://bbchallenge.org/antihydra : [collatz-like] "It's Collatz-likeness seems to indicate that..."; arXiv "the Collatz-like map H"; Ligocki "Consider the Collatz-like function"
MINOR not equivalent to the classical Collatz conjecture. https://arxiv.org/html/2509.12337v2 : [not-classical-collatz] No source states non-equivalence. The arXiv says only "Collatz-reminiscent" and "because of resemblance to the notoriously hard Collatz conjecture"; Ligocki and bbchallenge say "Collatz-like". The searches work (each finds those phrases); none finds an equivalence or a non-equivalence. The README's own narrowing is weaker: "The page does not claim equivalence with the classical Collatz conjecture." Tier B: listed, not edited.
CONFIRMED The page uses no mutable remote data and makes no third-party request. git show e7fb271c26:public/strata/one-more-odd-number/index.html : [beacon-no-third-party] True of the committed file: its only src/href off this origin are the six source links; scripts and fonts are same-origin (/media/home-banner.js, /fonts/fonts.css), and home-banner.js names only https://artwaste.land. Not true of the served page: see beacon-served
CONFIRMED makes no third-party request. src/pages/privacy.astro : [beacon-served] Checker filed WRONG: the served HTML (curl 2026-10-18) carries Cloudflare's Web Analytics beacon, which the committed file does not (the site adds it to every page). Overturned by the directing instance: the human ruled on 2026-09-28 (board note 706d1d) that the host is not a third party and that "no third-party request" sentences stand, read through /privacy/. The fix that had landed was reverted.
CONFIRMED The offset of 4 is forced: any other offset fails to reproduce floor(3x/2) across the nine audited frames, so the tape and the recurrence would disagree. public/strata/one-more-odd-number/index.html : [offset-forced] recreated: run lengths a = 4, 8, 14, 23, 36, 56, 86, 131, 198; the first pair 8 + k = floor(3(4 + k)/2) allows only k = 4 or 5, and k = 5 fails at 23 -> 36 (41 against 42); my simulator finds only k = 4 in -3..11
CONFIRMED A frame is recognized only in state E, reading a 0, with a at least 4 and no ones outside the two runs. public/strata/one-more-odd-number/index.html : [recognise-rule] decodeFrame (index.html:670-692): "machine.state !== 'E' || readCell(machine, machine.head) !== 0", "if (a < 4 || ...", and the leftEdge/rightEdge loop
MINOR The raw machine simulation stops at a finite safety cap of 1,000,000 transitions, which is reached at frame 13 (x = 1021). public/strata/one-more-odd-number/index.html : [cap-frame-13] Cap and x are right: recreated, the last frame before the cap is x = 1021 at raw step 624,812, and the next (x = 1531) is at 1,405,105. But the page's own audit table numbers x = 8 as frame 0 (index.html:712 "frame:0"), under which x = 1021 is frame 12 and frame 13 is x = 1531, past the cap. "frame 13 (x = 1021)" mixes the two numberings. Tier B: listed, not edited.
CONFIRMED For the audit range, every transition is executed. Nothing is interpolated. public/strata/one-more-odd-number/index.html : [nothing-interpolated] runToNextFrame calls machineStep once per transition and auditRawFrames loops it; my independent step-by-step simulator reaches the same nine frames at the same raw steps (11 to 24,393)
CONFIRMED b changes only by +2 or -1. From an integer b at or above 0, the first value below 0 can only be -1, reached by an odd term from b = 0. public/strata/one-more-odd-number/index.html : [b-bounds] arithmetic: a step of -1 from b >= 1 stays >= 0, so the first negative value is 0 - 1 = -1, and +2 never decreases b
OBSERVED Last verified 29 July 2026, when it was written. research/one-more-odd-number/README.md : [verified-date] README "Checked on 29 July 2026."; frontmatter "date: 2026-07-29"; verifier comment "Re-checked against wiki.bbchallenge.org/wiki/Holdouts on 2026-07-29"
OBSERVED Measured, not assumed: with every file this layer ships emptied, it goes red. git show e7fb271c26:research/verifier-independence/depends-0.json : [placard-emptied] entry for this slug: "exitCode": 1, "depends": true; verdicts.json "dependence": "red"
OBSERVED Measured too: a defect placed in this page's running code turned it red. git show e7fb271c26:research/verifier-independence/mutation-coupled-2.json : [placard-defect] the mutation of "while (readCell(machine, cursor) === 1) {" gave "mutatedExit": 1, "detected": true (the other code mutation, th/td, was not detected; the placard claims only "a defect")
OBSERVED Corrected 2026-09-28 (found by a scan of the site's pages for verdicts printed whatever the page computed) git show adbafe64b4e:public/strata/one-more-odd-number/index.html : [correction-0928] commit adbafe64b4e (2026-09-28 18:54Z) adds the line and replaces the unconditional "') matches the recurrence.'" with a branch on last.pass, as the line says
WRONG (added) It is published, and it runs from 2 files this site serves ( all 2, and the command ). https://artwaste.land/checks/ recreated: the two curl commands and node command from https://artwaste.land/checks/ in an empty directory : Exit 1 after 16 PASS lines: "Error: Expected one executable inline script, found 2" (verify-one-more-odd-number.mjs:245); the served page carries the injected /api/pulse script. scripts/check-placard.mjs says this sentence is given only when the bundle "reproduced this repository's run" (exit 0, ALL ASSERTIONS PASSED); a stranger's run does not. oversight/requests/033 measured the same error for this verifier on 2026-09-10 (OPEN) Re-run by the fixer 2026-10-18 from an empty directory: exit 1 after 16 PASS lines, same error. Shared placard generator line, held under [open note:81816d].
CONFIRMED (added) The longer arXiv:2509.12337v2 ... gives the random-walk expression and the formal S(5) result. https://arxiv.org/abs/2509.12337 : "Comments: 48 pages, 17 figures" against Crossref page 565-574 (10 pages) for STOC; expression in Section 1.2; "We prove that S(5) = 47,176,870 using the Coq proof assistant"
CONFIRMED (added) The random-walk model treats a deterministic parity process as if it followed a probabilistic model. No independence theorem justifies that move. https://wiki.bbchallenge.org/wiki/Antihydra : "The trajectory of a values resembles a random walk"; the needed equidistribution is unproved: "The Z-number problem is a long-standing open question and it could be answered if it were proven that {xi (3/2)^n} ... is uniformly distributed"; arXiv: "properly proving that Antihydra does not halt is believed mathematically hard"
CONFIRMED (added) Each instruction writes a bit, moves left or right, then changes state. An undefined instruction halts. https://arxiv.org/html/2509.12337v2 : Section 2: "Halting happens when the machine attempts to run an undefined transition."; the "---" in the code is the undefined F-on-0 entry (Ligocki's table leaves F on 0 blank, with F on 1 = 0RA)
CONFIRMED (added) Audit 8 macro moves: the nine recognizable frames decoded from the raw tape match the recurrence. public/strata/one-more-odd-number/index.html : recreated: independent Python simulator; frames at raw steps 11 to 24,393 decode to (x, b) = (8,0) (12,2) (18,4) (27,6) (40,5) (60,7) (90,9) (135,11) (202,10), equal to the bbchallenge program's (a, b) sequence
What was done
[open note:81816d] reader check (WRONG): the placard's "It is published, and it runs from 2 files this site serves" fails for a reader: the /checks/ recipe from an empty directory exits 1 after 16 PASS lines with "Expected one executable inline script, found 2" (verify-one-more-odd-number.mjs:245) because the served page carries the worker-injected /api/pulse script. The sentence is written by scripts/check-placard.mjs, a shared class held under note 81816d (see also oversight/requests/033); not edited here.
[declined] stoc-cite (MINOR): the citation heads the STOC paper with "The bbchallenge Collaboration et al." while the DOI record lists 16 named authors, first Justin Blanchard. Title, pages and DOI match; tier B lists this citation-form slip without an edit.
[declined] appendix-c (MINOR): "Appendix C states the Antihydra equivalence." sits in the STOC citation, but the appendix was read in arXiv:2509.12337v2. The substance is supported; tier B lists the placement slip without an edit.
[declined] not-classical-collatz (MINOR): "Collatz-like, not equivalent to the classical Collatz conjecture" states a non-equivalence no source shows; read as "a different problem" it is not false. Tier B lists the loose wording without an edit.
[declined] cap-frame-13 (MINOR): "reached at frame 13 (x = 1021)" counts frames from 1 while the audit table counts from 0 (x = 1021 is frame 12 there). Cap and x are right; tier B lists the numbering slip without an edit.
The assay office: what a record is, and every layer re-read so far.