---
slug: the-star-that-wont-stay-north
passed: 2026-10-14
by: the assay line (research/assay-line/queue-1014c/), Codex fixer directed by claude-assaying-boyle-1014; source record by Codex, checked by Codex
checked: 100
confirmed: 52
wrong: 46
unverifiable: 0
observed: 0
stale: 2
---

This was the directed repair pass over all 49 triage items for this layer, read at bd78d78a14, with evidence re-read before editing and the checker's 51 CONFIRMED claims retained with their evidential scope. The 46 wrong-count entries comprise 30 WRONG and 16 MINOR claims; X32 is rejected as a finding and recorded CONFIRMED. The date records this assay, not an unconditional clean bill: the film source was corrected but the MP4 was not re-rendered, and the generated placard date still awaits the prohibited shared landing rebuild. Historical logs and other layers were left untouched. The corresponding claims remain WRONG, and the counts are unchanged.

## Claims
- WRONG | a great circle | https://science.nasa.gov/learn/basics-of-space-flight/chapter2-1/ | X1: NASA defines a great circle through the sphere centre and gives this pole path radius about 23.4°. Changed the authored HTML, Markdown and film source to small circle. The MP4 was not re-rendered.
- WRONG | the Egyptians built shafts in the Great Pyramid to sight it. | https://earthsky.org/brightest-stars/thuban-past-north-star/ | X2: EarthSky describes proposed stellar alignments, uncertain shaft purpose and a crooked passage unsuitable for sighting. Replaced the asserted sighting purpose in HTML and Markdown.
- WRONG | over a quarter of a hundred centuries. | https://science.nasa.gov/learn/basics-of-space-flight/chapter2-1/ | X3: NASA gives about 26,000 years; a quarter of 100 centuries is 2,500. Replaced the interval with about 26,000 years.
- MINOR | the constellation behind the Sun on your birthday is no longer the one the horoscope names. | git show f5e5f84f34:research/precession-equinoxes/verify.mjs | X4: Independent enumeration of the model gives 321 mismatches and 44 matches among 365 dates. Qualified the text to most birthdays; no page correction line for this MINOR item.
- WRONG | The closest it gets is exactly |66.6° − β|. | git show f5e5f84f34:research/precession-equinoxes/verify.mjs; research/precession-equinoxes/assay-2026-10-14.json | X5: The radial-gap identity is exact for a stationary star and fixed obliquity, not a rounded 66.6° constant applied eternally to moving stars. Scoped the formula in HTML, Markdown, model/verifier documentation and film source. MP4 unchanged.
- WRONG | How close a star ever comes to being the pole star is fixed by one number | git show f5e5f84f34:research/precession-equinoxes/verify.mjs; research/precession-equinoxes/assay-2026-10-14.json | X6: Vega latitude changes from 61.732865° at J2000 to 62.559450° at its searched approach. Present latitude does not fix an eternal minimum. Corrected with X5; MP4 unchanged.
- WRONG | The modern rate reproduces his 2° to the second decimal. | git show f5e5f84f34:research/precession-equinoxes/verify.mjs; research/precession-equinoxes/assay-2026-10-14.json | X7: The computed 150-year drift is 2.095297°, displayed as 2.10°, not 2.00°. Replaced precision agreement with a rough modern-rate illustration, including film source. MP4 unchanged.
- MINOR | the field spans ~33°. | git show f5e5f84f34:public/strata/the-star-that-wont-stay-north/index.html | X8: FIELD=33 is radial distance from the ecliptic pole to the map rim. Caption now explicitly calls 33° the radius; no correction line.
- WRONG | No star but Polaris happens to fall almost dead on the circle, which is luck: our age has an unusually good North Star, accurate to under half a degree. | git show f5e5f84f34:research/precession-equinoxes/verify.mjs; git show f5e5f84f34:public/strata/the-star-that-wont-stay-north/model.js; research/precession-equinoxes/assay-2026-10-14.json | X9: Recomputed Polaris at 0.628727° in 2026, its searched minimum 0.460439° near 2102, and Thuban at 0.326917°. Corrected the current distance and uniqueness in HTML, Markdown and film source. MP4 unchanged.
- WRONG | for a few thousand years around 11,000 CE there is effectively no bright pole star at all. | https://simbad.cds.unistra.fr/simbad/sim-id?Ident=del+Cyg&output.format=ASCII; https://raw.githubusercontent.com/Starlink/sofa/master/src/ltpequ.c; research/precession-equinoxes/assay-2026-10-14.json | X10: Re-read SIMBAD Fawaris V=2.87 and astrometry, then ran unchanged SOFA ltpequ.c with the page stellar motion: 3.221078° near 11546 CE. Named the omitted star and scoped the plotted sample to seven stars.
- MINOR | Every row is computed in your browser by the same model the offline verifier checks, then matched against the published epochs | git show f5e5f84f34:research/precession-equinoxes/verify.mjs | X11: All seven table rows are computed, but ANCHORS has only six entries and skips Kochab. The explanatory text now says so; no correction line.
- WRONG | for deep-time stars they run a little tighter than the full long-term integrations | https://raw.githubusercontent.com/Starlink/sofa/master/src/ltpequ.c; research/precession-equinoxes/assay-2026-10-14.json | X12: Independent comparison gives Alderamin 1.921892° versus 2.199976° and Vega 5.691239° versus 4.001546°. Replaced uniform tighter distances with both results and the shared stellar-motion assumption. New assay results are not attributed to the original author.
- WRONG | A star sits at its own fixed distance from the centre | git show f5e5f84f34:public/strata/the-star-that-wont-stay-north/model.js; research/precession-equinoxes/assay-2026-10-14.json | X13: starVecAt changes RA and Dec with proper motion, so distance from the ecliptic pole changes. Replaced the fixed-star assertion with an explicit stationary-star construction; film source scoped too, MP4 unchanged.
- WRONG | the closest the wandering pole can ever come to that star | git show f5e5f84f34:research/precession-equinoxes/verify.mjs; research/precession-equinoxes/assay-2026-10-14.json | X14: The verifier searches first, then evaluates latitude at that epoch. The prose and verifier now call this a moving-star approximation within the finite model window, not an eternal identity; MP4 unchanged.
- WRONG | Vega (β = 61.7°) can never do better than ~4.9°. | git show f5e5f84f34:research/precession-equinoxes/verify.mjs; research/precession-equinoxes/assay-2026-10-14.json | X15: The page model searches Vega to 4.001546° near 13679, contradicting a permanent 4.9° limit. Replaced that limit with the finite first-order model result.
- WRONG | is stuck at ~6.7° forever. | https://raw.githubusercontent.com/Starlink/sofa/master/src/ltpequ.c; research/precession-equinoxes/assay-2026-10-14.json | X16: The page model gives Deneb 6.653336° near 10208; SOFA pole plus the same stellar motion gives 7.456701° near 10202. Removed the timeless bound in favour of the finite model result.
- WRONG | crossed all of Pisces | https://earthsky.org/human-world/when-will-the-age-of-aquarius-begin/ | X17: EarthSky attributes Pisces entry at 68 BCE and Aquarius entry at 2597 CE to Meeus using IAU boundaries. Changed all of Pisces to most and explicitly said the equinox remains in Pisces.
- MINOR | the Sun on your birthday is almost never in front of the sign you were assigned. | git show f5e5f84f34:research/precession-equinoxes/verify.mjs | X18: The same independent birthday count as X4 gives 44 matching dates. Replaced almost never with most birthdays in HTML and Markdown; no correction line.
- MINOR | then a fortnight in Ophiuchus | https://earthsky.org/astronomy-essentials/dates-of-suns-entry-into-each-constellation/; https://earthsky.org/astronomy-essentials/dates-of-suns-entry-into-each-constellation-of-the-zodiac/ | X19: The model stores 18 days and the re-read EarthSky 2018 table gives November 30 to December 18. Replaced fortnight with about 18 days; no correction line.
- WRONG | one full constellation. | https://www.tumblr.com/nasa/150688852794/zodiac; research/precession-equinoxes/assay-2026-10-14.json | X20: NASA distinguishes unequal constellations from equal signs. The model gives 30.032594° in 2150 years. Replaced one full constellation with about 30°, one tropical-sign width.
- WRONG | taken during lunar eclipses | https://mathshistory.st-andrews.ac.uk/DSB/Hipparchus.pdf | X21: Toomer, PDF page 8, distinguishes Timocharis lunar occultations from Hipparchus measurements at two lunar eclipses over about 160 years. Corrected the narrative and retained 150 only as an explicitly approximate modern-rate illustration; film source corrected, MP4 unchanged.
- MINOR | exactly opposite the Sun | https://science.nasa.gov/moon/eclipses/ | X22: NASA explains partial eclipses with imperfect Sun-Earth-Moon alignment. Changed exactly opposite to approximately opposite; no correction line.
- MINOR | Every star he could check had moved the same way. | https://www.originalsources.com/Document.aspx?DocID=56JSR85842WZNLW | X23: The indexed Heath translation says practically all the other fixed stars he compared. Restored that qualification in the surrounding prose; no correction line.
- WRONG | not less than 1° per century | https://www.originalsources.com/Document.aspx?DocID=56JSR85842WZNLW | X24: The indexed Ptolemy/Heath text gives the 300-years/3° quotation, not the short rate paraphrase as direct speech. Restored that quotation with attribution in HTML, Markdown and film source. MP4 unchanged.
- WRONG | their distances to a fraction of a degree | https://raw.githubusercontent.com/Starlink/sofa/master/src/ltpequ.c; research/precession-equinoxes/assay-2026-10-14.json | X25: The independent SOFA comparison changes Vega by 1.689693°, so the selected anchors do not demonstrate uniform subdegree accuracy. Removed that guarantee and distinguished the wheel reference year 2026 from the verifier estimate reference year 2000.
- MINOR | between 22.0° and 24.5° | https://science.nasa.gov/science-research/earth-science/milankovitch-orbital-cycles-and-their-role-in-earths-climate/ | X26: NASA gives the recent obliquity range as 22.1° to 24.5° and an approximately 41000-year cycle. Corrected the lower endpoint and period scope in HTML and SOURCES.md; no correction line.
- WRONG | obliquity 23.4392911° | https://raw.githubusercontent.com/Starlink/sofa/master/src/obl06.c | X27: SOFA obl80.c uses 84381.448 arcseconds at J2000; obl06.c uses 84381.406. Retained 23.4392911° and corrected its attribution to IAU 1980 in HTML, model comments, SOURCES.md and verifier. Tightened the verifier assertion to test the actual 1980 constant.
- WRONG | obliquity 23.4392911° | https://raw.githubusercontent.com/Starlink/sofa/master/src/obl06.c | X28: Same obliquity attribution as X27. Retained the independently supported IAU 2006 general-precession rate citation and distinguished it from the IAU 1980 obliquity.
- WRONG | the equinox's Pisces/Aquarius entries from the literature (arXiv:1501.05534). | https://arxiv.org/pdf/1501.05534 | X29: Re-read the four-page arXiv paper on the ecliptic and reference systems; it gives neither Pisces/Aquarius entry date. EarthSky reports Meeus 68 BCE and 2597 CE dates. Corrected the HTML apparatus and SOURCES.md.
- MINOR | so they cannot drift apart. | git show f5e5f84f34:research/precession-equinoxes/verify.mjs | X30: The verifier imports model.js and does not read HTML. Scoped the promise to shared functions and selected outputs; no correction line.
- MINOR | Every claim has a check; every check is shown. | git show f5e5f84f34:research/precession-equinoxes/verify.mjs | X31: Selected numeric tests and six external anchors do not check every claim. Replaced the universal footer promise with a statement that model checks and sources are listed; no correction line.
- CONFIRMED | If a number here is wrong, verify.mjs is where to prove it. | git show f5e5f84f34:research/precession-equinoxes/verify.mjs | X32: REJECTED: the sentence invites inspection of model results, not automatic failure after any HTML mutation. closestApproach prints Vega 4.00°, which can expose the old 4.9° sentence. Kept the invitation unchanged; broader coverage claims are separately corrected.
- STALE | Last verified 20 June 2026, when it was written. | git show f5e5f84f34:research/self-check/verifier-runs.json | X33: STALE: the committed verifier-runs.json records a successful 27/27 run on 2026-08-01, after the placard date. The new assay record dates this pass. NOT FIXED in generated HTML: the placard is owned by the shared generator; the permitted layer work cannot run the required shared landing rebuild. No correction/update line, per the triage exception.
- MINOR | so the film cannot drift from the check. | git show f5e5f84f34:research/precession-equinoxes/film/film.html | X34: The original film source contains separately authored captions, including a present-era Polaris claim contradicted by model.js. Narrowed the surrounding caption to model-driven positions and numbers; no correction line for this MINOR guarantee.
- MINOR | so the sign on your horoscope no longer matches the stars behind the Sun. | git show f5e5f84f34:research/precession-equinoxes/verify.mjs | X35: The independent model enumeration leaves 44 matching dates. Qualified the Markdown dek and related-note birthday wording to most birthdays; no correction line for this MINOR item.
- WRONG | Vega next | https://earthsky.org/brightest-stars/thuban-past-north-star/ | X36: The model succession places Errai near 4129 and Vega near 13679. Changed Vega next to Vega later in description metadata and Article JSON-LD.
- WRONG | every zodiac sign a whole constellation off its stars. | https://www.tumblr.com/nasa/150688852794/zodiac | X37: NASA unequal constellation widths plus the 44 model matches contradict a uniform offset for every sign. Replaced the metadata and JSON-LD claim with about 30° per 2150 years and most-birthday mismatch.
- MINOR | every figure offline-verified. | git show f5e5f84f34:research/precession-equinoxes/verify.mjs | X38: The verifier does not read the static metadata. Replaced every figure offline-verified with selected model outputs checked offline; no correction line.
- WRONG | slid your zodiac sign one constellation off its stars. | git show f5e5f84f34:research/precession-equinoxes/verify.mjs | X39: The Open Graph one-constellation offset has the same false uniform unit as X20. Replaced it with approximate angular drift and qualified birthday mismatches.
- WRONG | slid your zodiac sign a whole constellation off its stars. | git show f5e5f84f34:research/precession-equinoxes/verify.mjs | X40: The page frame used the same false uniform constellation unit. Replaced it with about 30° per 2150 years and most-birthday mismatch.
- WRONG | cut sighting shafts to it ~2800 BCE | https://earthsky.org/brightest-stars/thuban-past-north-star/ | X41: EarthSky dates the Great Pyramid to the 26th century BCE; about 2800 BCE is the separate Thuban near-pole epoch. Corrected both the date attachment and sighting-purpose claim in Markdown and HTML.
- WRONG | the closest any star ever comes to being the pole star | git show f5e5f84f34:research/precession-equinoxes/verify.mjs | X42: The Markdown repeated the timeless latitude claim. Corrected with the stationary-star and fixed-obliquity assumptions and the epoch-specific moving-star approximation; the same film-source formula is scoped, MP4 unchanged.
- MINOR | the Sun spends a fortnight | https://earthsky.org/astronomy-essentials/dates-of-suns-entry-into-each-constellation/; https://earthsky.org/astronomy-essentials/dates-of-suns-entry-into-each-constellation-of-the-zodiac/ | X43: Same 18-day model and November 30 to December 18 EarthSky interval as X19. Corrected Markdown fortnight to about 18 days; no correction line.
- MINOR | Everything the page can display is recomputed in the browser | git show f5e5f84f34:research/precession-equinoxes/verify.mjs | X44: Static prose and captions are not recomputed by the shared model. Narrowed Markdown to the dynamic results and named separate authored text; no correction line.
- MINOR | Year 2,025 CE · pole star: Polaris · 0.6° from the true pole · (today) | git show f5e5f84f34:public/strata/the-star-that-wont-stay-north/index.html | X45: The old sky range min=-13000 and step=25 forced 2026 to 2025; the label accepted a 30-year interval. Set the sky step to 1 and today condition to equality with 2026. Also set the zodiac step to 1 so its existing Now action can select 2026; no correction line.
- STALE | "dateModified":"2026-06-20" | git log f5e5f84f34 -- public/strata/the-star-that-wont-stay-north/index.html | X46: STALE: repository history has revisions after June 20. Set Article dateModified to the requested assay date 2026-10-14 and preserved datePublished. No Updated line, per the explicit exception.
- WRONG | "uploadDate":"2026-07-09" | git show f5e5f84f34:memory/log.d/2026-06-20T0240Z-precession-film.md; git log f5e5f84f34 -- public/strata/the-star-that-wont-stay-north/film.mp4 | X47: Read-only git history shows d0d15c2f4be added the MP4 on 2026-06-20, and the June 20 shipping log records it shipped. Corrected VideoObject.uploadDate to 2026-06-20.
- WRONG | (added) the same leftover the calendar spends centuries chasing. | https://eaae-astronomy.org/workshops/hipparchus-and-the-precession-of-the-equinoxes | X48: EAAE year lengths differ by 20m24s; 365.25 days exceeds the cited tropical year by 11m14s. Replaced the related note equating these residuals with the separate calendar task of fitting the tropical year into whole days.
- WRONG | (added) Polaris now (under half a degree) | git show f5e5f84f34:public/strata/the-star-that-wont-stay-north/model.js; research/precession-equinoxes/assay-2026-10-14.json | X49: Corrected the film caption SOURCE to Polaris in 2026, about 0.63° in this model, using poleStarDist at years=26. NOT FIXED in the shipped MP4: rendering/re-encoding was explicitly forbidden. The page flags the unrevised film and the dated Polaris correction explains the numerical error.
- CONFIRMED | Sailors trust the North Star to hold still while everything else wheels around it. | https://science.nasa.gov/learn/basics-of-space-flight/chapter2-1/ | navigation: NASA describes estimating latitude from Polaris altitude above the horizon.
- CONFIRMED | But the Earth's axis is not fixed; | https://science.nasa.gov/learn/basics-of-space-flight/chapter2-1/ | axis-not-fixed: NASA describes precession of the axis with about a 26000-year circuit.
- CONFIRMED | The spin axis points at whatever star happens to lie under it. | https://science.nasa.gov/learn/basics-of-space-flight/chapter2-1/ | pole-definition: The named pole star is near the direction of the spin axis, not necessarily exactly on it.
- CONFIRMED | Today it is Polaris. | https://science.nasa.gov/learn/basics-of-space-flight/chapter2-1/ | polaris-today: NASA identifies Polaris near the present north celestial pole.
- CONFIRMED | In twelve thousand years it will be Vega, the summer beacon | https://www.star-facts.com/vega/ | vega-future: The star source gives the next Vega era near 13700 CE; approximate epoch only.
- CONFIRMED | The pole's path is a circle of radius 23.4° (the axial tilt) about a fixed point: the pole of Earth's orbit. | https://science.nasa.gov/learn/basics-of-space-flight/chapter2-1/ | fixed-pole-circle: Confirmed as the stated first-order fixed-ecliptic model, not an exact permanent Earth trajectory.
- CONFIRMED | Hipparchus discovered precession around 129 BCE from a roughly 2° Spica comparison over about 160 years. | https://mathshistory.st-andrews.ac.uk/DSB/Hipparchus.pdf | hipparchus-discovery: Toomer supports a late-career discovery around this date and a rough 2° comparison. The corrected text uses about 160 years; 150 remains only an illustrative calculation.
- CONFIRMED | Star positions are ICRS J2000 (SIMBAD), carried by their own proper motion | git show f5e5f84f34:public/strata/the-star-that-wont-stay-north/model.js; https://simbad.cds.unistra.fr/simbad/sim-id?Ident=Vega&output.format=ASCII | astrometry: Re-read all seven SIMBAD coordinate and proper-motion fields and compared them with STARS; they agree at the entered precision.
- CONFIRMED | which is why Vega, the fastest mover here, visibly shifts across the millennia. | https://simbad.cds.unistra.fr/simbad/sim-id?Ident=Vega&output.format=ASCII | vega-pm: Hypotenuse of Vega motions is 349.721 mas/yr, versus the next largest Errai at 183.036 in this seven-star list.
- CONFIRMED | Errai (γ Cephei, ~4000 CE) | https://earthsky.org/brightest-stars/thuban-past-north-star/ | errai-epoch: EarthSky places Errai near 4000 CE; approximate epoch.
- CONFIRMED | Alderamin (α Cephei, ~7500 CE) | https://earthsky.org/brightest-stars/thuban-past-north-star/ | alderamin-epoch: EarthSky places Alderamin near 7500 CE; approximate epoch.
- CONFIRMED | Deneb has a roughly 7° approach around 10000 CE in the page model. | https://www.star-facts.com/deneb/ | deneb-epoch: The star source describes a roughly 7° future approach near 9800 CE; this supports only that approximate future passage, not never. The corrected page scopes its finite model result.
- CONFIRMED | and arrives at brilliant Vega around 13,700 CE before the long return to Thuban. | https://www.star-facts.com/vega/ | vega-epoch: The star source and page model put that future approach near 13700 CE; no exact astrometric guarantee.
- CONFIRMED | Thuban ~2787 BCE | https://earthsky.org/brightest-stars/thuban-past-north-star/ | thuban-canon: EarthSky reports the approximate near-pole epoch 2787 BCE; the page model separately gives -2818.
- CONFIRMED | Polaris ~2100 CE | https://earthsky.org/brightest-stars/thuban-past-north-star/ | polaris-canon: EarthSky attributes an approximately 2100 CE approach of 0.4525° to Meeus.
- CONFIRMED | Vondrák et al. 2011 | https://www.archaeoastronomy.it/Archaeoastronomy%20Program/Vondrak/Vondrak.pdf | vondrak-citation: The paper is Vondrak, Capitaine and Wallace, A&A 534 A22 (2011), describing long-term precession expressions.
- CONFIRMED | exactly 90° − β, where β is its ecliptic latitude. | git show f5e5f84f34:public/strata/the-star-that-wont-stay-north/model.js | latitude-distance: The model defines latitude as 90° minus angular distance from the ecliptic pole at the specified epoch.
- CONFIRMED | A star is a good pole star only if its ecliptic latitude is near 66.6°. | git show f5e5f84f34:public/strata/the-star-that-wont-stay-north/model.js | good-latitude: Confirmed for proximity to the fixed-tilt circle, with latitude evaluated at the relevant epoch.
- CONFIRMED | the point where the Sun crosses the equator in March, the anchor of the whole calendar of signs | https://eaae-astronomy.org/workshops/hipparchus-and-the-precession-of-the-equinoxes | equinox-definition: EAAE defines the equinox at the ecliptic/equator intersection; tropical Aries is fixed to the seasonal point.
- CONFIRMED | It creeps backward along the zodiac at about 1° every 71.6 years, a full lap in the same 25,772 years. | https://raw.githubusercontent.com/Starlink/sofa/master/src/p06e.c | equinox-rate: SOFA p06e.c gives J2000 rate 5028.796195 arcseconds per century, about 50.29 per year; 71.6 years per degree is approximate.
- CONFIRMED | When Greek and Babylonian astronomers fixed the twelve signs, a little over two thousand years ago, the March equinox sat at the head of the constellation Aries, and sign and constellation agreed. | https://www.tumblr.com/nasa/150688852794/zodiac | sign-history: Confirmed only as rough ancient association of sign names and constellations. NASA and EAAE distinguish equal signs from unequal constellations; this is not exact boundary coincidence or a precisely dated joint Greek/Babylonian event.
- CONFIRMED | which is why people speak of the dawning of the Age of Aquarius, still some centuries off (≈ 2597 CE). | https://earthsky.org/human-world/when-will-the-age-of-aquarius-begin/ | aquarius-date: EarthSky reports Meeus IAU-boundary entry at 2597 CE, not a universal astrological definition.
- CONFIRMED | Inner ring: the thirteen IAU constellations the ecliptic truly crosses | https://www.tumblr.com/nasa/150688852794/zodiac | inner-ring: NASA describes 13 constellations on the annual solar path.
- CONFIRMED | unequal in width, including Ophiuchus, which the twelve-sign zodiac drops entirely. | https://www.tumblr.com/nasa/150688852794/zodiac | ophiuchus-omitted: NASA distinguishes the unequal constellations and Ophiuchus from the twelve-sign scheme.
- CONFIRMED | about 45 days in Virgo but only 7 in Scorpius | https://www.tumblr.com/nasa/150688852794/zodiac | virgo-scorpius: NASA gives about 45 days in Virgo and 7 in Scorpius.
- CONFIRMED | None of this needed a telescope. | https://www.originalsources.com/Document.aspx?DocID=56JSR85842WZNLW | no-telescope: Ptolemy/Heath and Toomer describe the ancient observations; this predates telescopic astronomy.
- CONFIRMED | Spica had shifted by roughly 2° relative to the equinox. | https://www.originalsources.com/Document.aspx?DocID=56JSR85842WZNLW | spica-shift: Ptolemy/Heath reports distances about 8° and 6° from the autumnal equinox; their difference is about 2°.
- CONFIRMED | a full cycle in at most 36,000 years. | https://www.originalsources.com/Document.aspx?DocID=56JSR85842WZNLW | hipp-max: A rate floor of 1° per century implies at most 36000 years for 360°, if taken as a uniform rate.
- CONFIRMED | A naked-eye astronomer in the second century BCE used a roughly 160-year baseline. | https://www.originalsources.com/Document.aspx?DocID=56JSR85842WZNLW | hipp-naked-eye: Toomer supports an ancient naked-eye comparison across about 160 years; the corrected text uses that approximate interval.
- CONFIRMED | the plane of Earth's orbit drifts as well (planetary precession). | https://www.archaeoastronomy.it/Archaeoastronomy%20Program/Vondrak/Vondrak.pdf | planetary-precession: The Vondrak paper introduction identifies precession of the ecliptic due to planetary perturbations.
- CONFIRMED | Both matter only in deep time, where they shift the figures by of order a degree | https://raw.githubusercontent.com/Starlink/sofa/master/src/ltpequ.c | deep-time-only: At this page degree scale, omitted terms matter over millennia: the independent Vega comparison differs by 1.69°. This is not a zero-error claim at shorter intervals.
- CONFIRMED | SIMBAD ICRS J2000 positions and proper motions for all seven stars | https://simbad.cds.unistra.fr/simbad/sim-id?Ident=Vega&output.format=ASCII | simbad-citation: All seven SIMBAD positions and proper motions were re-read and agree with the entered STARS precision.
- CONFIRMED | IAU-boundary constellation transits (EarthSky) | https://earthsky.org/astronomy-essentials/dates-of-suns-entry-into-each-constellation/ | earthsky-citation: Recovered the accessible EarthSky 2015 and 2018 tables; they use IAU boundaries. The 2015 dates match the model ranges; yearly dates can vary.
- CONFIRMED | Earth's axis leans 23.4° and swings that lean around a full circle once every 25,772 years | https://science.nasa.gov/learn/basics-of-space-flight/chapter2-1/ | dek-axis: NASA supports the rounded tilt and cycle; 25772 is the chosen first-order model period.
- CONFIRMED | Thuban for the pyramid-builders, Polaris now, Vega in twelve thousand years | https://earthsky.org/brightest-stars/thuban-past-north-star/ | dek-succession: EarthSky places Thuban in the pyramid era and Polaris today; the star source gives Vega near 13700 CE.
- CONFIRMED | Polaris is temporary. | https://science.nasa.gov/learn/basics-of-space-flight/chapter2-1/ | og-pole: NASA precession discussion makes the present pole-star alignment temporary.
- CONFIRMED | Polaris today (and unusually close | https://www.star-facts.com/polaris/ | md-now: Polaris is currently the close pole star; the corrected distance is separately modelled at 0.628727° in 2026. The source 0.66° value is dated 2018.
- CONFIRMED | Vega around 13,700 CE | https://www.star-facts.com/vega/ | md-vega: Star Facts gives Vega near 13700 CE; only an approximate future epoch is confirmed.
- CONFIRMED | Two skies built from the same two facts | git show f5e5f84f34:src/content/strata/the-suns-crooked-clock.md | related-two-skies: Read the pinned sibling Markdown: axial tilt and orbital eccentricity underlie its equation-of-time/analemma discussion. This layer uses the tilt for precession geometry.
- CONFIRMED | There the tilt and orbit bend the sundial's day into a figure-eight; here the tilt, swung slowly around, walks the pole from star to star. | git show f5e5f84f34:src/content/strata/the-suns-crooked-clock.md | related-analemma: The pinned sibling explains the tilt/orbit contribution to the figure-eight analemma; this model rotates the tilted pole.
- CONFIRMED | Precession is why there are two lengths of year. | https://eaae-astronomy.org/workshops/hipparchus-and-the-precession-of-the-equinoxes | two-years: EAAE distinguishes the equinox-to-equinox tropical year from the star-relative sidereal year.
- CONFIRMED | The tropical year (equinox to equinox) is ~20 minutes shorter than the sidereal year (star to star) precisely because the equinox slides backward | https://eaae-astronomy.org/workshops/hipparchus-and-the-precession-of-the-equinoxes | year-difference: Subtracting the EAAE sidereal and tropical values gives 20m24s; the moving equinox shortens the tropical year.
- CONFIRMED | SIMBAD ICRS J2000 positions and proper motions for all seven stars | https://simbad.cds.unistra.fr/simbad/sim-id?Ident=Thuban&output.format=ASCII | simbad-thuban: Re-read this SIMBAD ICRS coordinate and proper-motion record; its fields match the corresponding rounded STARS entry.
- CONFIRMED | SIMBAD ICRS J2000 positions and proper motions for all seven stars | https://simbad.cds.unistra.fr/simbad/sim-id?Ident=Kochab&output.format=ASCII | simbad-kochab: Re-read this SIMBAD ICRS coordinate and proper-motion record; its fields match the corresponding rounded STARS entry.
- CONFIRMED | SIMBAD ICRS J2000 positions and proper motions for all seven stars | https://simbad.cds.unistra.fr/simbad/sim-id?Ident=Polaris&output.format=ASCII | simbad-polaris: Re-read this SIMBAD ICRS coordinate and proper-motion record; its fields match the corresponding rounded STARS entry.
- CONFIRMED | SIMBAD ICRS J2000 positions and proper motions for all seven stars | https://simbad.cds.unistra.fr/simbad/sim-id?Ident=Errai&output.format=ASCII | simbad-errai: Re-read this SIMBAD ICRS coordinate and proper-motion record; its fields match the corresponding rounded STARS entry.
- CONFIRMED | SIMBAD ICRS J2000 positions and proper motions for all seven stars | https://simbad.cds.unistra.fr/simbad/sim-id?Ident=Alderamin&output.format=ASCII | simbad-alderamin: Re-read this SIMBAD ICRS coordinate and proper-motion record; its fields match the corresponding rounded STARS entry.
- CONFIRMED | SIMBAD ICRS J2000 positions and proper motions for all seven stars | https://simbad.cds.unistra.fr/simbad/sim-id?Ident=Deneb&output.format=ASCII | simbad-deneb: Re-read this SIMBAD ICRS coordinate and proper-motion record; its fields match the corresponding rounded STARS entry.
- CONFIRMED | Hipparchus discovered the whole thing ~129 BCE | https://mathshistory.st-andrews.ac.uk/DSB/Hipparchus.pdf | hipparchus-discovery-date: Toomer gives observations through 128 BCE and places precession late in the career; around 129 BCE is approximate.
- CONFIRMED | The first recorded detection of precession | https://mathshistory.st-andrews.ac.uk/DSB/Hipparchus.pdf | film-first-detection: Toomer and the translated Ptolemy account credit Hipparchus with the discovery; no exact day is established.
- CONFIRMED | (added) luni-solar precession | https://science.nasa.gov/learn/basics-of-space-flight/chapter2-1/ | added-lunisolar: NASA describes lunar and solar torques on Earth oblateness, supporting the term.

## Findings

- [open note:98f7d7] X1 NASA defines a great circle through the sphere centre and gives this pole path radius about 23.4°. Changed the authored HTML, Markdown and film source to small circle. The MP4 was not re-rendered.
- [fixed] X2 EarthSky describes proposed stellar alignments, uncertain shaft purpose and a crooked passage unsuitable for sighting. Replaced the asserted sighting purpose in HTML and Markdown.
- [fixed] X3 NASA gives about 26,000 years; a quarter of 100 centuries is 2,500. Replaced the interval with about 26,000 years.
- [fixed] X4 Independent enumeration of the model gives 321 mismatches and 44 matches among 365 dates. Qualified the text to most birthdays; no page correction line for this MINOR item.
- [open note:98f7d7] X5 The radial-gap identity is exact for a stationary star and fixed obliquity, not a rounded 66.6° constant applied eternally to moving stars. Scoped the formula in HTML, Markdown, model/verifier documentation and film source. MP4 unchanged.
- [open note:98f7d7] X6 Vega latitude changes from 61.732865° at J2000 to 62.559450° at its searched approach. Present latitude does not fix an eternal minimum. Corrected with X5; MP4 unchanged.
- [open note:98f7d7] X7 The computed 150-year drift is 2.095297°, displayed as 2.10°, not 2.00°. Replaced precision agreement with a rough modern-rate illustration, including film source. MP4 unchanged.
- [fixed] X8 FIELD=33 is radial distance from the ecliptic pole to the map rim. Caption now explicitly calls 33° the radius; no correction line.
- [open note:98f7d7] X9 Recomputed Polaris at 0.628727° in 2026, its searched minimum 0.460439° near 2102, and Thuban at 0.326917°. Corrected the current distance and uniqueness in HTML, Markdown and film source. MP4 unchanged.
- [fixed] X10 Re-read SIMBAD Fawaris V=2.87 and astrometry, then ran unchanged SOFA ltpequ.c with the page stellar motion: 3.221078° near 11546 CE. Named the omitted star and scoped the plotted sample to seven stars.
- [fixed] X11 All seven table rows are computed, but ANCHORS has only six entries and skips Kochab. The explanatory text now says so; no correction line.
- [fixed] X12 Independent comparison gives Alderamin 1.921892° versus 2.199976° and Vega 5.691239° versus 4.001546°. Replaced uniform tighter distances with both results and the shared stellar-motion assumption. New assay results are not attributed to the original author.
- [open note:98f7d7] X13 starVecAt changes RA and Dec with proper motion, so distance from the ecliptic pole changes. Replaced the fixed-star assertion with an explicit stationary-star construction; film source scoped too, MP4 unchanged.
- [open note:98f7d7] X14 The verifier searches first, then evaluates latitude at that epoch. The prose and verifier now call this a moving-star approximation within the finite model window, not an eternal identity; MP4 unchanged.
- [fixed] X15 The page model searches Vega to 4.001546° near 13679, contradicting a permanent 4.9° limit. Replaced that limit with the finite first-order model result.
- [fixed] X16 The page model gives Deneb 6.653336° near 10208; SOFA pole plus the same stellar motion gives 7.456701° near 10202. Removed the timeless bound in favour of the finite model result.
- [fixed] X17 EarthSky attributes Pisces entry at 68 BCE and Aquarius entry at 2597 CE to Meeus using IAU boundaries. Changed all of Pisces to most and explicitly said the equinox remains in Pisces.
- [fixed] X18 The same independent birthday count as X4 gives 44 matching dates. Replaced almost never with most birthdays in HTML and Markdown; no correction line.
- [fixed] X19 The model stores 18 days and the re-read EarthSky 2018 table gives November 30 to December 18. Replaced fortnight with about 18 days; no correction line.
- [fixed] X20 NASA distinguishes unequal constellations from equal signs. The model gives 30.032594° in 2150 years. Replaced one full constellation with about 30°, one tropical-sign width.
- [open note:98f7d7] X21 Toomer, PDF page 8, distinguishes Timocharis lunar occultations from Hipparchus measurements at two lunar eclipses over about 160 years. Corrected the narrative and retained 150 only as an explicitly approximate modern-rate illustration; film source corrected, MP4 unchanged.
- [fixed] X22 NASA explains partial eclipses with imperfect Sun-Earth-Moon alignment. Changed exactly opposite to approximately opposite; no correction line.
- [fixed] X23 The indexed Heath translation says practically all the other fixed stars he compared. Restored that qualification in the surrounding prose; no correction line.
- [open note:98f7d7] X24 The indexed Ptolemy/Heath text gives the 300-years/3° quotation, not the short rate paraphrase as direct speech. Restored that quotation with attribution in HTML, Markdown and film source. MP4 unchanged.
- [fixed] X25 The independent SOFA comparison changes Vega by 1.689693°, so the selected anchors do not demonstrate uniform subdegree accuracy. Removed that guarantee and distinguished the wheel reference year 2026 from the verifier estimate reference year 2000.
- [fixed] X26 NASA gives the recent obliquity range as 22.1° to 24.5° and an approximately 41000-year cycle. Corrected the lower endpoint and period scope in HTML and SOURCES.md; no correction line.
- [fixed] X27 SOFA obl80.c uses 84381.448 arcseconds at J2000; obl06.c uses 84381.406. Retained 23.4392911° and corrected its attribution to IAU 1980 in HTML, model comments, SOURCES.md and verifier. Tightened the verifier assertion to test the actual 1980 constant.
- [fixed] X28 Same obliquity attribution as X27. Retained the independently supported IAU 2006 general-precession rate citation and distinguished it from the IAU 1980 obliquity.
- [fixed] X29 Re-read the four-page arXiv paper on the ecliptic and reference systems; it gives neither Pisces/Aquarius entry date. EarthSky reports Meeus 68 BCE and 2597 CE dates. Corrected the HTML apparatus and SOURCES.md.
- [fixed] X30 The verifier imports model.js and does not read HTML. Scoped the promise to shared functions and selected outputs; no correction line.
- [fixed] X31 Selected numeric tests and six external anchors do not check every claim. Replaced the universal footer promise with a statement that model checks and sources are listed; no correction line.
- [dated] X33 STALE: the committed verifier-runs.json records a successful 27/27 run on 2026-08-01, after the placard date. The new assay record dates this pass. NOT FIXED in generated HTML: the placard is owned by the shared generator; the permitted layer work cannot run the required shared landing rebuild. No correction/update line, per the triage exception.
- [fixed] X34 The original film source contains separately authored captions, including a present-era Polaris claim contradicted by model.js. Narrowed the surrounding caption to model-driven positions and numbers; no correction line for this MINOR guarantee.
- [fixed] X35 The independent model enumeration leaves 44 matching dates. Qualified the Markdown dek and related-note birthday wording to most birthdays; no correction line for this MINOR item.
- [fixed] X36 The model succession places Errai near 4129 and Vega near 13679. Changed Vega next to Vega later in description metadata and Article JSON-LD.
- [fixed] X37 NASA unequal constellation widths plus the 44 model matches contradict a uniform offset for every sign. Replaced the metadata and JSON-LD claim with about 30° per 2150 years and most-birthday mismatch.
- [fixed] X38 The verifier does not read the static metadata. Replaced every figure offline-verified with selected model outputs checked offline; no correction line.
- [fixed] X39 The Open Graph one-constellation offset has the same false uniform unit as X20. Replaced it with approximate angular drift and qualified birthday mismatches.
- [fixed] X40 The page frame used the same false uniform constellation unit. Replaced it with about 30° per 2150 years and most-birthday mismatch.
- [fixed] X41 EarthSky dates the Great Pyramid to the 26th century BCE; about 2800 BCE is the separate Thuban near-pole epoch. Corrected both the date attachment and sighting-purpose claim in Markdown and HTML.
- [open note:98f7d7] X42 The Markdown repeated the timeless latitude claim. Corrected with the stationary-star and fixed-obliquity assumptions and the epoch-specific moving-star approximation; the same film-source formula is scoped, MP4 unchanged.
- [fixed] X43 Same 18-day model and November 30 to December 18 EarthSky interval as X19. Corrected Markdown fortnight to about 18 days; no correction line.
- [fixed] X44 Static prose and captions are not recomputed by the shared model. Narrowed Markdown to the dynamic results and named separate authored text; no correction line.
- [fixed] X45 The old sky range min=-13000 and step=25 forced 2026 to 2025; the label accepted a 30-year interval. Set the sky step to 1 and today condition to equality with 2026. Also set the zodiac step to 1 so its existing Now action can select 2026; no correction line.
- [dated] X46 STALE: repository history has revisions after June 20. Set Article dateModified to the requested assay date 2026-10-14 and preserved datePublished. No Updated line, per the explicit exception.
- [fixed] X47 Read-only git history shows d0d15c2f4be added the MP4 on 2026-06-20, and the June 20 shipping log records it shipped. Corrected VideoObject.uploadDate to 2026-06-20.
- [fixed] X48 EAAE year lengths differ by 20m24s; 365.25 days exceeds the cited tropical year by 11m14s. Replaced the related note equating these residuals with the separate calendar task of fitting the tropical year into whole days.
- [open note:98f7d7] X49 Corrected the film caption SOURCE to Polaris in 2026, about 0.63° in this model, using poleStarDist at years=26. NOT FIXED in the shipped MP4: rendering/re-encoding was explicitly forbidden. The page flags the unrevised film and the dated Polaris correction explains the numerical error.

## Checks and limits

The layer verifier passed 27/27 after the repairs. The independent SOFA comparison and all replacement model figures are retained in research/precession-equinoxes/assay-2026-10-14.json, reproducible with the adjacent Python script and its documented inputs. Seventeen WRONG correction lines are mirrored between the HTML note and Markdown body; MINOR fixes have no page correction lines. The two STALE entries are the triage's no-line exceptions. The source-only film fixes remain incomplete in the shipped MP4, and its poster and all audio/image assets are unchanged. The public verifier copy is regenerated from the authored verifier; shared manifests and placards are left for the landing rebuild.

At the fixer stage, `node scripts/check-claims-pass.mjs` exited 1 only for unrelated coupon-collector and kapitza-pendulum records. The record then passed the exported checkRecord function with zero errors (100 claims). A Node DOM stub exercised the page script, seven table rows and both Now controls; the 2026 label was absent at adjacent years. Film JavaScript was syntax-parsed only. Playwright was unavailable, so no browser/visual validation is claimed.

Repair audit for queue-1014c: all seventeen dated correction lines were checked against the old claims and evidence. The cycle note now states the documented 25,772-year model period instead of inferring an intended interval. The verifier passed 27/27 and the em-dash check passed. The current claims gate rejects this record only because its eleven required open board notes finding lines await real board note ids; its counts reconcile. The full gate also reported unrelated held-at-the-door findings, left untouched. The report and proposed board note are in research/assay-line/queue-1014c/the-star-that-wont-stay-north/repair-report.md.