---
slug: the-star-next-to-the-moon
passed: 2026-11-06
by: claude-tender-pascal-d8fkqa, one independent agent on oversight/claims-pass.md that did not write the page, before publish, layer and tool at commit 5e394280e5; every finding re-read at its source by the author before fixing; four claims the fixes added were checked by the author and are marked so
checked: 80
confirmed: 40
wrong: 12
unverifiable: 0
observed: 28
stale: 0
---
# Claims pass: The Star Next to the Moon, and the tool it describes (Which light is that?)

Run before publish, so step 2 (the live page) and the download half of step 3 could not be done: neither page is
live and nothing is served under artwaste.land/checks/ for it yet. In their place, from the repository:
`node research/bright-light/verify.mjs` exits 0 with "91 passed, 0 failed (86 checks, then 5 controls)", byte-identical
to research/bright-light/out/verify-output.txt. Every number in the layer's "The check" agrees with that run (0.0025,
0.0014, 0.0033, 0.0089, Mars 0.126; 0.074 and 0.025 from maxima 0.0742 and 0.0254, the Moon 0.0169; 0.03 per cent and
0.0003; 0.2811, 0.1761, 0.1719; 24 events within 0.9 minutes; 86 checks then 5 controls). The three tables were redone
by hand: every percentage rounds as printed, and each table-2 row sums to table 1's count (55 N: 1,894 plus Betelgeuse).

`node assay/sources.mjs the-star-next-to-the-moon --online`, first run: 45 entries, 32 in the repository, 4 UNREADABLE
(HTTP 503 from nssdc.gsfc.nasa.gov), 6 FOUND, 3 NOT FOUND. During the pass the author committed 5e394280e5: a 46th
entry (dek-five-degrees), the tool's daytime sentence and the hood's "Positions" paragraph; the layer did not change.
The rerun found the four NASA entries and the new one. The three NOT FOUND are in their documents: mallama-hilton and
mars-terms are in the arXiv PDF, not on the abs page the record names; green-028's equations are split by markup ("A Ray").

Step 6, modified (verify.mjs family G2 reads the layer's tables). In a scratch copy of research/bright-light,
public/tools/sky-now and the layer, the unchanged copy passes 91/0. One number changed in each table in turn (45 N
"960" to 961; 45 N Venus 163 to 164; 45 N "Venus 33%" to 34%) fails G each time: "23 of 24 rows", 90 passed 1 failed,
exit 1. A changed prose number (the dek's "36 to 38" to "46 to 48", or "0.0089 for Saturn" to "0.0019") still passes
91/0: the tables are guarded, the prose is not, so all prose numbers were compared by hand (they agree).

Also run: `node research/the-sky-above-you/verify.mjs` (42 passes / 0 failures; the two ephem.mjs copies identical).
Not run: baserate.mjs. Instead, scratch scripts that import the copied engine re-implement the count; at the page's
own settings they reproduce 45 N exactly (5,719 Moon nights, 2,140 with one of the ten, table 2's row light by light),
which makes the recounts in findings 1 and 11 comparable with the page. Fetched by this pass: CycleCalcs (tool page,
privacy page, sky-id.js, main.js), LiveEarthViewer (identify page, the London city page and its scripts), Stellarium
Web (index, bundle, engine wasm, engine source), azmth, Heavens-Above (with the "<" postback), Astroviewer (ISS and
Tiangong pages), the Spot the Station App Store listing and NASA page, the NASA fact sheets, the arXiv PDF, Crossref,
Green's ICQ page, the Horizons manual, the IAU 2015 report, eCFR, NOIRLab, StarDate, NASA Imagine, Ian Ridpath,
Webvision and the APOD, NASA, RMG pages cited below. VizieR (an outside count of Hipparcos stars) reset the connection
twice. Records 1-46 below follow assay/sources/the-star-next-to-the-moon.json in order; the rest are claims it leaves
out, from the layer and then from the tool (index.html, app.mjs, sky.mjs, magnitude.mjs, mcp.js which_light).

After the pass (the author, 2026-11-06). All twelve findings were fixed before publish; each is answered below
with what changed. The fixes added a count at the tool's own cut (stars brighter than 2.15) to
research/bright-light/baserate.mjs and two families' worth of checks to verify.mjs, which now reads "95 passed, 0
failed (90 checks, then 5 controls)", saved as research/bright-light/out/verify-output.txt; the layer says 90. The four
claims those fixes put on the page are the last four lines under Claims, checked by the author, not by the agent; the
agent's own scratch recount for finding 1 (757 nights at 45 N, 806 at the equator, 811 at 30 S; Nunki 285, Elnath
243, Alhena 235, Algieba 160 at 45 N) is the base-rate file's, night for night, which is the independent part of
their check. The three NOT FOUND entries now carry verbatim extracts in research/bright-light/sources/
excerpts-2026-11-06.txt (the arXiv PDF's title page and section 4.4; Green's equations and his Table Ia, whose first
row gives 0.28 at sea level), so check-sources reads them offline.

## Claims
- MINOR | dek and plain: "among everything of magnitude 1.5 or brighter, there are only ten possible answers" (dek-ten) | research/bright-light/out/verify-output.txt | the excerpt lists eleven lights, Betelgeuse among them; the dek concedes Betelgeuse in its last sentence, the plain never mentions it; findings 1 and 3
- OBSERVED | dek: one of the ten within 10 degrees of the Moon on 36 to 38 per cent of the nights it is up, only stars on more than half of those (dek-share) | recreated: node research/bright-light/verify.mjs, 91 passed, output byte-identical to research/bright-light/out/verify-output.txt | 36.2 to 38.2 per cent; only stars 51.2 to 55.1; by hand 2,069/5,722 = 36.2 up to 2,343/6,128 = 38.2
- OBSERVED | dek: Betelgeuse got in once, from 55 degrees north, by four hundredths of a degree that the Moon's parallax supplied (dek-betelgeuse) | recreated: node research/bright-light/verify.mjs (family H) | Horizons' Moon 9.958 degrees from Betelgeuse, 0.042 inside 10 (the engine's own Moon 9.948); without parallax 10.75; research/bright-light/horizons/raw/edge-2034-04-22-moon-55n.txt kept
- CONFIRMED | dek: the Moon never strays more than about five degrees from the Sun's yearly path (dek-five-degrees, added to the record during this pass) | https://www.astronomy.com/observing/does-the-moon-follow-the-same-path-across-the-sky-as-the-sun/ | "So, the Moon can appear anywhere in a band extending 5.1° north (above) and south (below) of the ecliptic." verbatim; geocentric, as the record's note says
- OBSERVED | plain: one of them within 10 degrees on about a third of the nights the Moon is up, a star at least as often as a planet (plain-third) | recreated: node research/bright-light/verify.mjs (family G) | 36.2 to 38.2 per cent; only stars 51.2 to 55.1 per cent
- CONFIRMED | NASA's fact sheet gives the Moon's inclination to the ecliptic as 5.145 degrees (moon-inclination) | https://nssdc.gsfc.nasa.gov/planetary/factsheet/moonfact.html | "Inclination to ecliptic (deg) 5.145"; the checker's first run got HTTP 503, its rerun and this pass's fetch found it
- OBSERVED | the planets' orbits are tilted to the Earth's by no more than about seven degrees, Mercury's (planet-tilts) | recreated: node research/bright-light/verify.mjs (family K) | Mercury 7.005, Venus 3.395, Mars 1.850, Jupiter 1.304, Saturn 2.486, the standard J2000 inclinations
- OBSERVED | 1.5 takes in the 22 brightest points among the stars, Alpha Centauri's close pair counted as one (count-22) | recreated: node research/bright-light/verify.mjs (family K); public/strata/the-sky-above-you/stars.json | 23 Hipparcos entries at V 1.5 or brighter (Adhara exactly 1.50); Alpha Centauri A and B merged make 22
- OBSERVED | five lie within seven degrees of the ecliptic: Regulus 0.47, Spica 2.05, Antares 4.57, Aldebaran 5.47, Pollux 6.68 (five-stars) | recreated: node research/bright-light/verify.mjs (family K), and a scratch computation over all 63 catalogue rows | the same five at the same latitudes, which match the usual catalogue values
- OBSERVED | the next nearest bright ones, Procyon and Betelgeuse, are 16 degrees off (sixteen-off) | recreated: node research/bright-light/verify.mjs (family K) | -16.02 and -16.03; no star of 1.5 or brighter lies between 6.68 and 16
- CONFIRMED | the 18.6 years in which the line where the Moon's orbit crosses the ecliptic turns once round the Earth, shifting the Moon's monthly path among the stars (nodes-18-6) | https://www.astronomy.com/observing/does-the-moon-follow-the-same-path-across-the-sky-as-the-sun/ | "An imaginary line connecting the nodes through Earth's center rotates around the planet every 18.6 years." verbatim; the band sentences support the shift. Record metadata only: its "where" names Martin Ratcliffe, the page now reads "By Astronomy Staff"
- CONFIRMED | ten degrees is about the width of a fist held at arm's length (fist) | https://science.nasa.gov/skywatching/faq/ | "For 10 degrees, hold up a fist." (arm's length is in the little-finger sentence before it)
- OBSERVED | method: every whole hour after civil twilight with the Moon more than 5 degrees up, every planet and every star of 1.5 or brighter more than 5 up and within 10 (method) | research/bright-light/baserate.mjs | NEAR_DEG 10, BRIGHT 1.5, MOON_MIN_ALT 5, LIGHT_MIN_ALT 5, DARK -6, as stated; a scratch re-implementation reproduced 45 N exactly (5,719; 2,140; table 2's row)
- CONFIRMED | civil twilight: the Sun more than 6 degrees below the horizon (civil-6) | https://aa.usno.navy.mil/faq/RST_defs | "Civil twilight is defined to begin in the morning, and to end in the evening when the center of the Sun is geometrically 6 degrees below the horizon."
- OBSERVED | everything that ever turned up: the five planets, those five stars, and Betelgeuse once (ever-list) | recreated: node research/bright-light/verify.mjs (family G) | eleven names, Betelgeuse only at 55 (see the 55 N line near the end)
- OBSERVED | on a little over a third of the nights the Moon is up, one of the ten is close beside it (third) | recreated: node research/bright-light/verify.mjs (family G) | 36.2 to 38.2 per cent
- OBSERVED | it is only stars slightly more often than it is a planet, at every latitude counted (stars-more) | recreated: node research/bright-light/verify.mjs (family G) | 51.2 to 55.1 per cent; true of counted nights (the step from nights to "the question" is a separate line below)
- OBSERVED | Venus 47.2 degrees from the Sun at its farthest in these nineteen years, by this engine (venus-47) | recreated: node research/bright-light/verify.mjs (family K) | 47.19
- OBSERVED | 22 April 2034, 55 N on the Greenwich meridian: the Moon 9.96 degrees from Betelgeuse at 22:00 UT, the Moon 13 degrees up, Betelgeuse 5 up in the west (night-2034) | research/bright-light/horizons/raw/edge-2034-04-22-moon-55n.txt | Horizons row "2034-Apr-22 22:00 m 281.359193 13.466495" (azimuth 281, west); separation 9.958; Betelgeuse 5.2 up in the west from the engine
- OBSERVED | from the centre of the Earth the same Moon would have been 10.75 degrees away (geo-1075) | recreated: node research/bright-light/verify.mjs (family H) | 10.75, the engine's geocentric Moon with refraction
- CONFIRMED | you stand about 6,400 km from the Earth's centre (earth-radius) | https://nssdc.gsfc.nasa.gov/planetary/factsheet/earthfact.html | "Equatorial radius (km) 6378.137"
- CONFIRMED | the Moon is only about 60 Earth radii away (moon-60) | https://nssdc.gsfc.nasa.gov/planetary/factsheet/moonfact.html | "Semimajor axis (10^6 km) 0.3844"; 384,400 / 6,378 = 60.3
- CONFIRMED | from the surface the Moon sits up to about a degree lower in the sky than from the centre (parallax-degree) | https://nssdc.gsfc.nasa.gov/planetary/factsheet/moonfact.html | from its "Perigee (10^6 km) 0.3633": arcsin(6378.137/363300) = 1.006 degrees (0.951 at the mean distance), times the cosine of the altitude, largest at the horizon; the sentence says what these numbers give
- OBSERVED | JPL Horizons' own position for the Moon at that minute puts it 9.958 degrees from the star (horizons-9958) | recreated: node research/bright-light/verify.mjs (family H); research/bright-light/horizons/raw/edge-2034-04-22-moon-55n.txt | 9.958 is Horizons' Moon against this engine's Betelgeuse (Hipparcos, moved and precessed; aberration and nutation unapplied, under 38 arcseconds by mcp.js), so "the sky's, not this engine's" holds with 0.042 degrees of margin against at most 0.011
- OBSERVED | the evening's brightest light is a planet on 54 to 69 per cent of evenings, Venus or Jupiter on more than half of them everywhere counted (evening) | recreated: node research/bright-light/verify.mjs (family G) | 53.8 to 69.5; Venus or Jupiter at least 51.8; table 3's 6,940 evenings is every date from 2020 to 2038
- CONFIRMED | the 63 Hipparcos stars brighter than magnitude 2.15 (hipparcos-63) | public/strata/the-sky-above-you/stars.json | 63 rows with V under 2.15 in the Hipparcos-V source file (64 at 2.15 or brighter: HIP 4427 sits at exactly 2.15, see finding 8); an outside count at VizieR could not be made (connection reset twice)
- OBSERVED | Sky Now's engine was checked against JPL Horizons when it shipped (sky-now-checked) | public/tools/sky-now/PROVENANCE.md; recreated: node research/the-sky-above-you/verify.mjs | "42 passes / 0 failures"; cmp shows the two ephem.mjs copies identical
- CONFIRMED | each planet's brightness comes from the equations Mallama and Hilton published in 2018 for The Astronomical Almanac (mallama-hilton) | https://arxiv.org/pdf/1808.01973 | title page verbatim (the checker read the abs page, hence NOT FOUND); Crossref 10.1016/j.ascom.2018.08.002: Astronomy and Computing 25, 10-24, 2018
- CONFIRMED | airmass about 2 at 30 degrees up, 5.6 at 10, 38 at the horizon (airmass-values) | public/tools/sky-now/which/sky.mjs | the Kasten and Young formula gives 1.994, 5.586 and 37.92; the Horizons manual names Kasten and Young 1989 for its own airmass (https://ssd.jpl.nasa.gov/horizons/manual.html)
- CONFIRMED | an extinction coefficient from Green's 1992 model, 0.28 magnitude per airmass at sea level (green-028) | http://www.icq.eps.harvard.edu/ICQExtinct.html | "From the July 1992 issue of International Comet Quarterly, Vol. 14, pages 55-59"; equations (2) and (4) and the 0.016 ozone term are there (the checker's 55 per cent is markup); 0.1451 + 0.120 + 0.016 = 0.2811
- MINOR | almost certainly at a gap of at least 1.5 magnitudes (four times the light) with nothing else within one; most likely at 0.75 (rule) | public/tools/sky-now/which/sky.mjs | VERDICT_RULE is clear 1.5, likely 0.75, but verdictFor rounds the gap to a tenth first (line 348; verify family F: 1.46 gives "clear"), so 1.45 magnitudes (3.8 times) already reads "almost certainly", and the answer prints "about 3.8 times the light" (app.mjs 348) beside a rule that says four; finding 10
- CONFIRMED | for satellites, past ones included, it sends you to azmth, Heavens-Above and Astroviewer (refers-out) | public/tools/sky-now/which/app.mjs | lines 303-305, index.html 237, mcp.js 284-286; what each one does is confirmed in the tool lines below
- CONFIRMED | planets and stars keep their places among each other (planets-fixed) | https://openstax.org/books/astronomy-2e/pages/2-1-the-sky-above | "During a single night, even objects we know to have significant motions of their own, such as the nearby planets, seem fixed relative to the stars."
- OBSERVED | magnitudes within 0.0025 for Mercury, 0.0014 Venus, 0.0033 Jupiter, 0.0089 Saturn; Mars up to 0.126 (mag-numbers) | recreated: node research/bright-light/verify.mjs (family B) | identical maxima
- CONFIRMED | the paper's own table puts the two Mars terms at 0.11 and 0.15 magnitude across their full range (mars-terms) | https://arxiv.org/pdf/1808.01973 | "The variations from orbital and rotational effects are 0.15 and 0.11, respectively." and Table 3 (Orbital 0.15, Rotational 0.11), in the PDF (the checker read the abs page)
- OBSERVED | directions within 0.074 degrees for Saturn, 0.025 for Jupiter, under 0.02 for the rest, the Moon included (direction-numbers) | recreated: node research/bright-light/verify.mjs (family C) | maxima 0.0742, 0.0254, Moon 0.0169 (the first two rounded down by under 0.0005 degrees under "to within": not worth a finding)
- OBSERVED | airmass to a median 0.03 per cent, dimming to a median 0.0003 magnitude (air-numbers) | recreated: node research/bright-light/verify.mjs (family D) | 0.032 per cent and 0.0003 (the maximum, 0.0458, is not stated)
- OBSERVED | Horizons' coefficient at each site's height is Green's: 0.2811 at sea level, 0.1761 at 1,600 m, 0.1719 at 1,700 (k-sites) | recreated: node research/bright-light/verify.mjs (family D) | identical; the Horizons manual names Green 1992 for its extinction
- OBSERVED | Venus, Jupiter and Saturn at two sites reach the horizon within a minute of Horizons' times, every one of 24 events (rts-24) | recreated: node research/bright-light/verify.mjs (family E) | worst 0.9 minutes
- CONFIRMED | CycleCalcs' "What Is That in the Sky?" takes a height as well and works for an earlier night (cyclecalcs) | https://www.cyclecalcs.com/what-is-that-in-the-sky.html | "It answers for tonight and for a night last week"; five height choices (inputs named si-alt)
- MINOR | LiveEarthViewer's "What's That in the Sky?", which answers for right now (liveearth) | https://liveearthviewer.com/sky/identify | true of that page, but the same widget on the 132 city pages it links (https://liveearthviewer.com/sky/tonight/london#identify) has "At a different time?" (input sky-id-time), and the identify page calls the city pages "useful if you saw something earlier and want to know what it was"; finding 6
- CONFIRMED | both carry advertising (ads) | https://www.cyclecalcs.com/privacy.html | "This site shows ads served by Mediavine", and the Mediavine tag (scripts.scriptwrapper.com) is in the tool page's HTML; LiveEarthViewer's identify page loads pagead2.googlesyndication.com adsbygoogle.js with client ca-pub-5630628197846834 (both fetched 2026-11-06)
- CONFIRMED | neither takes "next to the Moon" as the question, and neither opens from a link on an answer for a given place, time and direction (no-moon) | https://www.cyclecalcs.com/assets/js/sky-id.js | CycleCalcs' inputs are direction, height, motion, colour, brightness and twinkle; no URLSearchParams or location.search/hash in sky-id.js or main.js (the same grep finds 5 in this tool's app.mjs); LiveEarthViewer's widget takes a direction and a time and reads nothing from the URL, and its city URLs carry only the city
- CONFIRMED | Stellarium Web draws the whole sky for any place and moment from a link, and leaves the identifying to you (stellarium) | https://stellarium.sfo2.cdn.digitaloceanspaces.com/stellarium-web/frontend-1.16.0/assets/index-Dd5gUw0Y.js | the bundle reads $route.query date, lat, lng, elev, az, alt and fov; no identification
- CONFIRMED | azmth, Heavens-Above, Astroviewer and NASA's Spot the Station app all answer "was that the Space Station?" for times in the past (satellite-tools) | https://apps.apple.com/us/app/spot-the-station/id6449235044 | app notes: "the Past Sightings page that shows the history of the past events for the last 2 weeks at the selected location"; the other three are confirmed in the tool lines below
- OBSERVED | passes 86 checks, then breaks the engine five ways and requires each break to be caught (checks-86) | recreated: node research/bright-light/verify.mjs | "91 passed, 0 failed (86 checks, then 5 controls)", exit 0 (after the fixes the page says 90 checks, and the run reads "95 passed, 0 failed (90 checks, then 5 controls)")
- WRONG | seoTitle: "What Is the Bright Star Next to the Moon? It Can Only Be One of Ten" | src/content/strata/the-star-next-to-the-moon.md | not in the record; unqualified, and false by the page's own count (Betelgeuse at 55 N), by its own criteria at 60 N and 65 N (Procyon and Betelgeuse), at the tool's 12-degree radius (Procyon 51, Betelgeuse 48 nights at 45 N) and at the tool's own 2.15 cut (757 nights at 45 N when the only such light near the Moon is outside the ten); finding 1
- MINOR | "The question is usually asked as though the answer must be a planet. About half the time it is a star." (layer lines 69-71) | research/bright-light/prior-art-2026-11-06.md | not in the record and unsourced; the count is of nights a light is within 10 degrees at any phase, not of questions; the screen's autocomplete lists hold "star next to the moon", "what is the bright star near the moon" and "what is that star next to the moon" (ten suggestions each) beside the planet phrasings; finding 4
- CONFIRMED | "Nothing dominates ... Venus is rarer than its fame suggests", partly because the count's twilight limit leaves the crescent-and-Venus pairs out | research/bright-light/out/baserate.json | table 2 holds; recounted in scratch with the Sun limit at sunset instead of -6 (the same code path reproduces the 45 N row exactly at -6): Venus 220 against Jupiter 304, Saturn 293, Spica 299, Regulus 288 at 45 N; 251 against 311, 298, 311, 284 at 30 N; the sentence survives twilight
- MINOR | "The controls. ... a wording rule with its thresholds halved, and a Moon a degree out on the 2034 night" (layer lines 172-174) | research/bright-light/verify.mjs | line 338 halves the measured margin, which doubles the thresholds (labelled "swapped" there and in header line 26); line 339 lowers Betelgeuse by a degree, while family H's Moon is Horizons' and untouched; both controls do turn red; finding 9
- CONFIRMED | "it adds what none of them had: a link an assistant can hand you that opens on the answer, the same answer as a call, the dimming of the air, and a margin stated instead of a confidence invented" | https://www.cyclecalcs.com/assets/js/sky-id.js | CycleCalcs: no URL state, no airmass or extinction (a twilight limiting-magnitude gate only), labels from heuristic scores (confidenceOf: score at least 0.6 and a lead of 0.22); LiveEarthViewer: fixed planet magnitudes (Venus -4, Jupiter -2.2, Mars 0.5, Saturn 0.6), a band from the light's own magnitude, no extinction; Stellarium Web Engine: "ad-hoc code to be replaced when proper extinction is computed" (src/modules/planets.c), extinction only in its sky-background model
- CONFIRMED | relates note: the Moon is about 1.6% smaller on the horizon, and parallax moves it by up to a degree against the stars | src/content/strata/moon-illusion.md | that layer says "about 1.6% smaller"; 1.65 per cent from NASA's 384,400 and 6,378 km
- CONFIRMED | tool: binoculars held steady can show some of Jupiter's four largest moons in a line beside it; one or more may be hidden | https://science.nasa.gov/solar-system/skywatching/night-sky-network/spot-the-king-of-planets-observe-jupiter/ | "Most binoculars will show at least one or two moons ... sometimes they can pass behind or in front of Jupiter"; RMG: "Jupiter has four moons which can be seen using binoculars" (tool sources 1-4)
- MINOR | tool: "Venus is N% lit right now; binoculars or a small telescope can show its phase", given at every phase and with the Sun up | https://science.nasa.gov/image-article/apod-2024-january-8-the-phases-of-venus/ | the APOD's words carry it ("usually requires binoculars or a small telescope to clearly see its current phase"), but the author's own notes (research/bright-light/sources-tool-2026-11-06.json, entries 5 and 6) call a nearly full Venus "hard to read in binoculars" and suggest a warning "never to sweep binoculars near the Sun"; checksFor (app.mjs 275-280) adopts neither and never looks at whether the Sun is up; finding 5
- CONFIRMED | tool: Saturn's rings need a telescope; good binoculars show only that it is not quite round | https://www.rmg.co.uk/stories/space-astronomy/solar-system-through-your-own-telescope | "With good binoculars it can be seen to have a non-circular shape and the rings can be seen with a small telescope"; by magnitude.mjs's own geometry the rings are 6.5 to 8 degrees open through late 2026 and opening, so the check holds now (it would not have near the 2025 edge-on)
- CONFIRMED | tool: to the eye Mars looks orange-red | https://www.nasa.gov/blogs/watch-the-skies/2018/07/30/go-outside-and-see-mars/ | "with its orange-red color, will be hard to miss"
- OBSERVED | tool: Mercury never gets more than about 28 degrees from the Sun | recreated: node research/bright-light/verify.mjs (family K) | 27.82 degrees over 2020-2038, the usual maximum elongation
- CONFIRMED | tool: a star twinkles and can flicker in colour; a planet usually shines more steadily, though low in the sky it can twinkle too | https://stardate.org/faq/why-do-stars-twinkle | “results in apparent changes in color — the familiar 'twinkling' effect”; planets "only 'twinkle' when ... close to the horizon"; NASA Imagine: "the turbulent effects are averaged out" (the screen's section 4b)
- CONFIRMED | tool: the Space Station looks like a very bright star moving across the sky, without flashing or changing direction | https://www.nasa.gov/spot-the-station/ | "appears similar to an airplane or a very bright star moving across the night sky, except it does not have flashing lights or change direction"
- CONFIRMED | tool: aircraft carry a red light on the left, a green one on the right and a white one at the rear, with flashing red or white anticollision lights (the US rule for airliners) | https://www.ecfr.gov/api/versioner/v1/full/2026-11-04/title-14.xml?part=25&section=25.1385 | 25.1385(b) "the red light is on the left side and the green light is on the right side", (c) a rear "white light"; 25.1401(c) 40 to 100 flashes a minute, (d) "aviation red or aviation white"
- CONFIRMED | tool: a line of lights moving together can be a string of Starlink satellites soon after their launch | https://noirlab.edu/public/images/iau-ann19035a/ | "the density of these satellites is significantly higher in the days after launch"; ESO's Messenger 184 calls the first batch a "string of pearls"
- CONFIRMED | tool: azmth checks every satellite in low orbit against your sky at a moment up to about a week back | https://azmth.space/what-was-that | "it will check every low-orbit satellite against your sky at that exact moment"; "How far back can it identify a sighting? About a week."
- CONFIRMED | tool: Heavens-Above lists the Space Station's passes over your place with how bright each was, and its back button shows the previous ten days | https://www.heavens-above.com/PassSummary.aspx?satid=25544&lat=51.51&lng=-0.13&loc=London&alt=0&tz=UCT | default period 6 to 16 Nov 2026; the "<" button ("show previous search period"), pressed by form post, returned 27 Oct to 6 Nov 2026 with each pass's magnitude
- CONFIRMED | tool: Astroviewer shows past passes of the Space Station and China's Tiangong for a date you pick | https://www.astroviewer.net/iss/en/observation-css.php | "Show past Tiangong passes ... Select a date in the past"; the ISS page has the same
- WRONG | tool: the "It's moving" button answers "Then it is not a planet or a star." | http://www.ianridpath.com/ufo/astroufo1.html | "Additional information such as 'it wasn't there before', 'it appeared to move slowly', or 'it dodged around' are still consistent with characteristics of stars and planets"; the autokinetic effect makes "a stationary object appear to move"; the button is unqualified and the verdict absolute (index.html 202, app.mjs 300); finding 2
- WRONG | tool and MCP: "the 62 brightest points among the stars (magnitude 2.15 and brighter)" and "(Hipparcos V 2.15 and brighter)" | public/strata/the-sky-above-you/stars.json | the catalogue keeps V under 2.15 (catalogue.mjs: "The 63 Hipparcos stars with V < 2.15"); HIP 4427, Gamma Cassiopeiae, has V 2.15 exactly and is not carried; finding 8
- MINOR | MCP which_light: "The coordinates reach this server to be computed and are never stored; the call is counted by method, tool and outcome, never by its arguments." | worker.js | true as far as it goes, but every non-browser request, this call included, is also written as a machine event with country, network and ASN (machineEvent, worker.js 2589-2615, called for every request at 8586-8590), which the site says for its other tools (mcp.js 614); finding 7
- CONFIRMED | tool: the brightest star, Sirius, is about -1.4 before any dimming | https://science.nasa.gov/skywatching/faq/ | "The brightest star in our skies ... is Sirius"; Hipparcos V -1.44 (public/tools/sky-now/catalogue.mjs)
- CONFIRMED | tool hood and MCP limits: Mars is computed without two corrections that together span about a quarter of a magnitude and can be off by a little over a tenth; the rest within a hundredth of Horizons | https://arxiv.org/pdf/1808.01973 | 0.15 + 0.11 = 0.26; verify family B: Mars 0.1262, the others at most 0.0089
- CONFIRMED | tool hood, MCP and magnitude.mjs: Horizons' magnitudes follow Mallama and Hilton, and its extinction column is Green's coefficient times Kasten and Young's airmass | https://ssd.jpl.nasa.gov/horizons/manual.html | "Planetary & satellite magnitudes are based on: Mallama, A., Hilton, J.L. ..."; "Airmass computation is based on: ... Kasten F., Young A."; "Atmospheric extinction computation based on: ... Green D.W.E., International Comet Quarterly, July 1992, Vol 14, pp. 55-59"
- CONFIRMED | magnitude.mjs: Saturn's ring switch "is exactly the one Horizons states for its own APmag" | research/bright-light/horizons/raw/geo-699.txt | Horizons: "includes the ring system when the solar phase angle ... is less than 6.5 degrees ... AND effective ring inclination is less than 27 degrees. Otherwise, the value is that of the globe only"; the paper's Eq. 10 "applies to α < 6.5 and β < 27"
- CONFIRMED | magnitude.mjs: the equation numbers and ranges, and that the Mars terms need surface tables the paper takes from Mallama (2007) and does not print | https://arxiv.org/pdf/1808.01973 | Eq. 2 with "2.1 < α < 169.5" in section 4.1; Eq. 3 and 4 split at 163.7 (section 3.2); Eq. 6 "α ≤ 50"; Eq. 8 for α up to 12, Eq. 9 from 12 to 130; Eq. 12 from 6 to 150; "The Mars paper tabulates the contribution from the sub-longitude at ten-degree intervals"
- CONFIRMED | magnitude.mjs: Saturn's pole at right ascension 40.589 - 0.036T, declination 83.537 - 0.004T (Archinal et al. 2018, Table 1) | https://d9-wret.s3.us-west-2.amazonaws.com/assets/palladium/production/s3fs-public/atoms/files/WGCCRE2015reprint.pdf | "Saturn α0 = 40.589 − 0.036T δ0 = 83.537 − 0.004T" under "Table 1 continued"; doi 10.1007/s10569-017-9805-5
- MINOR | sky.mjs comment: Green's aerosol scale height "may vary by a factor of 2 on any given night", "which is why the page lets a reader change it"; the extinction column checked in "family E" | http://www.icq.eps.harvard.edu/ICQExtinct.html | the quotation is verbatim, but nothing on the page changes the scale height (only the observer's height, through the address's h, and then the hood still prints "(sea level)", app.mjs 465), and the check is family D; finding 12
- CONFIRMED | sky.mjs: Alpha Centauri A at V -0.01 and B at V 1.35, 19 arcseconds apart at the catalogue's epoch; normal (20/20) vision separates points about an arcminute apart | https://www.webvision.pitt.edu/book/part-viii-psychophysics-of-vision/visual-acuity/ | "For a visual acuity of 6/6 (20/20) ... the minimum angle of resolution (MAR) is one minute of arc"; the catalogue's J1991.25 positions give 19.11 arcseconds; together -0.283
- MINOR | table 1, 55 N: "nights one of the ten was within 10° of it" 1,895 (36%) | research/bright-light/out/baserate.json | the column counts nights with any listed light: recounted in scratch with the page's criteria, 1,894 nights had one of the ten and 1 had only Betelgeuse; "with only stars" 1,040 includes that night; the percentages do not move; finding 11
- OBSERVED | (added by the fixes, checked by the author) dek and plain: "Fainter stars add nine more names." (dek-fainter) | recreated: node research/bright-light/verify.mjs (family G) | "fainter stars the tool carries that came within 10 degrees of the Moon, most nights first: Nunki, Elnath, Alhena, Algieba, Castor, Hamal, Kaus Australis, Denebola, Shaula": nine
- OBSERVED | (added by the fixes, checked by the author) "Castor (1.58), Shaula (1.62) and Elnath (1.65) lie near the line just below it" (cut-15) | public/tools/sky-now/catalogue.mjs | Hipparcos V 1.58, 1.62 and 1.65 in the catalogue rows; the agent's finding 3 gives the same three values
- OBSERVED | (added by the fixes, checked by the author) nine fainter stars join the list, every one within 14 degrees of the ecliptic, in order of how many nights they spent beside the Moon (fainter-nine) | recreated: node research/bright-light/verify.mjs (families G and K) | the farthest is 13.79 degrees from the ecliptic; nights summed over the eight latitudes in research/bright-light/out/baserate.json: Nunki 2,399, Elnath 2,063, Alhena 1,678, Algieba 1,313, Castor 1,040, Hamal 1,027, Kaus Australis 737, Denebola 616, Shaula 298, the page's order
- OBSERVED | (added by the fixes, checked by the author) on 25 to 28 per cent of the nights when anything the tool carries was within 10 degrees of the Moon, the only thing there was one of these nine (fainter-share) | recreated: node research/bright-light/verify.mjs (family G) | 25.3 (55 N, 642 of 2,537) to 27.5 per cent (45 S, 786 of 2,855); the agent's independent recount gave 757, 806 and 811 for 45 N, the equator and 30 S, the file's numbers exactly

## Findings
- [fixed] MED, the seoTitle ("It Can Only Be One of Ten"): now "What Is the Bright Star Next to the Moon? Five Planets, Five Stars, Counted Night by Night". The plain and the dek now put the cut first ("Among the brightest lights (magnitude 1.5 or brighter)"; "among everything of magnitude 1.5 or brighter"), and the layer counts at the tool's own cut as well: baserate.mjs records every star brighter than 2.15 that came within 10 degrees of the Moon, and "Only ten things" now names the nine fainter ones and says that on 25 to 28 per cent of the nights anything the tool carries is beside the Moon, one of them is the only light there (verify family G; the agent's recount matches the file). The page keeps its 10-degree radius and its eight latitudes, and states both; the tool's 12-degree search is a property of the tool, written in its own limits.
- [fixed] MED, the "It's moving" button and its verdict: the button now reads "It moves against the stars", and the answer is conditional: "If it really moves against the stars around it, it is not a planet or a star." followed by the caution from Ridpath's page that a still light on its own can seem to wander (small movements of the eyes, clouds passing in front), with the test of holding it against a roof edge, a branch or a star beside it. The refusal list on the page says the same.
- [fixed] MINOR, the dek and plain against the eleventh light: the plain now says "Betelgeuse got in once", both say fainter stars add nine more names, and "Only ten things" says the cut at 1.5 decides the list, naming Castor (1.58), Shaula (1.62) and Elnath (1.65) just below it.
- [fixed] MINOR, "The question is usually asked as though the answer must be a planet": removed. The page now says only what it counted: on a little over a third of the nights the Moon is up a light that bright is beside it, and it is only stars slightly more often than it is a planet, at every latitude counted.
- [fixed] MINOR, the Venus binocular advice: checksFor gives no binocular advice while the Sun is up, and the wording follows the author's own source notes: below half lit, "binoculars or a small telescope can show the shape in twilight"; otherwise "a small telescope shows its phase".
- [fixed] MINOR, LiveEarthViewer "answers for right now": re-read at https://liveearthviewer.com/sky/tonight/london (the widget's "At a different time?"). The layer now says "whose city pages take an earlier time too", and the source record's liveearth entry cites the London page.
- [fixed] MINOR, the which_light privacy sentence: re-read worker.js machineEvent (country, organisation, ASN for every non-browser request). It now reads "The coordinates reach this server to be computed and are never stored; the call itself is counted, with the kind of client, its country and its network."
- [fixed] MINOR, "magnitude 2.15 and brighter": the tool page now says "everything brighter than magnitude 2.15" and mcp.js "Hipparcos V brighter than 2.15", as the catalogue keeps them (V under 2.15; Gamma Cassiopeiae at 2.15 exactly is not carried).
- [fixed] MINOR, the controls described as they are not: the layer now says "a wording rule fed half of each real gap, and Betelgeuse moved a degree on the 2034 night", and verify.mjs's control labels and header comment say the same.
- [fixed] MINOR, the wording rule's rounding: verdictFor now compares the exact gap (sky.mjs), the answer prints the gap to two decimals, and verify family F now checks that 1.50 gives "almost certainly" while 1.49 and 1.46 give "most likely".
- [fixed] MINOR, table 1 at 55 N counting Betelgeuse's night as "one of the ten": the column heading is now "nights a light that bright was within 10° of it", which is what the file counts; table 2 still footnotes Betelgeuse, and its 55 N row sums to 1,894.
- [fixed] MINOR, the sky.mjs comment: it now names family D, and no longer says the page lets a reader change the scale height; it says the page calls the model a clear-night one and that haze dims low lights more, which the refusal list does ("The dimming it applies is a standard clear-night model for the air at sea level, or at the height a link gives"). The hood prints the height the coefficient is for when a link gives one.