Scoreboard

Every test we ran, and how many passes luck alone would give.

46tests and checks run

25fair tests: decided before looking, with a chance number

3of those passed (chance by luck under 0.05)

1.2passes luck alone would give from 25 tests

A few more passes than luck would give, but not many. With every idea wrong we would still get 3 or more passes 13 times in 100.

Our passes, and what happened when we checked them a second way:

Did Picardy sweat years follow a big beech-nut year at Seesen, Germany? A small rise, but just as big the year AFTER outbreaks, and it fades when slow changes are allowed for (0.08-0.11).
Do departments with more Picardy sweat outbreaks have more hantavirus cases today? Yes across France, but both are simply northern (see the next test).
US 1922-38: do sleepy sickness and flu deaths change together in the same year? Yes, modestly (r +0.43), but it does not repeat in England and Wales or in Scotland.

Fair tests by their chance-by-luck number

If none of our ideas were real, the bars would sit close to the dashed marks: about 5 in every 100 tests under 0.05, 50 in every 100 over 0.50. Real discoveries would pile up in the left-hand bar.

under 0.05 3 (luck: 1.2)

0.05 to 0.10 1 (luck: 1.2)

0.10 to 0.25 5 (luck: 3.8)

0.25 to 0.50 7 (luck: 6.2)

0.50 or more 9 (luck: 12.5)

Our testsWhat luck alone would give

Tests chosen after looking (6)

We picked these after seeing the data, so they do not count above. They are hints to test again with fresh data, not findings.

No sign 0.47 Same match, inside the north-east only (so north versus south cannot fake it)
Passed 0.03 Are the rings narrow two years after each of the 11 eruptions?
Close 0.069 Were the two summers before England's 5 sweating-sickness outbreaks wetter than usual?
Close 0.063 Were the two summers before Picardy sweat years wetter in northern France?
Passed 0.008 Were cocoliztli epidemic years wet in central Mexico (drought atlas)?
Close 0.06 Do more cities stop in drier 20-year periods?

Every test, newest first

Date	What we asked	Decided	Result
2026-10-27	Did Picardy sweat years follow a big beech-nut year at Seesen, Germany? Sweating sickness followed seed-boom years, as modern European hantavirus does · scripts/sweat_mast.py	Before looking	Passed 0.04 A small rise, but just as big the year AFTER outbreaks, and it fades when slow changes are allowed for (0.08-0.11).
2026-10-27	Was the summer 2 years before Picardy sweat years warm at Dijon (grape harvests)? Sweating sickness followed seed-boom years, as modern European hantavirus does · scripts/sweat_mast.py	Before looking	No sign 0.86 Slightly cool, not warm.
2026-10-27	Was the summer 2 years before warm in the English temperature record? Sweating sickness followed seed-boom years, as modern European hantavirus does · scripts/sweat_mast.py	Before looking	No sign 0.72 Slightly cool, not warm.
2026-10-27	Seed-boom rule with Dijon temperatures (summer t-2 warmer than t-3) Sweating sickness followed seed-boom years, as modern European hantavirus does · scripts/sweat_mast.py	Before looking	No sign 0.73 No sign.
2026-10-27	Do departments with more Picardy sweat outbreaks have more hantavirus cases today? The Picardy sweat struck the same places where hantavirus infects people in France today · scripts/sweat_map.py	Before looking	Passed 0.001 Yes across France, but both are simply northern (see the next test).
2026-10-27	Same match, inside the north-east only (so north versus south cannot fake it) The Picardy sweat struck the same places where hantavirus infects people in France today · scripts/sweat_map.py	After looking	No sign 0.47 No match at all: the north-south divide explained the first result.
2026-10-26	Do the Bamboo Annals' sky and weather entries land on ice-core eruption years more than chance? The 'three suns' of 1561 BC fall in the same year as the 2022 Thera candidate, but only by the current text's own year count, which was built centuries later · scripts/xia_annals.py	Before looking	No sign 0.23 3 of 8 entries hit; chance would give that about 1 time in 4.
2026-10-26	Was 1561-1546 BC unusually dry in the Qilian juniper rings? The early books remember failed harvests, mixed cold and heat and a five-year drought at the Xia's end; the ice cores show four eruptions in 1561-1550 BC, about when Erlitou ended · scripts/xia_rings.py	Before looking	No sign 0.1 Somewhat: 50th driest of 486 sixteen-year windows.
2026-10-26	Are the rings narrow two years after each of the 11 eruptions? The early books remember failed harvests, mixed cold and heat and a five-year drought at the Xia's end; the ice cores show four eruptions in 1561-1550 BC, about when Erlitou ended · scripts/xia_rings.py	After looking	Passed 0.03 Yes at 2 years, but 7 lags were tried, so about 0.2 really.
2026-10-26	Check Known answer: do the rings react to Aniakchak (1628 BC)? Juniper rings (Qilian Mountains, 1,400 km west of Erlitou): 1561-1546 BC was somewhat dry, with sharp dry years in 1558 and 1548 BC, from only 12 cores · scripts/xia_rings.py	Before looking	No chance number Yes: 1626 BC is one of the driest years (z -2.5), two years after.
2026-10-26	Could all five planets be seen together in 1576 BC? Pankenier's five-planet omen of 1576 BC: our sky check agrees with Keenan, only four planets could be seen · scripts/xia_planets.py	Before looking	No chance number No: all five were close only when 4 degrees from the Sun; only four could be seen (agrees with Keenan).
2026-10-26	Check Known answer (Richardson 2007): do the real 1872 winds carry the ship from Santa Maria to where she was found? Mostly known: the drift fits abandonment on 25 Nov (Richardson 2007); only the storm of 30 Nov to 2 Dec against the 1873 'no heavy weather' argument may be new · scripts/celeste_drift.py	Before looking	No chance number Yes, easily, but only if she kept sailing across the wind.
2026-10-26	How far away can an object be and still look sharper than the Sun in Bonilla's telescope? Bonilla's own photograph puts the bodies within a few km: birds or insects close to the observatory, not a comet · scripts/bonilla_focus.py	Before looking	No chance number Under about 25 km, and within about 4 km for the blur he describes; rules out the 2011 comet distances.
2026-10-26	How often did north-west China get hail by the 3rd lunar month? What fell at Qingyang: hail or a meteorite shower? Both fit the sizes; early-spring hail is uncommon there, and showers of countless meteorites are rare everywhere · scripts/qingyang_texts.py	Before looking	No chance number Rarely: 1 of 14 dated north-west hailstorms.
2026-10-25	Do Korean 'dim Sun' years (Goryeosa) come just after big eruptions in the ice? Korean dim-Sun records track volcanic eruptions (Goryeo period, 918 to 1392) · scripts/goryeosa_dimsun.py	Before looking	No sign 0.25 2 of 21 dim-Sun years fall in the 2 years after a big eruption, about what chance gives; most are spring dust-storm months.
2026-10-25	Is the 'former sky has fallen' myth more common within 1,000 km of a young impact crater? Sky-fall myths are no more common near young craters than anywhere else · scripts/myth_craters.py	Before looking	No sign 0.12 4.7% near craters vs 3.0% elsewhere, shuffled within countries: not enough to trust.
2026-10-25	Is the 'world destroyed by fire' myth more common near young craters? Sky-fall myths are no more common near young craters than anywhere else · scripts/myth_craters.py	Before looking	No sign 0.17 14.9% vs 14.2%: no difference.
2026-10-25	Is the 'patched sky' myth more common near young craters? Sky-fall myths are no more common near young craters than anywhere else · scripts/myth_craters.py	Before looking	No sign 0.35 0.7% vs 0.6%: no difference.
2026-10-25	Is the 'objects and stones revolt' myth more common near young craters? Sky-fall myths are no more common near young craters than anywhere else · scripts/myth_craters.py	Before looking	No sign 0.73 0.7% vs 4.2%: the other way round.
2026-10-25	Check Control: does an unrelated myth (earth-diver) look 'near craters' once regions are allowed for? Sky-fall myths are no more common near young craters than anywhere else · scripts/myth_craters.py	Before looking	No sign 0.94 Raw count said twice as common near craters; shuffling within countries removed it, as it should.
2026-10-25	Were the two summers before England's 5 sweating-sickness outbreaks wetter than usual? Sweating sickness outbreaks followed slightly wet summers: a weak hint that did not strengthen · scripts/sweat_drought.py	After looking	Close 0.069 Wetter, but the lag and region were picked after looking; 1517 goes the other way.
2026-10-25	Were the two summers before Picardy sweat years wetter in northern France? Sweating sickness outbreaks followed slightly wet summers: a weak hint that did not strengthen · scripts/sweat_picardy.py	After looking	Close 0.063 Slightly wetter; measure chosen after the English result.
2026-10-25	Repeat, fairer: each French outbreak measured at its own department, pattern slid along the record Sweating sickness outbreaks followed slightly wet summers: a weak hint that did not strengthen · scripts/sweat_local.py	Before looking	Close 0.062 Same lean, but tiny: 0.06 of a typical year-to-year swing.
2026-10-25	Seed-boom rule: was the summer 2 years before French outbreaks warmer than the summer 3 years before? Sweating sickness followed seed-boom years, as modern European hantavirus does · scripts/sweat_local.py	Before looking	No sign 0.87 Slightly the opposite.
2026-10-25	Same seed-boom rule for England's 5 outbreaks Sweating sickness followed seed-boom years, as modern European hantavirus does · scripts/sweat_local.py	Before looking	No sign 0.48 No difference.
2026-10-25	Check Known answer: do the 117 recorded ergot epidemics fall in cold summers? Check that works: ergot epidemics come in cold summers · scripts/ergot_climate.py	Before looking	Passed 0.001 Yes, clearly (less than 1 in 1,000 by luck). The method finds a real link when there is one.
2026-10-25	Did the 9 dancing plagues fall in cold, ergot-type summers? The dancing plagues were ergot poisoning · scripts/ergot_climate.py	Before looking	No sign 0.98 The opposite: warm summers after dry ones.
2026-10-25	Was grain dear in the years of the dancing plagues? Strasbourg 1518 came after a great drought and two years of dear bread · scripts/ergot_climate.py	Before looking	No sign 0.4 No price signal across the dances; only 1516-17 Strasbourg was dear.
2026-10-25	Did cocoliztli epidemics start when rain returned after a drought? Cocoliztli broke out in central Mexico after drought (within two years) · scripts/coco_drought.py	Before looking	No sign 0.14 Weak; and the year before epidemics was not dry (mean +0.34, wet side).
2026-10-25	Were cocoliztli epidemic years wet in central Mexico (drought atlas)? Cocoliztli broke out in wet years in central Mexico · scripts/coco_drought.py	After looking	Passed 0.008 Yes in the drought atlas, but this was spotted after looking.
2026-10-25	Repeat with a second record: Queretaro bald cypress Cocoliztli broke out in wet years in central Mexico · scripts/coco_drought.py	Before looking	No sign 0.4 No year-by-year wet signal; the first result did not repeat.
2026-10-25	Was the decade before each Maya city's last monument drier than usual (Yok Balum cave)? Each Maya city fell silent during a local dry spell · scripts/maya_drought.py	Before looking	No sign 0.46 No.
2026-10-25	Do more cities stop in drier 20-year periods? Each Maya city fell silent during a local dry spell · scripts/maya_drought.py	After looking	Close 0.06 Only because of 780-799; the share of cities ending runs the wrong way.
2026-10-25	US 1922-38: do sleepy sickness and flu deaths change together in the same year? Sleepy sickness and flu rise in the same years · scripts/el_flu.py	Before looking	Passed 0.03 Yes, modestly (r +0.43), but it does not repeat in England and Wales or in Scotland.
2026-10-25	Repeat for England and Wales 1921-37 Sleepy sickness and flu rise in the same years · scripts/el_flu.py	Before looking	No sign 0.4 Does not repeat (r +0.07).
2026-10-25	Repeat for Scotland 1922-37 Sleepy sickness and flu rise in the same years · scripts/el_flu.py	Before looking	No sign 0.8 Does not repeat (r -0.23).
2026-10-25	US: does sleepy sickness follow flu a year later (Ravenholt and Foege)? Sleepy sickness follows flu waves by about a year (Ravenholt and Foege) · scripts/el_flu.py	Before looking	No sign 0.5 No (r -0.01).
2026-10-25	England and Wales: one year after flu? Sleepy sickness follows flu waves by about a year (Ravenholt and Foege) · scripts/el_flu.py	Before looking	No sign 0.85 No, slightly the other way (r -0.26).
2026-10-25	Is the six-star count explained by Pleione's position 100,000 years ago? The lost-sister stories date from 100,000 years ago, when Pleione stood clear of Atlas · scripts/pleiades_lost_sister.py	Before looking	No chance number No: the six visible stars are the six bright ones; the count is a brightness step, not glare.
2026-10-25	Do the Odyssey's sky clues single out 1178 BC (the paper says 1 in 2,000)? The Odyssey's sky clues single out the eclipse day, 16 April 1178 BC · scripts/odyssey_dates.py	Before looking	No chance number No: a fitting new Moon comes about once every 64 years, and the eclipse was probably not total at Ithaca (chance 0.31).
2026-10-25	Did a ring-of-fire eclipse cross Gibeon on 30 October 1207 BC? Joshua's 'Sun stand still' was the annular eclipse of 30 October 1207 BC · scripts/eclipse_local.py	Before looking	No chance number Yes, near sunset, whatever the Earth-rotation correction (chance about 1); matches Khalisi 2021, so not new.
2026-10-25	Was the 5 March 1223 BC eclipse total at Ugarit? KTU 1.78 records the total eclipse of 5 March 1223 BC · scripts/eclipse_local.py	Before looking	No chance number Probably not: chance 5-11 percent with standard Earth-rotation values.
2026-10-25	Can the Korean daily annals date the 1453 eruption to the month? Daily Korean records can date the 1452/53 eruption to the month · scripts/sillok_sky.py	Before looking	No chance number No: they record no Sun colours at all in 1447-1459 and miss Kuwae 1459 too.
2026-10-25	How often is there an all-day fog in Istanbul in late spring? Base rate: an all-day fog in Istanbul in late spring almost never happens	Before looking	No chance number 0 in 593 comparable days of modern weather reports: the 1453 fog is truly unusual (but rests on one writer).
2026-10-24	Which tree ring rises first: 1053 (fits a December 1052 event) or 1055 (fits the 1054 supernova)? The 1052 white vapour matches the 1052 radiocarbon candidate	Before looking	No chance number A draw: 3 tree series say 1053, 3 say 1055.
2026-10-24	Was the 1261 glow unusual for its time? 1261: no sky record falls in the right window, and the obvious one was already considered	Before looking	No chance number No: glows were logged in 8 of 12 years around then, so it shows an active Sun, not one storm.