[{"data":1,"prerenderedAt":879},["ShallowReactive",2],{"layer:exploration:deepen":3},{"layer":4,"contentHash":860,"dependencyHashes":861,"approval":872,"releaseId":878},{"schemaVersion":5,"conceptId":6,"locale":7,"depth":8,"revision":5,"title":9,"subtitle":10,"summary":11,"objectives":12,"estimatedMinutes":18,"plate":19,"blocks":43,"sourceIds":855,"reviewStatus":856,"authoring":857},1,"exploration","en","deepen","Mechanism, harder numbers, and how historians know","Why the monsoon reverses, how clock drift compounds, and the method behind contested figures","Go beneath Discover's facts into mechanism and method: why the monsoon reverses, how clock drift compounds over a long voyage, an edge case in kamal readings, how historians back-project contested figures, how to weigh one account against another, and what shipwreck years teach about mean vs median.",[13,14,15,16,17],"Explain the pressure mechanism that reverses the monsoon wind each year.","Calculate how clock drift compounds over a long voyage, and why a chronometer reduces but never removes longitude error.","Work through an edge case where an unnoticed instrument fault biases a measurement in a predictable direction.","Describe two methods historians use to reconstruct a contested number, and show how changing one assumption changes the answer.","Judge which kind of surviving account to trust more for exact quantities versus for the general shape of events.",45,{"title":20,"rows":21},"Lesson plate",[22,25,28,31,34,37,40],{"label":23,"value":24},"Depth","Deepen",{"label":26,"value":27},"Reading time","≈ 45 minutes",{"label":29,"value":30},"Prior knowledge","Discover: exploration",{"label":32,"value":33},"Chapters","7",{"label":35,"value":36},"Labs","Voyage map, evidence sort, shipwreck data lab",{"label":38,"value":39},"Hard parts","None new; builds on Discover's chapter 7",{"label":41,"value":42},"Numbers","Every figure computed and asserted in gen_deepen.py",[44,48,54,60,63,84,104,109,127,150,155,158,174,190,223,228,240,245,258,263,277,282,285,297,310,315,320,323,339,350,354,357,370,376,381,384,419,424,427,505,556,569,574,577,589,619,623,627,632,657,661,689,710,832,846],{"id":45,"type":46,"markdown":47},"intro-deepen","prose","Discover told you *what* happened: the monsoon reverses, clocks helped find longitude, historians disagree about the 1492 population of the Americas. This layer asks the harder question about each of those: **why**, exactly, and **how do we know**?\n\nExpect more arithmetic than before, some of it several steps long, and a few places where the honest answer is *it depends on an assumption, and here is what happens if you change it*. That is not historians being sloppy. It is what careful reasoning about the past actually looks like.",{"id":49,"type":50,"variant":51,"title":52,"markdown":53},"how-to-read-deepen","callout","observation","What is different about this layer","Discover gave you facts and ranges. Here you will rebuild some of those ranges yourself, one assumption at a time, and see exactly why a well-kept clock still was not a perfect clock, why a kamal string that has stretched by two centimetres can quietly mislead a navigator, and why a wide range of historians' estimates is not the same thing as historians shrugging.",{"id":55,"type":56,"title":57,"eyebrow":58,"navLabel":59},"ch01","chapter","Why the monsoon wind actually reverses","Chapter 01","1 Why it reverses",{"id":61,"type":46,"markdown":62},"monsoon-mechanism-intro","Discover gave you the monsoon's timetable: south-west from about June to September, north-east from about November to February. It did not tell you **why** the wind flips. The cause is not magic and it is not the ocean \"running out\" of wind. It is a simple, repeatable fact about how land and water respond to sunlight.\n\n**Land heats up faster than the ocean, and it cools down faster too.** A patch of ocean absorbs the Sun's energy through metres of water that mixes and moves, so its surface temperature barely changes through the year. A patch of land heats and cools right at its surface, so it swings far more between seasons.",{"id":64,"type":65,"title":66,"items":67},"steps-summer-monsoon","steps","How the summer (south-west) monsoon is built, step by step",[68,72,76,80],{"title":69,"tag":70,"text":71},"The Sun climbs high over India","May-June","The Indian subcontinent heats rapidly. Its land surface gets far hotter than the Arabian Sea and Bay of Bengal at the same latitude.",{"title":73,"tag":74,"text":75},"Hot air over the land rises","continuing","Warm air expands and rises, and as it rises it leaves lower pressure at the surface beneath it: a broad low-pressure zone over the heated land.",{"title":77,"tag":78,"text":79},"Air is pulled in from the cooler sea","June onward","Air always moves from higher pressure to lower pressure. Moist, relatively cooler air is drawn in off the ocean to replace the rising air over land.",{"title":81,"tag":82,"text":83},"That inflow becomes the monsoon wind","June-September","The incoming sea air carries huge amounts of moisture picked up over the warm ocean. As it is forced to rise over land and mountains, it cools, and the moisture falls as monsoon rain.",{"id":85,"type":65,"title":86,"items":87},"steps-winter-monsoon","How the reversal happens by winter",[88,92,96,100],{"title":89,"tag":90,"text":91},"The Sun moves south, land cools fast","October-November","With less direct sunlight, the Indian landmass loses its stored heat quickly, because land holds heat far worse than deep water.",{"title":93,"tag":94,"text":95},"The ocean is still comparatively warm","same months","The Arabian Sea and Bay of Bengal are still warmer than the now-cooling land, because water changes temperature slowly in both directions.",{"title":97,"tag":98,"text":99},"Pressure over land rises","November","Cooler air is denser and sinks, building higher pressure over the subcontinent instead of the summer's low.",{"title":101,"tag":102,"text":103},"The wind reverses to blow from land to sea","Nov-Feb","Air now flows from the high pressure over the cool land out towards the lower pressure over the still-warm sea: the dry, land-to-sea north-east monsoon.",{"id":105,"type":50,"variant":106,"title":107,"markdown":108},"aha-differential-heating","aha","One idea explains both halves of the year","Notice that both seasons follow from the *same* fact: land's temperature swings further than the ocean's, in **both** directions. In summer, hotter land pulls in wind from the sea. In winter, colder land pushes wind out towards the sea. You do not need two separate explanations, only one, applied twice.\n\nThis is exactly the same physical idea that makes a sandy beach scorching at noon and cool water still comfortable, or that makes a coastal breeze blow inland on a summer afternoon and reverse after sunset, just acting over an entire subcontinent and six months instead of a beach and a few hours.",{"id":110,"type":111,"prompt":112,"options":113,"explanation":126},"pred-monsoon-mechanism","prediction","Suppose land and ocean actually warmed and cooled at exactly the same rate through the year. What would happen to the monsoon?",[114,117,120,123],{"id":115,"label":116},"a","It would still reverse twice a year, just more gently",{"id":118,"label":119},"b","There would be no reliable seasonal wind reversal at all, because there would be no lasting pressure difference to drive it",{"id":121,"label":122},"c","It would blow from the sea to the land all year round",{"id":124,"label":125},"d","It would only affect the coast, not the interior","**b.** The whole mechanism runs on land and ocean heating and cooling at **different** rates. If they changed temperature identically, there would be no seasonal pressure difference between them, so there would be nothing to pull wind steadily in one direction for months and then reverse it. The Indian Ocean's monsoon exists because land and water are physically different materials, not because of anything special about the wind itself.",{"id":128,"type":129,"itemId":130,"prompt":131,"check":132,"hints":144,"feedback":147},"p-monsoon-mechanism","practice","exploration.deepen-monsoon-cause","Which of these correctly explains why the wind blows from sea to land in summer?",{"kind":133,"options":134,"correct":143},"choice",[135,137,139,141],{"id":115,"label":136},"The ocean pushes air away because it is too hot",{"id":118,"label":138},"Hot land creates a low-pressure zone, and air flows in from the higher-pressure air over the cooler sea",{"id":121,"label":140},"The Earth's rotation reverses the wind for six months",{"id":124,"label":142},"Ships travelling to India pull the wind with them",[118],[145,146],"Air moves from high pressure to low pressure.","Ask which surface, land or sea, is hotter in June, and what hot air does to the pressure beneath it.",{"correct":148,"incorrect":149},"Right: hot land makes rising air and low surface pressure; the sea, still cooler, has relatively higher pressure, so wind flows in from the sea to fill the gap.","Hot air over the land rises, leaving low pressure at the surface. Wind always flows towards lower pressure, so sea air, at relatively higher pressure, flows inland.",{"id":151,"type":56,"title":152,"eyebrow":153,"navLabel":154},"ch02","The mathematics of longitude error, taken further","Chapter 02","2 Longitude error",{"id":156,"type":46,"markdown":157},"longitude-recap","You have already seen that a clock wrong by one minute puts you wrong by a quarter of a degree of longitude, about **27.8 km** at the equator, and that a clock losing 3 seconds a day drifts about **58 km** off over a 42-day Atlantic crossing.\n\nNow stretch the same idea over a much longer voyage: the Magellan-Elcano circumnavigation, **1,082 days** from leaving Spain to the Victoria's return. What happens to the error if nobody resets the clock the whole way?",{"id":159,"type":160,"items":161},"formulas-longitude-error","formulas",[162,165,168,171],{"expression":163,"caption":164},"drift(s) = rate(s\u002Fday) × days","Total seconds a clock has drifted after a given number of days at a constant rate.",{"expression":166,"caption":167},"error(min) = drift(s) ÷ 60","Seconds of drift converted to minutes of time.",{"expression":169,"caption":170},"error(°) = error(min) ÷ 4","The Earth turns 360° in 24 hours = 1,440 minutes, so 4 minutes of time error is 1° of longitude.",{"expression":172,"caption":173},"error(km) = error(°) × 111.3","Degrees of longitude error converted to kilometres at the equator.",{"id":175,"type":176,"title":177,"problem":178,"steps":179,"help":185},"we-magellan-drift","worked_example","How far off could an ordinary clock put the whole expedition?","Assume a fairly ordinary spring-driven clock of the early 1500s drifts by a constant **10 seconds a day**, and nobody has any way to check or reset it against a known longitude for the whole voyage. Over the **1,082-day** Magellan-Elcano expedition, how far wrong could its longitude reading become?",[180,181,182,183,184],"Total drift = 10 s\u002Fday × 1,082 days = **10,820 seconds**.","Convert to minutes: 10,820 ÷ 60 = **180.3 minutes** of time, about 3.0 hours.","Convert to degrees: 180.3 ÷ 4 = **45.1° of longitude**, using 4 minutes of time per degree.","Convert to distance: 45.1 × 111.3 ≈ **5,020 km** at the equator.","For scale: 45.1° is about **12.5%** of the way round the entire planet. A clock nobody could check would have made the expedition's own maps roughly one-eighth of a world turn out of true.",{"simplerExplanation":186,"hints":187},"Multiply the daily drift by the number of days to get total seconds lost, turn that into minutes, divide by 4 to get degrees, then multiply by 111.3 to get kilometres.",[188,189],"Do the unit conversions one at a time: seconds, then minutes, then degrees, then kilometres.","1,082 × 10 is the first number you need.",{"id":191,"type":192,"title":193,"items":194},"timeline-longitude-history","timeline","How the longitude problem was actually solved",[195,199,203,207,211,215,219],{"time":196,"title":197,"text":198},"1610s","Galileo's moons","Galileo proposes timing eclipses of Jupiter's four largest moons as a portable clock. It works well on land; almost impossible to time through a telescope on a rolling ship.",{"time":200,"title":201,"text":202},"1670s","Greenwich founded","Britain sets up the Royal Observatory mainly to build the precise star and Moon tables a rival method, lunar distance, would need.",{"time":204,"title":205,"text":206},"1714","Longitude Act","Parliament offers a prize of up to £20,000 for a method accurate to within half a degree after a voyage to the West Indies.",{"time":208,"title":209,"text":210},"1730s","Harrison's H1","John Harrison, a self-taught carpenter and clockmaker, builds his first sea clock. It works, but is too large and delicate for routine naval use.",{"time":212,"title":213,"text":214},"1759","H4 completed","A fourth, far smaller design: roughly pocket-watch sized, with a temperature-compensated balance wheel and a gimbal mounting against the ship's motion.",{"time":216,"title":217,"text":218},"1761-62","Jamaica trial","H4 loses only about 5 seconds over 81 days at sea, comfortably inside the prize's half-degree limit.",{"time":220,"title":221,"text":222},"1773","Harrison paid","Parliament finally pays Harrison most of the prize, after decades of delay and a rival camp that favoured the lunar-distance method instead.",{"id":224,"type":50,"variant":225,"title":226,"markdown":227},"model-limit-constant-drift","model_limit","Real clocks do not drift at a constant rate","The worked example above assumes the clock loses exactly 10 seconds every single day, which makes the arithmetic clean but is not how real mechanical clocks behaved at sea. A pendulum or balance-wheel clock's rate changes with **temperature** (metal parts expand and contract) and with the **ship's motion** (rolling and pitching disturb a swinging pendulum badly). A clock might lose 5 seconds on a calm, cool day and 20 on a hot, rough one.\n\nThis is exactly the problem John Harrison's H4 was built to solve: a temperature-compensated balance wheel (so heat and cold roughly cancel out) mounted on gimbals (so the ship's motion does not throw off the mechanism). H4 did not stop drifting altogether. It made the drift small and far more predictable.",{"id":229,"type":176,"title":230,"problem":231,"steps":232,"help":238},"we-h4-same-voyage","How would Harrison's H4 have done on the same voyage?","H4's 1761-62 sea trial lost only **5 seconds** over **81 days**, a rate of about 0.062 seconds a day. Apply that same rate to the 1,082-day Magellan-Elcano voyage.",[233,234,235,236,237],"Rate = 5 ÷ 81 ≈ **0.062 seconds a day**.","Over 1,082 days: 0.062 × 1,082 ≈ **66.8 seconds** total drift.","In minutes: 66.8 ÷ 60 ≈ **1.11 minutes**; in degrees: ÷ 4 ≈ **0.278°**.","In kilometres: 0.278 × 111.3 ≈ **30.9 km**, compared with **5,020 km** for the ordinary clock.","Ratio: 5,020 ÷ 30.9 ≈ **162 times** better, on this voyage length, than the ordinary clock. Still not zero.",{"anotherExample":239},"Try the same two clocks over a 300-day voyage instead: the ordinary clock gives about 1,391 km of error and H4 about 8.6 km, a ratio of about 162 again. The absolute errors shrink with a shorter voyage, but the improvement factor between the two clocks does not.",{"id":241,"type":50,"variant":242,"title":243,"markdown":244},"nuance-ratio-invariant","nuance","Why H4's advantage does not depend on the voyage length","Look again at the two ratios just computed: about **162×** over 1,082 days, and about **162×** over 300 days. They come out the same, and that is not a coincidence.\n\nEach clock's error is drift rate × days, converted by the same fixed steps into kilometres. So the **ratio** of the two clocks' errors is just the ratio of their two drift rates, 10 seconds a day against about 0.062 seconds a day, and the number of days cancels out completely. A longer voyage makes **both** errors bigger by the same multiple, so their ratio stays put. This is why \"H4 is about 162 times better than an ordinary clock\" is a fair thing to say about the *instrument*, without having to add \"...but only on this particular voyage.\"",{"id":246,"type":111,"prompt":247,"options":248,"explanation":257},"pred-ratio-length","Two clocks are compared on a 1,082-day voyage: an ordinary one drifting steadily, and H4. If they were compared instead on a much shorter 50-day voyage, what would happen to H4's advantage over the ordinary clock (how many times more accurate it is)?",[249,251,253,255],{"id":115,"label":250},"It would shrink, because there is less time for H4 to prove itself",{"id":118,"label":252},"It would grow much larger",{"id":121,"label":254},"It would stay about the same, because both clocks' errors shrink by the same multiple",{"id":124,"label":256},"It cannot be predicted without knowing the ship's speed","**c.** Both errors are directly proportional to the number of days, so shortening the voyage shrinks both errors by the same factor and leaves their ratio almost unchanged. What changes with voyage length is the *size* of the error each clock produces, not how many times better one clock is than the other.",{"id":259,"type":50,"variant":260,"title":261,"markdown":262},"mis-chronometer-perfect","misconception","“A marine chronometer is perfectly accurate”","Harrison's H4 was a huge leap, not a magic fix. Over a voyage as long as Magellan and Elcano's, H4's own trial rate would still have produced roughly **30.9 km** of longitude error, about **162 times** smaller than an ordinary clock's **5,020 km**, but not none.\n\nEvery instrument in this topic, the kamal, the astrolabe, the compass, the chronometer, reduces error. None of them removes it. Good navigation after 1760 meant combining a much better clock with careful star and Sun sights and dead reckoning, cross-checking one against another, not trusting any single tool completely.",{"id":264,"type":129,"itemId":265,"prompt":266,"check":267,"hints":271,"feedback":274},"p-longitude-shorter","exploration.deepen-clock-drift-shorter","A different ship's clock drifts at the same 10 seconds a day, but the voyage lasts only 200 days instead of 1,082. About how many degrees of longitude error would build up? Round to the nearest whole degree.",{"kind":268,"answer":269,"tolerance":5,"unit":270},"number",8,"degrees",[272,273],"Total drift in seconds = 10 × 200.","Convert seconds to minutes, then minutes to degrees by dividing by 4.",{"correct":275,"incorrect":276},"**8°**. 10 × 200 = 2,000 seconds = 33.3 minutes ≈ 8.3° of longitude, showing the error grows in direct proportion to how long the clock has been running.","Multiply 10 seconds by 200 days to get total seconds, divide by 60 for minutes, then by 4 for degrees.",{"id":278,"type":56,"title":279,"eyebrow":280,"navLabel":281},"ch03","An edge case: a kamal cord that has stretched","Chapter 03","3 A stretched cord",{"id":283,"type":46,"markdown":284},"kamal-edge-intro","A kamal only works if the knotted cord stays exactly the length it was cut to. Cord, historically a natural fibre, can stretch a little when it gets wet, or simply with age and use. Suppose a navigator never notices, and keeps using the same knot as if the cord were still its original length. What happens to the latitude reading?",{"id":286,"type":176,"title":287,"problem":288,"steps":289,"help":295},"we-kamal-stretch","What a two-centimetre stretch does to a latitude reading","A kamal card is **5 cm** tall. Its cord was originally knotted at **30 cm**, giving a known angle. Over a long, humid voyage the cord stretches to **32 cm**, but the navigator keeps using the same knot, believing it is still 30 cm.",[290,291,292,293,294],"True angle at 30 cm: 2 × arctan((5 ÷ 2) ÷ 30) ≈ **9.5°**. This matches Discover's kamal figure exactly.","Angle now actually being measured, at 32 cm: 2 × arctan((5 ÷ 2) ÷ 32) ≈ **8.9°**.","Difference: 9.5 − 8.9 = **0.6°**, all because the cord grew by 2 cm, about 6.7% of its length.","In distance: 0.6 × 111.3 ≈ **66.8 km** of latitude error, purely from a stretched cord the navigator never inspected.","The direction of the error matters too: a longer cord makes the same star look lower, so the navigator underestimates the angle and believes the ship is further from the pole than it really is.",{"simplerExplanation":296},"A longer cord makes the card look smaller from your eye, so the angle it covers shrinks, even though the star has not moved. The navigator reads a smaller angle and thinks they are closer to the equator than they actually are.",{"id":298,"type":111,"prompt":299,"options":300,"explanation":309},"pred-kamal-stretch","A navigator's kamal cord has quietly stretched, but they keep reading it as before. In which direction will their estimated latitude be wrong?",[301,303,305,307],{"id":115,"label":302},"It will show a higher latitude than the true one",{"id":118,"label":304},"It will show a lower latitude than the true one (closer to the equator than they really are)",{"id":121,"label":306},"It will show no error at all; only the card size matters",{"id":124,"label":308},"The error is completely random and could go either way","**b.** A longer cord holds the card further from the eye, so the same star's height appears to cover a **smaller** angle, about 8.9° instead of the true 9.5°. A smaller measured angle is read as a lower latitude. So an unnoticed stretch quietly makes a navigator believe the ship is closer to the equator than it actually is, an error of roughly 66.8 km in this example.",{"id":311,"type":50,"variant":312,"title":313,"markdown":314},"careful-recalibrate","careful","Why sailors re-checked their instruments against known ports","The fix for this kind of slow, unnoticed drift is not a cleverer formula, it is **routine checking**. A prudent navigator re-measured a kamal's known angles against a familiar port whenever the chance came up, exactly the way you would check a kitchen scale against a known weight. The same principle applies to any measuring instrument today: a ruler that has been left in the sun, a scale that has not been zeroed, a stopwatch that has not been checked against a clock. Small, silent errors compound the longer you go without checking.",{"id":316,"type":56,"title":317,"eyebrow":318,"navLabel":319},"ch04","How historians attempt contested numbers","Chapter 04","4 Contested numbers",{"id":321,"type":46,"markdown":322},"methodology-intro","Discover told you that the 1492 population of the Americas is disputed, and gave you a range. This chapter explains **how** historians and demographers actually arrive at numbers like these, because the method explains why the range is as wide as it is.",{"id":324,"type":65,"title":325,"items":326},"steps-methods","Three ways historians reconstruct a population nobody counted",[327,331,335],{"title":328,"tag":329,"text":330},"Later written records","records","Colonial tax rolls, mission baptism registers and tribute lists, taken from decades after contact, give a floor: at least this many people were still alive and counted then.",{"title":332,"tag":333,"text":334},"Archaeology and land","archaeology","Counting known settlement sites, estimating how many people each site's houses could hold, and working out how much food the surrounding farmland could grow, gives an independent estimate of how many people the land could have supported before contact.",{"title":336,"tag":337,"text":338},"Back-projection from a later count","back-projection","Take a reasonably confident later population, assume a mortality rate for the epidemic decades in between, and calculate backwards to what the earlier population must have been.",{"id":340,"type":176,"title":341,"problem":342,"steps":343,"help":348},"we-backprojection","Why the mortality-rate assumption matters so much","Suppose historians are fairly confident the population of the Americas had fallen to about **6 million** by 1600. Back-project to 1492 using two different assumed mortality rates for the epidemics and wars in between: **90%** and **95%**.",[344,345,346,347],"If 90% died, then 10% = 10% survived. So the 1492 population = 6 million ÷ 0.10 = **60 million**.","If 95% died, then only 5% survived. So the 1492 population = 6 million ÷ 0.05 = **120 million**.","Compare: 120 ÷ 60 = **2.0**. Moving the mortality assumption from 90% to 95%, a change that sounds small, **doubles** the implied 1492 population.","This is exactly why published estimates spread so widely: 60 million sits near the low-to-middle of Discover's cited range; 120 million sits near the top. Both come from the *same* 1600 figure. The disagreement is really a disagreement about the mortality rate, not about arithmetic.",{"anotherExample":349},"A 92% mortality assumption on the same 1600 figure gives 6 ÷ (1 − 0.92) = 75 million, a third answer from a third reasonable-sounding assumption.",{"id":351,"type":50,"variant":260,"title":352,"markdown":353},"mis-range-ignorance","“A wide range means historians don't really know anything”","It is tempting to think that if experts cannot agree between, say, 50 and 112 million, they know nothing. That is wrong. A range still **rules out** almost every possible number: nobody serious argues the 1492 population of the Americas was 1 million, or 500 million. The genuine argument sits inside a much narrower band than \"anything is possible.\"\n\nThe worked example above shows exactly where the disagreement comes from: not sloppiness, but a real, specific, arguable choice, the mortality rate, that different careful historians make differently from the same evidence. Naming *which* assumption drives a disagreement is more useful than either pretending certainty or throwing up your hands.",{"id":355,"type":46,"markdown":356},"mfg-share-methodology","The same pattern shows up in India's disputed share of world manufacturing output. Paul Bairoch's often-quoted estimates put India at about **25%** of world manufacturing output around 1750, falling to about **2%** by 1900, a drop by a factor of about **12.5**.\n\nNobody in 1750 measured \"world manufacturing output\" the way a statistics office would today. These figures are built by economic historians from incomplete customs records, tax assessments and scattered production estimates for different countries, at different levels of detail, from different years, then combined and compared. Change which incomplete records you weight most heavily, or how you estimate output for a country with thin records, and the resulting share shifts. That is why different respected economic historians publish different numbers for the same rough shape of story: a large, skilled Indian textile industry losing world market share over the 1700s and 1800s.",{"id":358,"type":129,"itemId":359,"prompt":360,"check":361,"hints":364,"feedback":367},"p-backprojection","exploration.deepen-backprojection","Using the same 6 million estimate for 1600, what 1492 population would a **92%** mortality assumption imply? Answer in millions, to the nearest whole million.",{"kind":268,"answer":362,"tolerance":5,"unit":363},75,"million",[365,366],"If 92% died, 8% survived.","1492 population = 6 million ÷ 0.08.",{"correct":368,"incorrect":369},"**75 million**. 6 ÷ 0.08 = 75, sitting between the 90% and 95% answers, exactly as you would expect for an assumption between theirs.","Survivors = 100% − 92% = 8%. Divide 6 million by 0.08.",{"id":371,"type":372,"conceptId":373,"relation":374,"explanation":375},"conn-data-handling","connection","data-handling","helps_understand","Back-projecting a population from an assumed mortality rate, and asking how sensitive the answer is to that assumption, is the same reasoning skill as checking how much a mean shifts when one value in a dataset changes.",{"id":377,"type":56,"title":378,"eyebrow":379,"navLabel":380},"ch05","Weighing one account against another","Chapter 05","5 Weighing accounts",{"id":382,"type":46,"markdown":383},"weighing-intro","Discover asked whose account survived. This chapter asks a sharper question: when two surviving accounts disagree, how does a historian decide which to lean on, and by how much?",{"id":385,"type":386,"caption":387,"columns":388,"rows":393},"table-evidence-trust","table","What different kinds of source are good for, and why",[389,390,391,392],"Source type","Best trusted for","Weaker for","Why",[394,399,404,409,414],[395,396,397,398],"Customs ledger or tax roll","Exact quantities: sacks, coins, duty paid","Motives, feelings, the shore's side of a meeting","A private tax record has no audience to impress. Nobody inflates a ledger nobody else was meant to read.",[400,401,402,403],"Traveller's journal or letter home","The order of events, personal impressions, what the writer noticed","Precise quantities; the writer often rounds, boasts or forgets","Written to be read, sometimes by a patron or a king, so it can shade towards the impressive.",[405,406,407,408],"Same-day ship's log","Dates, courses, weather, distances","Interpretation of what things meant","Recorded hour by hour while memory is fresh, but by a single side, for a practical purpose.",[410,411,412,413],"A traveller's account dictated decades later","The overall shape and range of a life's journeys","Exact dates, exact figures, precise sequence","Memory reorganises and compresses over decades, even when the broad picture stays genuine.",[415,416,417,418],"Archaeology (a wreck, a site)","Physical facts: what was actually carried, built or used","Motives, dates without other clues, anything the object cannot show","Objects cannot lie about their own existence, but they can only speak to what happened to survive.",{"id":420,"type":50,"variant":421,"title":422,"markdown":423},"ex-archaeology-confirms","example","When a wreck checks a written record","The **Belitung wreck**, found in 1998, carried about **60,000 Chinese ceramic pieces** in a hull built in an Arab, not Chinese, style, physical proof of exactly the mixed Indian Ocean trade that written sources describe, from a period with very few surviving documents of its own.\n\nThe **Mary Rose**, sunk in 1545 and raised in 1982 after **437 years** underwater, and the **Vasa**, sunk in 1628 and raised in 1961 after **333 years**, both let historians check written ship specifications and crew lists against the actual timbers, guns and skeletons recovered. Sometimes archaeology confirms the paper record closely. Sometimes it corrects it: the Vasa's own inquiry blamed poor stability, and the wreck itself showed exactly the top-heavy design the inquiry described.",{"id":425,"type":46,"markdown":426},"rihla-vs-log","Compare two kinds of account directly. Antonio Pigafetta kept a journal through the entire Magellan-Elcano voyage and brought it home himself in 1522, close to a same-day record. Ibn Battuta, by contrast, travelled for about 29 years and only **dictated** his account, the *Rihla*, in **1355**, to a scholar named Ibn Juzayy, years after his journeys had ended.\n\nThat does not make the Rihla worthless, far from it: it is one of the richest surviving descriptions of the fourteenth-century world, from Delhi to the Maldives to China. But a historian reads it differently from a same-day log. Exact sequences of years and some numbers in the Rihla have been shown, by cross-checking against other sources, to be compressed or reordered, the natural result of describing a three-decade life from memory. The overall shape of where he went and what he saw holds up well. The fine detail needs more caution than a log written the same evening.",{"id":428,"type":429,"component":430,"componentVersion":5,"config":431,"objective":503,"textAlternative":504},"lab-map-accounts","interactive","voyage-map",{"voyages":432,"showWinds":502},[433,475],{"id":434,"label":435,"year":436,"from":437,"to":437,"path":438,"why":473,"consequence":474},"magellan","Magellan and Elcano","1519-1522","Sanlucar",[439,442,445,448,451,454,457,460,463,466,469,472],[440,441],36.78,-6.35,[443,444],28.1,-15.41,[446,447],-22.9,-43.2,[449,450],-53.5,-70,[452,453],-10,-140,[455,456],13.44,144.79,[458,459],10.32,123.89,[461,462],0.69,127.4,[464,465],-30,80,[467,468],-34.36,18.47,[470,471],14.93,-23.51,[440,441],"To reach the Spice Islands by sailing west, so Spain could trade in cloves and nutmeg without crossing Portuguese routes.","The first voyage all the way round the world, finished by Elcano after Magellan was killed. It proved the oceans connect, and showed how deadly such a crossing was.",{"id":476,"label":477,"year":478,"from":479,"to":479,"path":480,"why":500,"consequence":501},"ibn-battuta","Ibn Battuta's travels","1325-1354","Tangier",[481,484,487,490,493,496,499],[482,483],35.78,-5.81,[485,486],30.04,31.24,[488,489],21.42,39.83,[491,492],28.61,77.21,[494,495],3.2,73.2,[497,498],24.87,118.68,[482,483],"He set out at 21 on pilgrimage to Mecca and then kept going, working as a judge and an envoy. Curiosity, faith and employment kept him moving for about 29 years.","His book, the Rihla, is one of the richest descriptions we have of the fourteenth-century world, from Delhi to the Maldives to China, written by a traveller, not a conqueror.",false,"Compare a voyage recorded almost day by day with a set of travels only written up decades afterwards.","Two routes on a world map. **Magellan and Elcano, 1519-1522**: Sanlucar, the Canaries, down the coast of South America, through the strait, across the Pacific to Guam and the Philippines, on to the Moluccas, round Africa and home, all recorded close to the day it happened in Pigafetta's journal.\n\n**Ibn Battuta, 1325-1354**: Tangier, Cairo, Mecca, Delhi, the Maldives, Quanzhou and home again, covering about 29 years, but only written up as the Rihla in 1355, after the travelling had finished.\n\nThe point of putting them side by side: both routes are broadly trustworthy for *where* the traveller went. The Magellan-Elcano route is far more trustworthy for the *exact date* of any single point on the line, because it was written down within hours of happening, not reconstructed from memory years later.",{"id":506,"type":429,"component":507,"componentVersion":5,"config":508,"objective":554,"textAlternative":555},"lab-sort-evidence","sort-game",{"prompt":509,"bins":510,"items":520,"seconds":553},"Sort each claim by how much a historian should trust it, and for what.",[511,514,517],{"id":512,"label":513},"quantity","Trust for exact numbers",{"id":515,"label":516},"shape","Trust for general shape only",{"id":518,"label":519},"caution","Treat with real caution",[521,525,529,533,537,541,545,549],{"id":522,"label":523,"bin":512,"why":524},"e1","A Malabar port's customs ledger records 200 sacks of pepper loaded on a named ship","A private tax record with no audience to impress. Excellent for exact quantities.",{"id":526,"label":527,"bin":518,"why":528},"e2","A captain's letter home boasts that his ship carried 500 sacks of the finest pepper","Written to impress a patron. Numbers in letters written for effect are often rounded up.",{"id":530,"label":531,"bin":515,"why":532},"e3","Ibn Battuta's Rihla describes visiting Delhi and later the Maldives, dictated decades afterwards","The overall route and experiences hold up under cross-checking; exact years and some figures do not.",{"id":534,"label":535,"bin":512,"why":536},"e4","Pigafetta's journal records the strait was reached on a specific date in 1520","A same-day log by an eyewitness who survived the whole voyage. Strong for exact dates.",{"id":538,"label":539,"bin":512,"why":540},"e5","A shipwreck's cargo is recovered and counted piece by piece by archaeologists","Physical objects cannot exaggerate their own number. Very strong for exact quantities.",{"id":542,"label":543,"bin":518,"why":544},"e6","A royal chronicle written generations later praises a king's fleet as the greatest ever built","Written long after the fact, for praise rather than record. Treat superlatives with real caution.",{"id":546,"label":547,"bin":515,"why":548},"e7","An oral tradition, written down two centuries later, describes a famine and where people migrated","Likely captures the real shape of events even if exact years or numbers have shifted in retelling.",{"id":550,"label":551,"bin":512,"why":552},"e8","A ship's manifest lists the names and number of crew who signed on before departure","An administrative record made for a practical purpose, not for glory. Reliable for headcounts.",0,"Practise judging which kind of evidence to trust for exact figures and which only for the general shape of events.","Eight claims sort into three bins: **trust for exact numbers** (customs ledgers, same-day logs, archaeology, crew manifests: records with no audience to impress, or physical objects that cannot lie about their own count), **trust for the general shape, not exact numbers** (accounts dictated or written down long after the events, where the overall picture survives cross-checking better than fine detail), and **treat with real caution** (boastful letters and chronicles written to impress a patron or celebrate a ruler).\n\nThe pattern to notice: a source's reliability depends on **who wrote it, for whom, and how soon**, not on how confident or detailed it sounds.",{"id":557,"type":111,"prompt":558,"options":559,"explanation":568},"pred-ledger-vs-journal","A trader's private customs ledger says 200 sacks of pepper were loaded onto his ship. His letter home to his family says 500 sacks of the finest pepper. A historian can only use one figure. Which should it be, and why?",[560,562,564,566],{"id":115,"label":561},"500, because more detail (\"the finest\") makes it more believable",{"id":118,"label":563},"200, because a private tax record has no reason to exaggerate, while a letter home has every reason to impress",{"id":121,"label":565},"Average the two to 350",{"id":124,"label":567},"Neither; without a third source nothing can be said","**b.** The ledger was written for the trader's own tax records, seen by officials, not family, with no benefit to inflating the number. The letter was written to impress people back home, exactly where boasting creeps in. Averaging two sources of very different reliability, as in option c, treats a boast and a tax record as equally trustworthy, which they are not. The honest move is to prefer the source with no incentive to exaggerate, and say so.",{"id":570,"type":56,"title":571,"eyebrow":572,"navLabel":573},"ch06","Shipwrecks: measuring deep time","Chapter 06","6 Deep time",{"id":575,"type":46,"markdown":576},"wrecks-intro","A shipwreck is a kind of clock. The moment it sinks, a piece of one particular day is sealed away; the moment it is found and raised, that sealed record re-enters the world. The gap between those two moments, years spent on the seabed, is itself a fact worth working with.",{"id":578,"type":176,"title":579,"problem":580,"steps":581,"help":587},"we-wreck-stats","Working out the average time a famous wreck spends underwater","Five well-documented wrecks and how many years each spent underwater before being found or raised: Mary Rose 437, Vasa 333, Titanic 73, the San Jose galleon 307, and the Belitung wreck about 1,168. Find the mean, median and range.",[582,583,584,585,586],"List the years underwater in order: 73, 307, 333, 437, 1168.","Mean = sum ÷ count = (437 + 333 + 73 + 307 + 1168) ÷ 5 = 2,318 ÷ 5 = **463.6 years**.","Median (the middle value of 5, sorted) = **333 years**.","Range = largest − smallest = 1,168 − 73 = **1,095 years**.","Notice the mean (463.6) is pulled well above the median (333) by the Belitung wreck's huge 1,168 years. One unusual value can drag a mean a long way; the median barely moves.",{"anotherExample":588},"Drop the Belitung wreck and recompute with just the other four (437, 333, 73, 307): mean = 287.5, much closer to the median of 320, because the extreme value is gone.",{"id":590,"type":429,"component":591,"componentVersion":5,"config":592,"objective":617,"textAlternative":618},"lab-data-wrecks","data-lab",{"datasets":593,"valueRange":603,"step":5,"challenges":605},[594],{"label":595,"values":596,"unit":602},"Years underwater before recovery",[597,598,599,600,601],437,333,73,307,1168,"years",{"min":553,"max":604},1300,[606,610,614],{"measure":607,"target":608,"prompt":609},"mean",350,"Change one wreck's years so the mean comes out near 350.",{"measure":611,"target":612,"prompt":613},"range",500,"Change one value so the range becomes about 500 years.",{"measure":615,"target":600,"prompt":616},"median","Change one value so the median becomes 307 years.","Explore how mean, median and range respond differently when one unusual value is very large.","An editable dot plot holds five values, the years each wreck spent underwater: Mary Rose 437, Vasa 333, Titanic 73, San Jose galleon 307, Belitung wreck 1,168. Live readouts show the mean (463.6), median (333) and range (1,095).\n\nThree challenges ask you to change a single value to hit a target mean, range or median, showing directly how sensitive each measure is to one unusual point, exactly the same lesson the mortality-rate worked example taught with population figures instead of years.",{"id":620,"type":50,"variant":242,"title":621,"markdown":622},"nuance-belitung-date","Even a wreck's own date can be an estimate","Mary Rose (sank 1545), Vasa (1628), the Titanic (1912) and the San Jose galleon (1708) all sank on dates recorded at the time, by witnesses, in official reports. The Belitung wreck's sinking date, used above as about **830 CE**, is different: nobody wrote it down. It is an estimate, mainly from a dated inscription on one bowl in the cargo and the style of the ceramics, refined since the find in 1998.\n\nSo even inside one small dataset, some numbers are counted facts and one is a careful reconstruction with its own margin of error, the exact distinction this whole chapter has been about.",{"id":624,"type":372,"conceptId":373,"relation":625,"explanation":626},"conn-data-wrecks","applied_in","Mean, median and range on a small, real dataset, and noticing how one unusual value pulls the mean but not the median, is data handling doing real historical work.",{"id":628,"type":56,"title":629,"eyebrow":630,"navLabel":631},"ch07","Standing back","Chapter 07","7 Standing back",{"id":633,"type":634,"tone":635,"items":636},"spec-deepen-recap","spec","amber",[637,641,645,649,653],{"label":638,"big":639,"value":640},"Clock drift, long voyage","45.1° \u002F 5,020 km","An ordinary 10 s\u002Fday clock, uncorrected over 1,082 days (Magellan-Elcano).",{"label":642,"big":643,"value":644},"H4, same voyage","0.278° \u002F 30.9 km","Harrison's real trial rate applied to the same voyage length: about 162× better, whatever the length.",{"label":646,"big":647,"value":648},"Kamal cord stretch","0.6° \u002F 66.8 km","A cord growing from 30 cm to 32 cm under the same 5 cm card, unnoticed.",{"label":650,"big":651,"value":652},"Mortality assumption","60-120 million","The 1492 population implied by 90% versus 95% mortality, from the same 6 million in 1600.",{"label":654,"big":655,"value":656},"Wreck years, mean vs median","463.6 vs 333","One extreme value (the Belitung wreck) pulls the mean far more than the median.",{"id":658,"type":659,"prompt":660},"reflect-assumption","reflection","Pick one contested number from this topic (the 1492 population of the Americas, or India's manufacturing share in 1750\u002F1900). Name the single assumption a historian could change that would move the number the most, and explain, in your own words, why that assumption is hard to pin down.",{"id":662,"type":663,"title":664,"terms":665},"gloss-deepen","glossary","New words for this layer",[666,669,673,676,679,682,685],{"term":667,"meaning":668},"Differential heating","Land and water warming and cooling at different rates under the same sunlight; the underlying cause of the monsoon.",{"term":670,"meaning":671,"example":672},"Chronometer","A very accurate clock built to keep time at sea, compensating for temperature and motion.","Harrison's H4 was the first successful marine chronometer.",{"term":674,"meaning":675},"Back-projection","Working backwards from a later, better-known figure using an assumed rate of change, to estimate an earlier, unknown one.",{"term":677,"meaning":678},"Carrying capacity","How many people a given area of farmland could support, used by archaeologists to estimate past populations.",{"term":680,"meaning":681},"Mortality rate","The share of a population that died over a given period, used as the key assumption in back-projecting past populations.",{"term":683,"meaning":684},"Customs ledger","An official record of goods and duty paid, kept for tax purposes rather than to impress anyone.",{"term":686,"meaning":687,"example":688},"Gimbal","A pivoted mounting that keeps an instrument level even as the object carrying it rolls and pitches.","H4 was gimbal-mounted so a ship's rolling would not disturb its mechanism.",{"id":690,"type":429,"component":691,"componentVersion":5,"config":692,"objective":708,"textAlternative":709},"lab-match-deepen-terms","match-pairs",{"prompt":693,"mode":694,"pairs":695},"Match each term from this layer to its meaning.","connect",[696,698,700,702,704,706],{"a":667,"b":697},"Land and sea warming and cooling at different rates",{"a":670,"b":699},"An accurate clock built to survive a ship's motion and heat",{"a":674,"b":701},"Working backwards from a later count using an assumed rate",{"a":677,"b":703},"How many people a stretch of farmland could feed",{"a":683,"b":705},"A private tax record with no audience to impress",{"a":686,"b":707},"A pivoted mount that keeps an instrument level","Check that the new vocabulary from this layer has stuck before the quiz.","Six terms must be connected to their meanings: **differential heating** (land and sea warming and cooling at different rates, the cause of the monsoon), **chronometer** (an accurate clock built to survive a ship's motion and temperature swings), **back-projection** (working backwards from a later, better-known count using an assumed rate of change), **carrying capacity** (how many people a stretch of farmland could support), **customs ledger** (a private tax record with no audience to impress), and **gimbal** (a pivoted mount that keeps an instrument level as a ship rolls).\n\nEvery one of these terms did real work earlier in this layer: they are not decoration, they are the vocabulary the reasoning depended on.",{"id":711,"type":712,"title":713,"questions":714},"quiz-deepen","quiz","Mechanism, method and edge cases",[715,728,741,754,767,780,793,806,819],{"itemId":716,"prompt":717,"options":718,"correct":118,"why":727},"exploration.deepen-q-monsoon","What actually causes the monsoon wind to reverse direction between summer and winter?",[719,721,723,725],{"id":115,"label":720},"The Earth's rotation changes speed",{"id":118,"label":722},"Land heats and cools faster than the ocean, reversing the pressure difference between them",{"id":121,"label":724},"Ships travelling in different directions push the wind",{"id":124,"label":726},"The Moon's gravity pulls the wind seasonally","Land's temperature swings much further than the ocean's through the year, which reverses which side has lower pressure, and wind always flows towards lower pressure.",{"itemId":729,"prompt":730,"options":731,"correct":121,"why":740},"exploration.deepen-q-drift","An ordinary clock drifting 10 seconds a day, uncorrected for the whole 1,082-day Magellan-Elcano voyage, would end up wrong by roughly how much longitude?",[732,734,736,738],{"id":115,"label":733},"About 4.5 km",{"id":118,"label":735},"About 45 km",{"id":121,"label":737},"About 5,020 km",{"id":124,"label":739},"About 500,000 km","10 × 1,082 seconds converts through minutes and degrees to about 5,020 km, roughly an eighth of the way around the world.",{"itemId":742,"prompt":743,"options":744,"correct":118,"why":753},"exploration.deepen-q-h4-perfect","Did Harrison's H4 make longitude-finding perfectly accurate?",[745,747,749,751],{"id":115,"label":746},"Yes, it eliminated all error",{"id":118,"label":748},"No; it was far more accurate than an ordinary clock, but still had roughly 30.9 km of error over a voyage this long",{"id":121,"label":750},"No, it was no better than an ordinary clock",{"id":124,"label":752},"It only worked in the Atlantic, not other oceans","H4 cut the error by a factor of about 162 compared with an ordinary clock on the same voyage, a huge improvement, but not zero error.",{"itemId":755,"prompt":756,"options":757,"correct":118,"why":766},"exploration.deepen-q-kamal","A kamal's cord stretches slightly and the navigator does not notice. What happens to the measured latitude?",[758,760,762,764],{"id":115,"label":759},"It reads higher than the true latitude",{"id":118,"label":761},"It reads lower than the true latitude, closer to the equator than reality",{"id":121,"label":763},"Nothing changes; only the card size matters",{"id":124,"label":765},"The reading becomes completely random","A longer cord makes the star's height look like a smaller angle, and a smaller angle is read as a lower latitude.",{"itemId":768,"prompt":769,"options":770,"correct":121,"why":779},"exploration.deepen-q-range-meaning","What does a wide range of published population estimates for 1492 actually tell you?",[771,773,775,777],{"id":115,"label":772},"That historians know essentially nothing",{"id":118,"label":774},"That the true number could be absolutely anything, including a few thousand",{"id":121,"label":776},"That the argument is really about a specific, arguable assumption, such as an epidemic mortality rate, not about guessing blindly",{"id":124,"label":778},"That the question should be dropped as unanswerable","A range still rules out almost every possible number. The genuine disagreement traces back to specific choices, like the mortality rate used in back-projection.",{"itemId":781,"prompt":782,"options":783,"correct":118,"why":792},"exploration.deepen-q-backproject","Back-projecting from the same 6 million people in 1600, why does assuming 95% mortality give a much bigger 1492 population than assuming 90%?",[784,786,788,790],{"id":115,"label":785},"Because 95% is simply a bigger number than 90%",{"id":118,"label":787},"Because a higher mortality rate means fewer survivors, so the same 6 million survivors implies a much larger starting population",{"id":121,"label":789},"Because the two assumptions use different 1600 figures",{"id":124,"label":791},"There is no real difference between the two answers","With 90% dying, 10% survive, giving 60 million. With 95% dying, only 5% survive, so the same 6 million survivors implies double, 120 million.",{"itemId":794,"prompt":795,"options":796,"correct":118,"why":805},"exploration.deepen-q-ledger","Why do historians usually trust a private customs ledger more than a proud letter home for exact quantities?",[797,799,801,803],{"id":115,"label":798},"Ledgers are always written in better handwriting",{"id":118,"label":800},"A private tax record has no audience to impress, while a letter home is written to look good",{"id":121,"label":802},"Letters are always shorter than ledgers",{"id":124,"label":804},"Ledgers are always older than letters","The ledger was made for the writer's own tax purposes; nobody exaggerates a record only officials will check.",{"itemId":807,"prompt":808,"options":809,"correct":118,"why":818},"exploration.deepen-q-rihla","Why does a historian read Ibn Battuta's Rihla, dictated in 1355, differently from a same-day ship's log?",[810,812,814,816],{"id":115,"label":811},"The Rihla is fictional and should be ignored",{"id":118,"label":813},"It was dictated years after the journeys ended, so its overall shape holds up well but exact dates and figures need more caution",{"id":121,"label":815},"Ship's logs are always more interesting to read",{"id":124,"label":817},"There is no real difference; both should be trusted equally","Memory reorganises detail over decades even when the broad picture stays genuine, unlike a log written the same day.",{"itemId":820,"prompt":821,"options":822,"correct":118,"why":831},"exploration.deepen-q-wrecks","In the five-wreck dataset, why is the mean (463.6 years) so much higher than the median (333 years)?",[823,825,827,829],{"id":115,"label":824},"A calculation mistake was made",{"id":118,"label":826},"The Belitung wreck's unusually large value pulls the mean upward, while the median is unaffected by how extreme one value is",{"id":121,"label":828},"The mean should never be used for this kind of data",{"id":124,"label":830},"Older wrecks are always undercounted","One very large value drags a mean a long way, while a median only looks at the middle-ranked value and barely moves.",{"id":833,"type":834,"title":835,"points":836},"summary-deepen","summary","The short version",[837,838,839,840,841,842,843,844,845],"The **monsoon reverses** because land heats and cools faster than the ocean: hot land pulls in sea air in summer (low pressure), cold land pushes air out in winter (high pressure). One mechanism explains both seasons.","Clock drift **compounds with voyage length**: 10 s\u002Fday over 1,082 days gives about 45.1° (5,020 km) of longitude error, about 12.5% of the way round the world.","Harrison's H4 was far better, about **162 times** more accurate on the same voyage, but still had roughly 30.9 km of error. No instrument in this topic reaches zero error, only smaller error.","A kamal cord that stretches by just 2 cm (30→32 cm) changes the measured angle from 9.5° to 8.9°, misreading latitude by about 66.8 km, always in the direction of appearing closer to the equator.","**Back-projecting** a past population needs an assumed mortality rate. Changing that one assumption from 90% to 95%, applied to the same 6 million people in 1600, **doubles** the implied 1492 figure from 60 to 120 million.","A **wide range is not ignorance**. It rules out almost every possible value and usually traces back to one specific, arguable assumption, not to guessing blindly.","For **exact quantities**, trust records with no audience to impress (customs ledgers, same-day logs, archaeology) over writing meant to impress someone (boastful letters, praise-chronicles, accounts dictated decades later).","**Archaeology can confirm or correct** written history. The Belitung wreck (found 1998) proved mixed Indian Ocean trade with almost no surviving documents; the Mary Rose (437 years down) and Vasa (333 years down) let historians check paper records against real timbers.","In a small dataset, one unusual value (the Belitung wreck's roughly 1,168 years underwater) pulls the **mean** (463.6) well above the **median** (333), which barely moves.",{"id":847,"type":848,"sourceIds":849},"sources-deepen","sources",[850,851,852,853,854],"exploration-rmg-longitude","exploration-wiki-americas-population","exploration-wiki-deindustrialisation","exploration-britannica-ibn-battuta","exploration-wiki-magellan",[850,851,852,853,854],"needs_review",{"generatedBy":858,"notes":859},"claude-code","Draft generated locally; pending owner review. All new numbers (clock drift over a 1,082-day voyage, kamal cord stretch, mortality-rate back-projection, wreck-years statistics) computed and asserted in gen_deepen.py; numbers.py and voyages.py were read but not modified.","2ca1dcc2eee8fa486b970fcec0effd44275919f0bf8951675a5f24b9bd5cac96",{"logic:practice":862,"component:voyage-map@1":863,"component:sort-game@1":864,"component:data-lab@1":865,"component:match-pairs@1":866,"source:exploration-britannica-ibn-battuta":867,"source:exploration-rmg-longitude":868,"source:exploration-wiki-americas-population":869,"source:exploration-wiki-deindustrialisation":870,"source:exploration-wiki-magellan":871},"3d6b0fe1b15255975a32b0fcd94e8019bc959ad45cbf12e136e86149549c6878","39a0c4da8ea3e6637b0b63d8c38cb151ee52e167eae571f75affe7452f56f13a","b164f45a2c8ca08f26c450768ff0231e113e9fe45381eddb34dc6d0548596c38","466896cc37735f48db03875fe9c9ce42fc8bcb7e5f937c9779d70513703b91bd","2a8ee4ac87460b4e1175a4bb13c96b03d577db06dde95670eb7fcfe4ad787899","a13d35e0e7729889a1bcc756f6b9516a3814aae0d7a056ec8b9e8d489972b581","ec4cb21136ecd92e54fb3b49d7a9d333603496550426997b368fbb0b91b3b3ce","0b0e6f85230ee89d0d7dc39590470f0282baad1c9e65aa9e7b0a3c8dfd0c021c","521046b1cb07e0d9b180db6edd4ebc9f50dd2ae6a0f4ffc0e7ce8734b89c5444","f35819f113af55113ffe01fc8ea9e7bdda0703d09c24081ad9cf70dfffc99efa",{"state":873,"reviewer":874,"selfReview":875,"reviewedAt":876,"method":877},"approved","The library owner",true,"2026-09-20T10:18:37.581Z","owner_bulk","preview-7e1cbbcc4f",1789899598077]