Steffen’s own game: his signature and counter in the footer. Author title: Lehmann’s Shadow (her name and the thing the player runs into). Ideas given: “the looks should feel both earthy and scientific, and maybe for one we could have a few animated SVGs.” This binds the look (section 13): a geological section plate in earth colours on a dark ground, and the picture itself is a live SVG, animated where the mechanism moves (the travelling wave, the fan, the drum). It overrode nothing in the precedent rule; the last three games were cream, buff and grey, and all drawn on canvas or in HTML, so a dark earthy SVG plate moves away from all three anyway.
Martina Kölbl-Ebert (2001), “Inge Lehmann’s paper: ‘P′’ (1936)”, Episodes 24(4):262–267 (Classic Paper in the History of Geology). It summarises and quotes, with interpolations in italics, Inge Lehmann (1936), “P′”, Publications du Bureau Central Séismologique International, Série A, 14:87–115. Available: the typeset Episodes PDF only, read from inbox/, six pages. Episodes DOIs follow the pattern 10.18814/epiiugs/<year>/v<vol>i<issue>/<article>; this article’s would be 10.18814/epiiugs/2001/v24i4/007, which could not be verified from this session (the journal’s site and Crossref were not reachable), so the game’s About screen cites the volume and pages and links the journal, not a DOI. Lehmann’s original is not attached; everything below is her model as the commentary reports it, and her own figures as redrawn in the commentary’s Figure 2a and 2b.
Primary: a model, ported. Lehmann’s simplified Earth: radius 6370 km; a core of radius 5/9 of the Earth’s; a constant P-wave speed of 10 km/s in the mantle and 8 km/s in the core, so rays are straight lines inside each shell and bend at the boundary by the law of refraction. From this she computes, by graphical ray tracing, where a ray leaving the epicentre at take-off angle i₀ comes up: a ray grazing the core (i₀ = 33.75°) emerges at 112.5°; the same ray, refracted into the core, at 186.2°; as i₀ falls to about 26° the core rays run back to 153.9° and then forward again to 180° at i₀ = 0. Nothing lands between 112.5° and 153.9°: the shadow zone. Then the three-shell Earth: an inner core of radius 8/10 r₀ sin 16° = 0.2205 r₀ (about 1400 km), with a higher speed, 8.6 km/s. The ray with i₀ = 16° just touches it (emerging at 160.7°); rays with i₀ between 16° and 14.85° are totally reflected off it and come up between 160.7° and 118.8°; rays steeper than 14.85° pass through it and come up from 118.8° out to 180°. So the inner core throws waves into the shadow.
Secondary: an argument, reported. The observations the model has to meet: P-waves beyond about 105° are weak; P′ is observed, with smaller amplitudes, inside the shadow down to about 105° and up to the real focal distance of 143°; diffraction round the core had been the explanation; Lehmann’s P′₃ waves have a large vertical component and grow with distance, which diffraction cannot give. And Part II, four stations near 150° recording the 1929 Buller earthquake, four pulses where two or three were expected. The paper’s judgement on itself: “a hypothesis which seems to hold some probability, although it cannot be proved from the data at hand.” The epilogue: Gutenberg and Richter’s 1200 km inner core two years later; Jeffreys 1939 showing diffraction untenable; twenty years of argument about the speed jump; Birch and Bullen on a solid inner core.
A mechanism made intelligible. How a ball of slow rock in the middle of the Earth bends waves away from a whole band of the far side, and how a smaller ball of faster rock inside it bounces them back into that band. The player operates the mechanism with Lehmann’s own control, the take-off angle, and then with her own hypothesis, the inner ball’s size and speed. There is a counterintuitive corner inside it: a wave that dives only a hair steeper than the one that grazes the core does not land a hair further on but seventy degrees further, on the far side of the shadow, and no angle at all lands in between.
The Earth is a ball. When the ground shakes somewhere, the shaking runs right through the ball and comes out on the far side, where listening posts feel it. But the middle of the Earth is made of slower stuff, and it bends the shaking the way a glass of water bends a straw, so there is a band on the far side where the shaking never arrives. It is a shadow. Posts in the shadow felt a faint shaking anyway. Inge Lehmann drew the paths the shaking takes and found that if there is a harder, faster little ball inside the slow middle, the shaking bounces off it and lands right in the shadow. So the Earth is a ball inside a ball inside a ball, and the little one in the middle was found by a woman in Copenhagen with cardboard cards in oatmeal boxes.
| Position | Format | Feedback device | Chassis | |
|---|---|---|---|---|
| Dear Household | controller (a town’s planner) | challenge with a discovery inside, a side-bet | sightlines between two streets, gardens filling with saplings | Retro: 1970s public-information poster, cream and flat greens, Fredoka, solid blocks and envelopes; light; canvas |
| Me and My Friends | agent, plural (a street’s parents answering the survey) | challenge the player cannot meet | the crossing: friendship lines on a crayon playground, a crayon ruler | Drawn: crayon on buff sugar paper, Gloria Hallelujah and Patrick Hand, wobbly rings; light; canvas |
| One Number, Please | experimenter (the lab’s designer) | discovery closed by a written reply | the subjects’ own screen redrawn, a notebook filling a row per session | Retro: a Windows 2000 z-Tree terminal, dialog grey and navy, Liberation Sans, bevelled buttons; grey |
What this game does differently. Position: the player is the wave, a single ray choosing its dive, and then Lehmann choosing what the wave dives into. The ray is a non-human agent in the paper’s own figures; Me and My Friends had agents too, but human ones answering a survey, and here the control is a physical angle, not a card. Format: discovery through aiming, with a target inside it (reach the posts): the player probes a hidden geometry shot by shot, conjectures where nothing can land, commits to that, and then tests Lehmann’s cure. Not a budgeted allocation, not a survey, not a design switch read into a notebook. Device: new to the table, the fan: every wave the player fires stays drawn, so the player’s shots accumulate into Lehmann’s own Figure 2a, and the shadow appears as the wedge nobody could draw into; with it a ring of posts that light when hit, and a small drum that writes the arrival’s up-and-down. No sightlines, no ruler, no notebook. Chassis: an atlas plate of the Earth’s interior in earth colours on a basalt ground, drawn as live SVG with Fraunces and DM Mono; after three light-or-grey canvases, a dark animated plate; nothing shared with any of the three.
Aim (start screen, top of the play screen, first line of How to play): Reach every listening post on the far side of the Earth with a wave. In the second act the line becomes: Build the inside of the Earth so the posts in the shadow hear the quake too.
Yardstick, in the aim’s units: posts reached, shown as the ring of posts itself, lit or dark, with “N of 18 posts have heard you” under it. In act two, the four shadow posts, lit or dark.
Verdict (the end screen leads with these, in words):
No number is the verdict. The posts-reached count is the yardstick during play; the end screen leads with the three paragraphs above.
A. The Wave (agent, then controller; discovery with a target inside). Controls: one dial, how steeply the wave dives (0° straight down, 90° along the ground); FIRE. In act two, two dials, the inner ball’s size and its speed against the core’s, and QUAKE, which fires at every angle at once. Deliberation: “my 40° shot came up at 100°, my 30° shot at 174°; the post at 135° should be somewhere between; 35°? 33°?” and later “the core bends waves too far round; I need something that bends them back, or bounces them.” After the press: the wave travels, straight in the mantle, kinks at the core, comes up somewhere on the rim; a post lights or does not; the shot stays drawn. Discovery moment: the hair’s breadth between 33.75° and 34°, a jump of seventy degrees with nothing in between; then the bounce off the faster ball landing exactly in the gap. Fun: anticipation (where will this one come up?), agency (the mapping is learnable and strange), discovery (the wedge appears in the player’s own drawing), mastery (the second time through, every reachable post in six shots, and the ball built in one quake). Replay: try a slower ball on purpose and watch waves go past the centre and round; try the smallest ball that still fills the shadow.
B. The Oatmeal Boxes (experimenter; discovery). Lehmann at her garden table with cards from stations at 149°, 150°, 150.4°, 150.6°: the player reads the four records, marks the pulses on each, and must decide whether they show two waves, three or four, with the Buller earthquake’s foreshocks complicating the count. Deliberation: which wiggle is a pulse. Discovery: four lines through the points means four waves, one more than two shells allow. Fun: discovery, some agency. Why it fails the walkthrough: a stranger cannot read a seismogram, the records are not in the paper beyond one figure, and every decision is “is that a pulse?”, which is a guess dressed as a reading. The paper’s own Part II is honest that the data were insufficient.
C. The Earth (subject; the reckless one: wordless, real time). The player is the Earth. Quakes go off round the rim every few seconds and fire their fans live; the player drags the inner ball bigger or smaller with a thumb and presses to harden or soften it, and the ring of posts glows where waves land, with the four shadow posts marked. No words on the play screen at all. Deliberation: continuous tuning by eye. Discovery: the moment the shadow posts glow. Fun: agency, discovery, a toy-like pleasure. Replay: none beyond fiddling. Walkthrough: lovely for a minute; but the stranger never learns why there is a shadow, because they never fired a single wave and watched it bend, and the jump at 33.75° is lost in the fan. Its gift, the whole fan firing at once with the player’s ball in the middle, is taken into A’s second act.
A, the pick. 0:00. A dark plate: a cut-open Earth in ochre and rust, a gold wave running through it on a loop and coming up on the far side where a small post lights. Lehmann’s Shadow. Three lines: “Turn a dial to choose how steeply your wave dives. Fire it, and watch it bend. Then put a ball inside the Earth and fire again.” The aim. BEGIN. 0:20. The play screen: the Earth, big, the quake at the top as a small burst, eighteen posts round the right-hand rim with degrees every thirty. A line says what this is: the mantle (ochre) where waves run fast, the core (rust) where they run slower and so get bent toward the middle, and a rule: a wave bends only where the colour changes. A dial: Dive: 45°. As the thumb moves it, a dotted line shows the first straight leg of the wave down to where it would meet the core. FIRE. 0:35. The wave runs down the dotted line, kinks at the rust, runs across, kinks again, comes up at 161°; the post at 165° stays dark, the post at 155° too; a line says “Dived at 45°. Came up at 161°, between two posts. Nobody heard it.” The shot stays as a thin gold thread. The stranger frowns, nudges to 47°, fires: 157°, the post at 155° lights with a small drum-trace; “1 of 18 posts have heard you.” A grin. 1:10. They try 60°: the dotted leg now misses the core entirely; the wave goes straight across the mantle and comes up at 60°, lighting the post at 65°? No, 60° is between 55° and 65°; dark. They learn the mantle rule: comes up at 180 minus twice the dive. Four quick shots light 25°, 45°, 85°, 105°. 2:00. The fan is forming. The posts at 115°, 125°, 135° and 145° are dark. They aim for 135°: a dive of 22.5° would give it if the straight rule held; the wave kinks and comes up at 158°. Hm. 30°: 170°. 34°: 112°. 33.5°? The wave grazes the rust and comes up at 184°, i.e. 176° on the other side. The stranger stares: 34° gives 112°, 33.5° gives 176°, nothing in between? They try 33.75 exactly, then 26°, which comes up at 154°, the nearest anyone gets. 3:00. The button That’s every post I can reach has been sitting there; they press it. The screen fires the whole fan, Lehmann’s figure, and the empty wedge from 112° to 154° is hatched and named: “the shadow: no angle reaches it.” The four dark posts are named. Then the twist card: “In 1936 the posts in the shadow had heard something: faint, and mostly up-and-down, as if it had come up steeply. Lehmann had an idea.” 3:40. Act two: the same Earth, and two dials: the ball’s size, and its speed against the core’s: slower, same, faster. The rule: into slower rock a wave bends toward the middle; into faster rock it bends away, and if it comes in at a shallow slant it bounces off. As the thumb moves, a pale ball grows in the middle of the rust. “You have four quakes.” The stranger thinks: the core bends waves too far round, past the shadow; something faster would bend them back. Faster, medium. QUAKE. The fan fires; several rays glance off the pale ball and come up in the wedge; the four posts light one after another; the drum writes a steep up-and-down. “All four shadow posts heard the quake.” 4:30. They try slower, out of curiosity: rays plunge through the ball and go past the centre; nothing in the wedge but the last post. Then a tiny faster ball: three of four. Last quake: faster, Lehmann’s size. 5:10. Verdict: what they did, what she said, what it meant; her ghost ball drawn over theirs.
C, the runner-up. 0:00. A dark plate, no words. Quakes flash round the rim; fans fire; posts glow. A pale ball in the middle with a handle. 0:30. The stranger drags it bigger; the fans redraw; some posts change. They press it; it hardens, goes bone-white; four posts that were dark glow. 1:00. “Did I do that?” They drag it smaller; they go dark again. 1:30. They understand that the ball matters, and that hard and big fills the band; they do not know what the band is or why it was dark. 2:00. They put the phone down, pleased and none the wiser about the shadow. Fun for ninety seconds; the discovery is the plot, and it did not land.
A, The Wave. The player is inside Lehmann’s figure, doing what she did: choosing a take-off angle and seeing where the ray comes up, then choosing an inner core and seeing whether the shadow posts hear. The deliberation is real and gets sharper (the mapping from dive to landing is learnable, and then it breaks), the discovery lands twice inside play (the jump, then the bounce), and the sentence afterwards has no “guess”: I fired waves at every angle and none could land in a band on the far side; then I put a faster ball inside the core and the waves bounced off it into the band. Originality: no game in the table has had the player aim a thing and watch it travel and bend, nor built one of the paper’s figures from the player’s own shots. Calibration is a gift: the model is tiny and every number in the paper comes out of it to the decimal.
Author test. The author of the commentary, and anyone who teaches this, would recognise Figure 2a forming shot by shot, the 33.75° grazing ray, the 153.9° caustic and the 118.8° reflected ray, and would enjoy the one thing a lecture cannot give: a stranger discovering the shadow by trying to shoot into it.
Runner-up: C, The Earth. Switch if the author wants a silent sixty-second piece for a lecture slide, accepting that the shadow is then shown, not found.
What the reckless concept would have given. Live play with the whole fan and a ball you shape by hand. Kept: the second act fires the whole fan at once, and the ball grows in the picture as the dial moves. Given up: real time, and silence.
Two acts, one long round each, four screens: start; act one (fire waves into the two-shell Earth, as many shots as the player likes, ended by the commitment button That’s every post I can reach, which fires the full fan and names the shadow, then the twist card); act two (the inner ball, four quakes, the fourth final, or an early That’s my Earth); verdict. No presets: the two regimes are the paper’s two figures, 2a then 2b, in the paper’s order because that is the order of the discovery. No score. The yardstick is the ring of posts. Target length five to seven minutes; a player who knows the answer takes three.
Register: a geological section plate from an atlas, drawn live (Instrument family, with Austere’s atlas plate as the picture’s manner; by the Given, earthy and scientific). From the paper’s own object: the cut-open Earth of Figure 2, which is already a plate. Earth colours for the earth; one bright colour, gold, for the waves and for nothing else; a dark basalt ground so the plate glows like a lit section in a museum case; fine ruled degree ticks round the rim and small-cap labels, as a plate has.
#1B1714 (the page); mantle ochre #B07A3C with a darker rim #8A5A2B; core rust #7A3024 with a lighter heart #9A4432; bone #EEE3C8 (the inner ball when faster, every word, the ticks); seismic gold #F2C14E (every wave, every lit post, the drum’s trace, the one primary button); slate #5C6670 (unlit posts, the shadow’s hatching, muted text). The inner ball’s colour runs from peat #3A2A20 (much slower) through rust (same as the core) to bone (much faster).Lehmann’s energy and amplitude calculations (section 2–3 of her paper) beyond the drum’s up-and-down and the twist card’s words; the modern Earth with curved rays (the commentary’s inset) beyond a line in About; the real shadow’s limits (105° and 143°) beyond the twist card; the S-wave discussion; the whole of Part II (the Buller records and the four pulses) beyond the verdict’s epilogue; the biography beyond the oatmeal boxes in the core narrative and one line on the verdict; diffraction as the rival explanation beyond one clause.
Ported. The game runs Lehmann’s simplified Earth as the commentary reports it: straight rays, one constant P-wave speed per shell (10, 8, and for the inner ball 8.6 km/s), the law of refraction at each boundary via the conserved perpendicular distance r·sin i, total reflection where the sine would exceed one, and Δ summed from the chord geometry. No randomness anywhere.
Where the discovery comes from. The shadow and its filling follow from the model’s geometry, not from our rules. The calibration check varies the ear (1°–4°) and shifts the posts by five degrees: the posts called unreachable always lie inside 112.5°–153.9°, and every shadow post that lies where her inner-core rays land is lit by her ball. With posts at 115° the sliver 112.5°–118.8° that her own model leaves dark shows up as a dark post under her ball too; the posts were placed at 10°, 20°, …, 180° so that the model’s own story is clean, and the memo and About say so. The “faster beats slower” lesson is the model’s: a ball a little slower (0.9) leaves the first three shadow posts dark; one a quarter slower fills the band from the far side, which the verdict names rather than hides.
Checked headlessly against the paper’s own numbers (calibration.js, 41 checks, result in calibration-result.txt): core radius 3539 km; the grazing ray at 33.75° coming up at 112.5°, and refracted at 186.2° (folding to 173.8°); the core rays’ minimum 153.9° at about 26°; the straight-down ray at 180°; nothing between 112.5° and 153.9°; the mantle chord rule 180 − 2·i₀; the inner radius 0.2205 r₀ = 1405 km; the 16° ray touching the ball and coming up at 160.7°; total reflection from 14.85° coming up at 118.8°, where the reflected and transmitted branches join; the three-shell coverage (reflect 118.8°–160.7°, through 118.8°–180°); the arrival angle equal to the dive on every branch; the posts’ shadow exactly {120, 130, 140, 150} and Lehmann’s ball lighting all 18; the robustness runs above; the words; a scripted game to a verdict. All 41 passed.
No score, so the table is of the second act’s outcomes: shadow posts lit by the ball’s radius and speed (from calibration-result.txt):
r1\v 0.75 0.8 0.85 0.9 0.95 1 1.05 1.075 1.1 1.15 1.25
0.1 1234 1234 234 34 4 34 234 1234 1234 1234
0.2 1234 234 34 4 234 1234 1234 1234 1234
0.2205 1234 234 34 4 234 1234 1234 1234 1234
0.3 1234 234 34 4 1234 1234 1234 1234 1234
0.4 1234 234 34 4 1234 1234 1234 1234 1234
0.5 234 34 4 4 4 34 1234 1234
Read: the same speed as the core does nothing; a little faster (1.05–1.1) fills the band for any ball between a fifth and two fifths of the Earth’s radius, and the smallest balls need to be faster still; a little slower lights only the far end; much slower (0.75–0.8) fills it from the far side. So the decisions matter in both dials, the sensible move (faster) works, the half-measure (same) fails, and the perverse move (much slower) succeeds for a reason the verdict explains. The first act has no table: the mapping dive → landing is the paper’s Figure 2a, and the headless loop (loop-log.txt) shows the jump at 33.75° and the fourteen reachable posts.
loop.js): the printout read as the stranger: 34° at 112°, 33½° at 175°, nothing between; the reveal naming the four posts; the slower ball lighting only 150°, the faster one all four.session-test.js): eighteen shots including the jump, the reveal, the twist, four quakes (slower, faster, tiny, Lehmann’s), the verdict, About, a retry, and the early-finish path; help opened mid-play and closed; no errors, no horizontal scroll. Screenshots in the folder. Session length is an estimate: five to seven minutes for a stranger, three for someone who knows the answer.Steffen, after playing the first build: the look is right, the wave as the player is the game, but the play needs simplifying, and after the first wave the player is not taken straight to the second. Changes:
WAVES), counted on the status line (Wave 3 of 12 · 2 of 18 posts lit). After the twelfth the whole fan fires and the shadow is named with no button to find; one Next leads to the twist card. The quiet commitment button is gone. The aim is now Light as many listening posts as you can with twelve waves, in the yardstick’s units.quake(game, speedRatio); a size can still be passed for an author who wants the dial back). Three quakes (QUAKES), the third final; the gold button then reads See the verdict; a quiet See the verdict is there after the first quake for anyone done early.Steffen’s second play: the look and the wave-as-player hold, but nearly every line was too cryptic (“Post 170° heard it. Up-and-down: strong.”; “The nearest post, 150°, is 4½° away”; the reveal; “set the speed in the ball”; the act-two results; the verdict), there was no cue that the next wave was ready, and dives in the decisive span between 33° and 34° were impossible to set. Changes:
DIVE_STEP 0.1, DIVE_SLIDER_STEP 0.5), and the first intro line says so. When a dive within a degree of the last one lands fifteen degrees or more away (JUMP_DEG), the game says so: Only 0.5° away from your last dive, and you came up 63° further round. Something big happens between 33.5° and 34°. Try dives in between: − and + move a tenth of a degree. When it lands in nearly the same place (SAME_DEG): Almost the same dive as last time, and almost the same place. Nothing new here: try a very different dive. This is the one place the game comments on the player’s choice, at the author’s request.Steffen’s third play. The start screen now says, in his words, that you are the earthquake, a wave sent through the ball, and that your listening posts record what comes out on the other side; and that you choose the direction (straight through the middle, through the core, just touching it, or past it) and learn from what the posts record how the Earth is made up, down to its inner core. The verdict is longer and written for a teenager: What you did retells the whole game (two layers, where waves that miss and waves that catch the core come out, the band that never gets a wave, the ball you built); What Lehmann found tells the story of the shadow posts, the creeping-wave explanation and why it fails, her ball and its bounce, and her own caution; What happened next gives the 1200 km, 1939, the twenty-year argument about the speed, 1946 and the picture we still have, and the garden in Copenhagen. Calibration and the browser session re-run; all clean.
Steffen’s fourth round, after a discussion and with a scan of Lehmann’s complete original (30 pages) in hand. His points: the frame “you are the earthquake” is the game and stays with everything that comes with it; his play was all about whether and how the wave bends, and the listening posts played no part in it, which he called a problem; the slower/same/faster choice for the ball was to go; that there must be a second core inside the core was not spelled out; the audience is a smart teenager. Brief: the best version from all of that.
What the original changed. The first three builds had only the Episodes abridgement. The original confirms the model to the decimal and adds three things the game now uses. (a) Her Table II (p. 98): rays through the inner core, a take-off angle and a distance each; the mechanism reproduces every row to the tenth of a degree, now a check. (b) Section 7 and Figure 3 (pp. 95–96): the evidence is records of the Buller earthquake of 16 June 1929 at Irkutsk (110.8°), Sverdlovsk (135.0°) and Baku (137.4°), small sideways and considerable up-and-down, growing with distance. (c) Section 9 (pp. 98–100): the totally reflected ray is “of quite small intensity”; it is the ray through the inner core that is steep, grows with distance and has the characteristics observed. Revision 3’s verdict credited the bounce with that; corrected. Also: the paper never says the inner core is solid, and closes by asking others to look (p. 115). Revision 3 said everyone agreed on a solid ball by 1946; the verdict now says Lehmann never claimed it, Birch proposed it in 1940, and it was shown in 1971 (Dziewonski and Gilbert, Nature 234:465).
The new shape. Two acts with the same control, the dive, and one card between them.
TARGET_POST), and sets a job: hit it with the six waves left. Each wave then says how far short of or past the ringed post it came up. No dive can do it; the player runs into the jump at 33.75° trying. The fan then opens with “You could not hit the post at 130°. Nobody can.” This gives the posts a part (a target) and makes the shadow the player’s own failure, not an announcement.QUAKES’ place in play are gone. quake() stays in the mechanism as the author’s tool behind the size-by-speed table.Why hidden. Asked “is there a second core?”, the old second act answered by handing the ball over and asking for its speed. Now the player finds it the only way a wave can: by its effect on the path. Claim boundary, stated in About and in the verdict’s first sentences on Lehmann: she could not see paths, only what the posts wrote, and inferred the ball; the game lets the player see what she had to reason to.
Layout. The dial and FIRE now sit directly under the picture and drum, with a one-line headline between them (“You came up at 175°: 45° past the ringed post.”; “Something bent you where no colour changed.”); the full sentences follow below the button. Picture, headline, dial and button fit one 380 × 760 screen, so a wave can be fired and watched without scrolling.
Words. Verdict rewritten for a teenager around the new play: what you did; what Lehmann found (she could not ride a wave; the records; her doubt; the three-layer drawing; the faint bounce and the steep transmitted waves; a hypothesis, not a proof); what happened next (1200 km, about 1220 km today, 1939, solidity 1940 and 1971, cardboard cards in oatmeal boxes). Fates: You found the ball. / You touched the ball. / The ball stayed hidden. The quoted line is cut to a phrase.
Checks. Calibration: 55 checks, all passed (new: Table II; the job set after the sixth wave and impossible; the act-two script lighting 150°, 120°, 140°, 130°; early end; the hidden-ball path; the verdict’s dates). Browser session at 380 × 760: twelve waves with the job and the jump, the records card with three posts, eight waves in act two with marks, the second reveal, the verdict, About, a retry, and a second game in which the player never dives steeply and the ball stays hidden; no errors, no horizontal scroll. Not played by strangers.
Divergences added. The job at 130° and its six waves; the records’ trace sizes; the marks; the early end of act two; in the game a bounced wave lights a post as well as a transmitted one, where the paper calls the bounce faint (said in About and the verdict).
Steffen, after playing revision 4 (“this is super now”), with three notes. (1) After a wave’s feedback nothing on the instruments invites the next wave. (2) “Hit the post at 130°” is too much on the nose, and since it cannot be done it is hard to understand and frustrating. (3) The second act read as a repetition of the first twelve rounds; he had expected the game to move straight to the idea that there might be a core inside the core. Changes:
FREE_WAVES and TARGET_POST removed).WAVES2) to send at it; the first lines say that only a dive steeper than 16° reaches it, and a wave that misses is told the same. Results name what the ball did: bent you away from the middle, or bounced you off, and where the same dive came up in the first Earth. The act ends when the four shadow posts are lit or after the eighth wave; then the fan fires through three shells.A rebuild after a long discussion in which Steffen worked out for himself what the paper achieves, and found that the game had never said it: the paper sets what the posts recorded against what an Earth with one core predicts, and shows that an Earth with a core within the core fits. His decisions: the game is about two possibilities, one core or a core within the core, “that’s all” (no talk of a true Earth, no rival explanations); the leading question is flagged from the first screen (if your waves can reach all the posts, one core is right; if not, there might be a core within the core); new title, his, Find Your Inner Core with Lehmann; the cue for the next wave much bigger; aiming on the picture; eight waves and four; a verdict a smart sixteen-year-old can follow. The previous build is in previous/lehmanns-shadow-rev5.html; lehmanns-shadow.html now redirects to find-your-inner-core-with-lehmann.html.
quake() and the speed range (the size-by-speed table in the check uses heardByQuake directly).Steffen after playing revision 6 (“we are very very close now”) and four more rounds of discussion. His notes: only the southernmost faint post could be reached in stage two; the quiz before the end was superfluous; and, more important, the faint posts were flagged with a broken ring from the start as a permanent label, which felt wrong: what is it that they felt, and when? He proposed that the stations should visibly answer (those that feel a strong shake light up, those that feel nothing stay faint, those that feel a small quiver blink). A bundle of waves per go was discussed and rejected by him as too complicated: the game is after the broad gist, that someone learns there must be a core within the core, and may take liberties if they are flagged in an optional read-further part of the debrief. Previous build: previous/find-your-inner-core-rev6.html.
The seven beats:
fire2): any wave that comes up inside the shadow makes the nearest ~ post quiver. Aim: Can you make the ~ posts quiver now?record()). It matches.Also: the quiz and its two buttons are gone; the broken rings are gone; the cue reads Now choose your next direction. and the button Fire · go 3 of 8; the degree labels at 120° and 150° made way for the ~ marks. Fates: You found the inner core. (a ~ post made to quiver) / The inner core was there all along.
Calibration: 53 checks, all passed. Browser session at 380 × 760 with touch: the watched earthquake (14 lit, 4 quivering, 4 marks), drag-to-aim, eight goes with the jump, the first fan (14 lit, 4 dead), the card, four goes (two into the shadow, two posts quivering), the second fan (14 lit, 4 quivering), the verdict, Read further opened (six notes), About, a retry from the watched earthquake, and a game with no steep dive; no errors, no horizontal scroll. Not played by strangers.
Steffen after revision 7, in a long exchange. His objection, and the reason the game had kept slipping: “we can only ever record data from real experiments”, yet the stages showed posts doing different things in different Earths without saying these were not experiments. He restated the task himself: ground the game in the puzzle the real data pose, and mark the two stages as counterfactuals. His outline, built here: stage 0 with no player action, just what the stations recorded, in a few steps, with the quiver band highlighted; stage 1, “now imagine you are an earthquake and the Earth has just one core”, at most four rounds with some nudging toward the band, then “you could not make the band quiver … what is wrong? one last experiment”, then the revelation that with one core it cannot be done: “it was a trap, sorry”; stage 2, the second counterfactual, three rounds, in which the right direction makes the band quiver; a verdict that depends a little on the last stage. He confirmed three real earthquakes for stage 0, the trap line as written, and that “every direction at once” is an animation of what could have been, never a go. Previous build: previous/find-your-inner-core-rev7.html.
WAVES 4 plus LAST_TRIES 1. Posts start dark; the post where the player comes up lights. Every go’s headline says how far short of or past the band it came up. After the first go the aim becomes Can you make the band quiver? After the fourth: You could not make the band quiver. and the cue What is wrong? One last try. (the button reads Fire · one last try). After the fifth, every direction at once, the answering animation with the band silent, and Sorry, that was a trap. With one core it cannot be done. … So the Earth cannot have just one core.WAVES2 3. The liberty of revision 7 stays (a wave anywhere in the band makes the nearest post there quiver). A go that leaves the band silent ends with Try a dive between 12° and 16°. Then every direction at once: the band quivers, the record exactly.Calibration: 53 checks, all passed. Browser session at 380 × 760 with touch: three earthquakes watched (headlines 1, 2, 3 of 3), the band marked only afterwards, four goes with the nudge and the jump, the question and the last try, the trap, the card, three goes ending with the band quivering, the second fan, the verdict, Read further (seven notes), About, a retry from the facts, and a game with no steep dive ending on the other verdict; no errors, no horizontal scroll. Not played by strangers.
Steffen after revision 8: the facts stage ran too fast and it was hard to see that it was about three events; the big question mark was dull and did not prepare for what follows (his idea: an animation that switches slowly between the one-core and the two-core picture); and a go that made no station notice anything sat oddly with a nudge about the band (his idea: one station should always answer). Previous build: previous/find-your-inner-core-rev8.html.
nearestOutside, ours, said in Read further and About). A wave that comes up outside the band lights the nearest post outside the band; inside the band the nearest band post quivers, as before. The three-degree ear still decides what “every direction at once” reaches and where the band is. The “No post is there” lines are gone.Steffen: in the facts stage the posts light up “but sadly we don’t see any waves going through the core”. Now each of the three earthquakes sends its waves visibly through the cut-open Earth: gold lines grow from the quake along their paths at one speed, bend at the core, and each post lights as the first wave reaches it (quakeRun). Only the paths on which the two imagined Earths agree are drawn (agreedRays: every direction that passes beside where an inner core would be, dives above 16°); the steepest ones, where the two pictures differ, are left undrawn while the inner core fades in and out. No drawn wave comes up in the band (checked), and the band’s four posts quiver last, with nothing visibly reaching them. The closing lines point at that: Look at the gold lines: not one of them leads into it. And yet … the posts in it feel a faint quiver. The expanding ring is gone. Calibration: 55 checks, all passed; browser session re-run, no errors; stage 0 takes about 33 seconds.
Steffen, a quick fix “important for human players”: Earthquake x of 3 belongs at the top of the screen, and, optionally, the stage should slow further with a button to press before each earthquake. Both done. In stage 0 the line at the top of the screen reads Earthquake 1 of 3 in large gold type while that earthquake runs; nothing runs until the player presses Show earthquake 1 of 3 (then 2, then 3); after each, a line says what happened and that its row is kept. After the third the top line reads Three earthquakes, the same thing three times. and the button becomes Next. Previous build: previous/find-your-inner-core-rev10.html. Calibration unchanged (55 checks, all passed); browser session re-run with the three presses, no errors.
Steffen: in the facts stage the band’s posts blink “but it’s not quite clear why”; should much fainter rays reach them? Drawing whole rays would give the answer away and show something nobody knew in 1936, so the band gets only the last stretch: as each band post quivers, a short dotted stub comes up into it from below, with a question mark at its lower end (addStub). The line after each earthquake says a faint wave reached them by a path nobody knew, and the closing text points at the stubs. The stubs belong to stage 0 only. Also, as offered and left to me (“just do your best”): “every direction at once” in both imagined Earths now uses the same running-waves animation as the facts stage (quakeRun): the waves travel, and each post lights or quivers as the first wave reaches it. With one core nothing reaches the band and nothing answers there; with a core within the core the waves from the inner core arrive in the band and its posts quiver, and the text says these are the paths the question marks were asking for. Previous build: previous/find-your-inner-core-rev11.html. Calibration unchanged (55 checks, all passed); browser session re-run, no errors.
Steffen: the colours of a lit post and a quivering one were far too close for the human eye. A quivering post is now slate grey with a pale rim (#8E99A3, rim #DDE3E8) against the gold of a lit one, on the picture, in the legend and in the three kept rows.