Each room of the Delta lab against results already known from the real world. The known answer is written down before each run is built; the run is not changed after it is seen, except to fix a fault (written in the fails list); every match and every miss stays here.
Why: a company sending an experiment up to low Earth orbit could have it run in the lab first, the prediction
stamped before launch, and compare after the flight. These records are what such a prediction stands on. The same
list, with the code for each run, is in the repository (lang/lab/TRACK.md).
| # | Known result (what really happens) | Lab | Lab's answer | Match? | Run |
|---|---|---|---|---|---|
| 1 | Wing nut / T-handle spun about its middle axis turns over and over (Dzhanibekov, 1985 in orbit; also on Earth, Mike at home) | space room | easiest and hardest axes steady; middle turned over 10 times in 200,000 steps | MATCH | runs/wingnut.delta |
| 2 | A candle flame with no gravity: round (a sphere), dimmer, blue (little or no yellow soot), and it may go out as the air round it is used up (shuttle, ISS and drop tower candle tests) | space room | no gravity: round, 32 x 31 mm (with gravity 81 x 23); burns about 3% as much; the same share burned short of oxygen (363 v 348 thousandths); still burning at 9 s | ROUND: MATCH. DIMMER: MATCH. BLUE: MISS (the soot rule knows nothing of cooling yet). GOES OUT: not run long enough | runs/candlespace.delta |
| 3 | Crystals grown in orbit come out larger and more perfect than on the ground (no convection, no settling) | space room | salt round a seed, 5 runs each: in orbit 78% full, nearly square (long/short 1.07); on Earth (brine moving down) 65% full, a lopsided column (2.28) growing toward the incoming brine; no overlap | MORE PERFECT: MATCH. LARGER: not tested yet | runs/saltplay.delta (2nd input: drift) |
| 4 | Liquid in free fall pulls itself into a ball; a floating ball of water wobbles when touched (ISS water demos) | space room | from a square: nearly round (1.16 edge to edge from round, the square 1.41), wobbled 23 times when poked and came back round; with gravity on a floor it slumps (1.4 wide) and a poke spreads it to a puddle (4 wide) | BALL: MATCH. WOBBLE: MATCH | runs/waterball.delta |
| 5 | The space station's orbit, about 400 km up: once round in about 92-93 minutes, at about 7.7 km a second (written 21 Nov 2026 ~13:32, before any run) | neither rising nor falling at 7.675 km/s; once round in 92.7 minutes, stepped and counted | MATCH | runs/knowns.delta | |
| 6 | Escape speed from the ground: about 11.2 km a second (written 21 Nov 2026 ~13:32, before any run) | 11.01 km/s (the slowest still going out after 2 days: a short watch counts near-misses as escapes) | CLOSE (2% low) | runs/knowns.delta |
The known answers here were written down before any run for this room. A column of air built up 10 m at a time
from measured constants only (gravity, the weight of air and water molecules, water’s heat of boiling, air’s
warmth per kg); every quantity stepped, no curves. Code: lang/lab/runs/src/atmos-src.delta.
| # | Known result (what really happens) | Lab's answer | Match? | Run |
|---|---|---|---|---|
| 1 | Air pressure falls by about half every 5.5 km up | air full of water vapour: half at 5,530 m (50,841 Pa at 5.5 km; real 50,540); dry air: half at 5,300 m | MATCH | runs/atmos.delta |
| 2 | Air cools about 6.5 C for every km up (on average, in the lowest part of the sky) | full of water vapour: 6.66 C a km on average over 9 km (4.8 in the first km); dry air 9.7 C a km | MATCH (the moist column; dry air alone misses: condensing water is what brings it to 6.5) | runs/atmos.delta |
| 3 | Water boils at about 70 C on top of Everest (8.8 km) | 70.1 C (moist column; 68.1 in the colder dry column) | MATCH | runs/atmos.delta |
| 4 | A sealed balloon grows as it rises, about doubling by 5.5 km | 1.77 times | CLOSE: the written "about doubling" was loose; the gas law gives about 1.75 | runs/atmos.delta |
| 5 | A weather balloon (about 2 m across at launch, rubber) swells as it rises and bursts at about 30-35 km up, about 6-8 m across, where the air is about 1% of the ground's pressure (Mike, 11:03: "rising balloon expands as it reaches the upper atmosphere then pops"; written before the run) | burst at 30,580 m, 8.63 m across, at 964 Pa (0.95% of the ground); with gravity from the Earth's mass at each height (Mike), 30,700 m at 965 Pa; above 11 km the air's warmth is fixed from measurements: MAGIC (fixed temps) | MATCH | runs/atmos.delta |
| 6 | The weather balloon's shape as it rises, and how it ends (Mike gave two real films rather than his words) | blind helper (skin physics): slightly egg-shaped with a funnel to the neck at launch, round high up; floats at ~31 km, tears beside the neck at 35.5 km if carried higher | SHAPE: MATCH (films). THE END: MISS (real ones burst at 29-32 km still climbing, the whole skin shattering) | proofs/blind-balloon-2026-11-21 |
| 7 | The speed of sound: about 340 m/s at the ground, about 295 m/s at 11 km (written 21 Nov 2026 ~13:32, before any run) | 340.1 m/s at the ground; 289.8 at 11 km (the column 210 K there, the real 217) | GROUND: MATCH. 11 KM: CLOSE (2% low) | runs/knowns.delta |
| 8 | Water boils at about 95 C in Denver, 1.6 km up (written 21 Nov 2026 ~13:32, before any run) | air 83.5 kPa (real 83.4); boils at 94.6 C | MATCH | runs/knowns.delta |
| # | Known result (what really happens) | Lab's answer | Match? | Run |
|---|---|---|---|---|
| 1 | Cold ping pong balls in a funnel under suction jam: they do not raft hex (Mike's own cold rigs, 2009) | jams at 2 degrees with ball-on-ball grip | MATCH (proofs/pingpong-2026-11-21/REAL-CHECKS.md) | runs/pingfunnel.delta |
| 2 | Salt grown slowly makes clear cubes; grown fast it branches and crusts, and makes hollow stepped hopper cubes | slow: a full square; fast: branching | cubes and branching: MATCH; hoppers: not seen yet | runs/saltplay.delta, saltcube.delta |
| 3 | Faraday's 1821 rotations: a wire turns round a magnet in mercury, and a magnet round a wire | checked 21 Nov: the turning is DRAWN at set speeds; only the energy book is worked out | NOT A TEST YET (magic, in MAGIC.md with what would make it real) | lab objects/faraday.delta |
| 4 | Basalt cooling cracks into columns, most of them six-sided (Giant's Causeway) | from a scatter, 3 in every 4 columns six-sided, fives and sevens in equal numbers (paper 5.3); no hexagon drawn | MATCH, but checked AFTER the run (the run came days before this row was written) | causeway.delta |
| 5 | Steel quenched from red hot is hard and brittle (it snaps); tempered (reheated gently) it gets its give back and bends (Mike, 11:56: "metal blade hardening"; written before any run) | not run yet | - | - |
| 6 | A pendulum 1 m long swings there and back once every 2.0 seconds (written 21 Nov 2026 ~13:32, before any run) | 2.009 s, stepped and counted (a small swing) | MATCH | runs/knowns.delta |
The sea from its surface (the wave tank’s sea) down to about 11 km. Known answers written before the run; the
column built 10 m at a time from sea water’s measured density and gravity from the Earth’s mass. Code:
lang/lab/runs/src/deep-src.delta.
| # | Known result (what really happens) | Lab's answer | Match? | Run |
|---|---|---|---|---|
| 1 | Water's pressure grows by about one more air pressure for every 10 m down | 1.99 at 10 m, 2.99 at 20 m, 3.98 at 30 m, 10.9 at 100 m | MATCH | runs/deep.delta |
| 2 | A balloon taken down is half its size by 10 m, a third by 20 m, a quarter by 30 m (the gas squeezed) | 501, 334, 251 thousandths | MATCH | runs/deep.delta |
| 3 | At the deepest (about 11 km) the pressure is about 1,100 times the air's at the surface | 1,122 times (real measurements about 1,070: about 4% high, because water's stiffness is kept at its surface value; corrected reason, diary 1F) | MATCH | runs/deep.delta |
| 4 | Polystyrene cups sent down on deep-sea dives come back shrunk to a fraction of their size (their air squeezed out) | at 11 km: 3% of its size, 31% of its height (foam 97% air, measured densities) | MATCH | runs/deep.delta |
| 5 | Sound travels through sea water at about 1,500 m/s (written 21 Nov 2026 ~13:32, before any run) | 1,511 m/s (from its measured stiffness and weight) | MATCH | runs/knowns.delta |
Rock under the sea floor, at a trench where one plate slides under another (the Mariana trench is the deep’s own
bottom). Its warmth comes out of the run: heat flowing up and made in the rock, carried only so fast. Known answers
written before the runs. Code: lang/lab/runs/src/crust-src.delta, plates-src.delta.
| # | Known result (what really happens) | Lab's answer | Match? | Run |
|---|---|---|---|---|
| 1 | Rock gets warmer going down, about 25-30 C for every km in continental crust | 25.8 C a km (average continent: heat flow 65 mW/m2, rock 2.5 W/m/K); 765 C at 40 km | MATCH | runs/crust.delta |
| 2 | In the deepest mines (about 4 km down) the rock is about 60 C | old cold rock (heat flow 45, quartzite 5.0): 49 C at 4 km (the average continent would give 116 C) | CLOSE: about 20% low | runs/crust.delta |
| 3 | Lava (basalt) comes out at about 1,100-1,200 C | in this model it would equal rock's melting point, which is put in | NOT A TEST here (circular) | runs/crust.delta |
| 4 | Hot water from vents on the deep sea floor stays liquid at about 400 C (the deep's pressure keeps it from boiling) | at 2,500 m (25.2 MPa) water boils only at 427 C | MATCH, with a caveat: real water there is past its critical point (374 C), with no boiling at all; the boiling rule does not know the critical point | runs/crust.delta |
| 9 | The sea floor deepens as it ages: about 2.5 km deep at the ridge, about 5 km at 50 million years (written 21 Nov 2026 ~13:32, before any run) | the cooled plate shrinks and settles: 2.78 km deeper at 50 million years, so 5.3 km (the 2.5 km ridge put in); 0.5 km at 1, 1.3 at 10, 3.9 at 100 | CLOSE (6% deep) | runs/knowns.delta |
Where the room sits: two plates, one diving under the other (known answers written before the run):
| # | Known result (what really happens) | Lab's answer | Match? | Run |
|---|---|---|---|---|
| 5 | The deepest sea is a trench where one plate slides under another (the Mariana trench, about 11 km: the deep's own bottom) | the trench is the shape the run starts with | NOT A TEST | runs/plates.delta |
| 6 | Volcanoes stand in a line about 100-200 km behind the trench, above where the sinking plate reaches about 100 km down | 100 km, straight from the chosen 45 degree slope | NOT A TEST (circular) | runs/plates.delta |
| 7 | Earthquakes get deeper the further from the trench, along the sinking plate, down to about 700 km | the plate's middle stays below 650 C (a chosen snapping limit) only to 445 km | MISS | runs/plates.delta |
| 8 | Heat flow through the sea floor is low over the trench (the cold sinking plate) and high under the volcanoes | over the trench 11 mW/m2 against 47 coming in (LOW); under the volcano line flat, about 38 (no flowing, melting rock above the plate in this model) | LOW: MATCH. HIGH: MISS | runs/plates.delta |
Also checked: the ocean plate cooled 50 million years by the lab’s steps matches the old maths’ cooling exactly (464 C at 25 km, 846 at 50, 1,142 at 80), and its heat flow (47 against 48 mW per square metre).