1,000 scaling claims: overview

1,000 new claims, each evaluated. Ten new families of 100. 983 hold, 17 fail. Every number comes from code; failures are kept.

Families

Family	Statement	Held
G1	Volume grows as the cube and surface as the square, so any area-based mechanism	99/100
G2	Losing one channel of n costs 1/n of flow, so redundancy is cheap at large n.	96/100
G3	Surface-to-volume rises as 1/d, so nanoscale chemistry is surface chemistry.	100/100
G4	Diffusion-limited growth makes a halved time window worth 29% less, not 50%.	100/100
G5	Independent stages add variance linearly, so stage count is an uncertainty multi	100/100
G6	Galvanic current is linear in the area ratio.	99/100
G7	Calibration degrees of freedom bound the number of identifiable parameters.	100/100
G8	Per-test failure probability compounds: 100 tests at 1% fail 63% of the time.	100/100
G9	A ratio's error is dominated by its denominator when the ratio is large.	89/100
G10	Thermal strain is linear in excursion and gives 12 microstrain per kelvin.	100/100

What is different from the first 1,000

The first set tested measurement pathologies. This set tests scaling relations: cube-versus-square, 1/n redundancy, surface-to-volume, square-root diffusion, variance addition, degrees of freedom, compounding probability, denominator dominance, thermal strain. These are the relations that decide whether a design survives a change of size.

Limits

Scaling relations are textbook; the specific numbers at each parameter value are the new content.
Verification is arithmetic, not experiment.

Incoprea Research · mirrored study, 3 November 2026. Composition and computation. No experiment performed.