DON’T ASSUME THE LAW. BREAK IT.
What is the smallest modification of the assumed rules that produces an experimentally distinguishable world—and which experiment exposes it first?
Four places where the rules have nowhere to hide.
Each apparatus changes one narrow assumption and keeps the comparison surface deliberately simple.
Hollow Shell
Build the sphere from discrete mass elements. Never call the shell theorem. Perturb the distance exponent and watch whether interior cancellation survives.
Earth Tunnel
Cross a planet through its center under different density and persistence rules. The center becomes a severe test of what “gravity inside matter” means.
Orbit Persistence
Keep central attraction while allowing free velocity to decay by a tunable amount. Ask what a planetary trajectory itself permits.
Deep-Space Coast
Let a synthetic outbound probe coast for a long baseline. Tiny changes that hide during one orbit can accumulate into navigation-scale residuals.
Which apparatus breaks first?
The engine searches parameter space for the smallest perturbation that crosses a declared discrimination criterion. The criteria are part of the experiment design, not claimed measurement limits.
Make assumptions replaceable.
Source rule
Represent the source explicitly enough that a shortcut cannot smuggle in the expected answer.
Spatial rule
Inverse square becomes a selectable geometry in the shell test rather than an invisible theorem.
Inertia rule
Persistent velocity and controlled decay are separated so “conservation” cannot close the question by definition.
Discriminator
Search for conditions under which rival models disagree maximally while the apparatus stays simple.
Rotate the simulation itself.
Every experiment becomes a deterministic bitmap sequence, then that sequence is projected live onto a draggable 3D surface. Raw state, GLB and OpenUSD remain independently inspectable.