What would count as a test?
Separate an analogue you can build, a signal you can observe, and a consistency condition you can formalize. A provocative claim is not yet a laboratory result.
Choose an evidence route
Define the analogue
A laboratory analogue would test shared dynamics, not recreate a black hole. The experiment needs a defined boundary, control variables, and a predicted observable.
A protocol note is not an experimental result
Read the explanation
The saved brief offers three evidence routes and selects one at a time: an analogue laboratory test, astronomical observation or formal consistency audit. Bars use one hundred fifty pixels per route. Choosing one route changes its explanation, proposed measurement and falsifier prompt. These route labels are protocol-design options, not three completed tests, proof that singularities are rejected or evidence of a Bitcoin vulnerability. The astronomical route describes one observable signal intended to distinguish model A from model B: two competing models and one discriminating observable. Bars use two hundred pixels per count. An observable is useful only when alternatives, noise and selection effects are made explicit. This tool supplies a planning prompt, not actual astronomical data, an uncertainty estimate or a derived physical prediction. Local state stores two fields: selected route and the users falsifier note. The JSON export has six fields, including the route, title, explanation, measurement, entered falsifier and evidence boundary. Bars share eighty pixels per field. Reset returns the laboratory route and an empty note and removes saved state. A typed falsifier is not automatically checked, authenticated or executed as an experiment.