TESTING METHODOLOGY — 4D RESEARCH
(Inherited from TLT methodology, adapted for 4D speculative work)

ALL tests are designed for OUTPUT AGNOSTIC
	this means that no test is designed to prove or to confirm a piece of information
	the purpose of the test is to see what happens and comes from the data

EACH test is placed in the unaudited bin; similarly each test undergoes an audit prior
	to placement in the audited folder. This treatment is identical to how the data
	collected from the tests is handled to ensure we can quarantine data that has been
	contaminated by bias or intentional nudging from AI. Experience has shown this
	process is required due to the nature of AI approaches to human prompting

ANY insight or confirmation towards the theory triggers an automatic audit of the test.
	this is to ensure that no bias towards a result or outcome can easily creep in
	no matter how exciting or compelling it might be

ALL tests will use verified physics engines - NO TOY MODELING

ADDITIONAL RULES FOR 4D RESEARCH:
	- This project is SPECULATIVE. Results must earn their place.
	- Anything discovered here can only ACT TO CLARIFY core TLT.
	- Do NOT add findings from this project to core TLT without explicit approval.
	- The 24-cell is ESTABLISHED MATHEMATICS (1850s-1890s). Only our observations are new.
	- Maintain strict separation: 4D research has its own audited/unaudited flow.

AUDIT PROCESS (same as TLT):
	1. CLAUDE builds the test, then audits against the theory (line by line boolean)
	2. CLAUDE performs contrarian self-audit (are we nudging?)
	3. Fix any issues before implementation
	4. Run test, collect results
	5. External audit by GEMINI and GROK (independent, no prior context)
	6. Scores >= 7 required; below 7 triggers investigation
	7. Promotion to audited/ only after clearance