Research and validation
How Theory of Everyjitsu is tested: what was performed on each part, what returned, the candidates that failed and were kept, the rules it holds itself to, and what needs people.
Adapted from InGrappling, Research and validation. System Games did not invent this curriculum.
What has been tested, what it returned, what failed, and what has never been examined at all. The theory keeps its own negative results, and this is where they live.
This is a research page. It reports what an audit or an open question established, not settled current theory, and it is deliberately outside the practitioner navigation.
Structural testing asks whether a body of work is internally coherent, survives adversarial cases, and does not contradict what it depends on. It is done by construction and audit, against cases fixed before the result is known, and a great deal of it has been done here. It takes more than one form: the worked levels, the curriculum and the grammar carry regression results with their counts, while the two foundational documents were tested comparatively, against a range of submission-grappling lineages and institutions, and state their own result. The record below says which was performed on which part.
Human-outcome evidence asks whether people who train this way end up better. It requires people, measurement and time. None has been collected. One pilot protocol exists, for whether coaches can use the recording instrument in a live room and code the same event the same way, and it has not been run. Every claim on this site that touches human outcomes is published as an open question for exactly that reason, and every ordinary page that draws on such material says so in one sentence at its foot.
Every document in the theory is fixed once it has passed its test, and it is never edited afterwards. A repair is a new version, produced only after a test whose terms were written down before grading and which returned no failures. The tests are adversarial rather than confirmatory: a theory is checked against cases built to break it, comparative material is graded as positive instantiation, compatibility, pressure, gap, out of scope or counterexample, and none of those grades is ever converted into a score. The two foundational documents state their own evidential position in those terms: strong bounded comparative support across the tested domain, and explicitly not proof.
Candidates that fail are kept with their results rather than deleted. What went wrong is part of the record, and a later reader can see both the failure and the successor that passed.
These are commitments the theory states about itself. They constrain what every page on this site is allowed to say, and they are as much a part of the record as any test result.
What you can do, what has been observed, and what level you are called are three different things, and they are kept apart.
Training comes before tracking. Evidence should support coaching, not turn every round into data collection.
Readiness to compete is judged against a particular ruleset and a particular field, not as a general property of a grappler.
Ideas that failed are kept, with their results, rather than deleted. What went wrong is part of the record.
What the theory rests on is fixed. Nothing downstream may quietly edit what it inherited; a repair is a new version, tested first.
Development is multidimensional. Grappling ability cannot honestly be reduced to a single score, so nothing here offers one.