Research and validation

Two kinds of support, and only one of them exists here

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.

How the theory tests itself

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.

The rules the theory holds itself to

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.
• What has not been seen stays unknown. It is recorded as its own thing, never scored as a zero and never read as poor performance.
• Safety is a hard constraint. Being strong somewhere else does not cancel a safety concern.
• Evidence about a capability sits in one of seven states, from first exposure to self-organised use. They are ordered, and they are not a score.

The audits

Four audits govern the material. Their verdict counts are published as recorded, including the failures, because an audit that only reported passes would not be worth running.

• B1-to-A1 Whole-Pathway Coherence, Compression and Temporal Audit 183 pass · 23 pressure · 4 redundancy · 0 fail Passed, with open pressure
• B1-to-A11 Whole-Pathway Coherence, Redundancy, Temporal and Credential Audit 411 pass · 74 pressure · 15 redundancy · 0 fail Passed, with open pressure
• B1-to-A11 Curriculum Upstream Congruence and Competitive Readiness Falsification Audit 631 pass · 144 pressure · 1 source gap · 4 fail Regression failed
• Post-A1 Advanced Regime Coupled Successor Targeted Regression 315 pass · 40 pressure · 5 redundancy · 0 fail Passed, with open pressure

The two whole-pathway audits established, at document level, that every adjacent transition in the twenty-level path keeps a distinct developmental object, that no belt rests on time, attendance, block completion, technique count, results, teaching or lineage, and that the arithmetic of the worked blocks is not a promotion schedule. The competitive-readiness audit is the one that returned failures, and its findings are the subject of the competition readiness page. Those four failures are the most useful thing in this section.

Failed candidates are kept

Two frozen results are failures, and both were superseded by successors that passed. They are not deleted and they are not offered as alternatives to the current material. One was blocked on an administrative status defect rather than a technical one, which is a distinction the record keeps carefully: an audit that completes and returns a governance defect is a different event from a test that fails on substance.

Where the record is thinnest

The two most upstream parts of the theory, the ontology and the theory of control, are also the two whose paperwork is least complete. Each is a clarification of an earlier version and presents itself as that version’s successor, and the record that would confirm the succession has not been recovered. For the ontology, the earlier version itself has since been found and checked against the fingerprint the current version cites for it, and what the current version states that it preserves from that version, the inventory of what a grappling exchange consists of, is now published on the ontology page from the earlier text itself. For the theory of control the earlier version remains absent.

That is a limitation about standing, not about substance. The documents themselves are present and checked against their recorded fingerprints, and what they say is published on their own pages. What cannot be stated is their formal place in the theory’s version history, and no page here states it. On the question of provenance, the parts nearest the foundation are less well evidenced than the parts built on them.

What has no page yet

Some material that exists has deliberately not been turned into pages. Each case and the reason for it:

• /theory/pathway/b2/ and the other nineteen worked-level pages Nineteen of the twenty level programmes have not been published. One of them rests on a record that is attested elsewhere but cannot be verified here, and the other eighteen were read closely enough to place them in the pathway but not closely enough to write a page about each. Publishing thin pages from a structural skim would look like coverage and would not be. Changes when: Each level is read in full and the read is recorded, and for the second level the missing governing record is recovered and checked against its fingerprint.
• /theory/research/sources/ There is no browsable index of the source documents. Everything a page relies on is listed in that page’s own provenance disclosure, with its identifier and fingerprint, which is what a reader checking a claim actually needs. Changes when: A reader needs to browse the documents themselves rather than check the sources behind a particular page.
• the twelve grammar concept pages The relations that make up the control vocabulary are sections of one page rather than twelve pages of their own. Splitting them before anything links to them would add routes without adding anything to read. Changes when: Something needs to link to an individual relation directly.

How a pressure becomes a change

Open pressures are not defects to be closed by argument. A pressure is resolved by a test whose terms are written down and fixed before the result is known, and if the result warrants it, a successor version is written. The predecessor is never edited and never deleted, so the trail from a difficulty to a change stays readable afterwards. The pressures the current material carries are listed at the foot of this page, in the words the record uses for them.

How the public pages are kept honest

Every passage on every theory page rests on statements read from those fixed documents, and the statements carry their own status: what is formally defined, what testing was performed on it and what returned, what limit it states, and what has and has not been examined with people. A page can explain, simplify, combine and order that material; it cannot assert anything the documents do not carry, strengthen a claim about outcomes, or drop a material uncertainty. Those checks run against the pages a reader actually sees. The ordinary pages keep the machinery out of the way; this one is where to look at it.