Adversarial Viability Control Theory

Holding still and controlling the exchange

Take one back control: seatbelt, both hooks, and a defender whose chin is down and whose hands are on the choking arm. Ask three questions about one declared target, the rear naked choke, and the answers come apart. Is it reachable? Yes: there is at least one way the exchange could go from here that ends with the choke locked, the defender reaches for a hook and the arm goes under the chin. Is it enforceable? Not yet: against the replies this defender actually has in this window, hands on the arm and chin down, the attacker has no sequence that forces the arm under the chin whatever the defender does. Is the defender viable? Yes, for now: the neck can be kept defended over this window without the defender improving anything at all. Most coaches would call the whole situation “back control”, and it is. The three answers are three separate facts about it.

Adversarial Viability Control Theory is the account of those facts and of how they change. Control, in it, has more than one dimension: what you can reach, what you can force against the answers this opponent actually has, what you can keep safe, what you can interrupt, and what you have made costly or predictable for them are separate facts, and no one of them stands in for the rest. A position can be strong on one of them and weak on another, which is why a dominant-looking position can leave the opponent with more live options than a loose-looking one.

None of this says positions are unreal or unimportant. Mount, side control and half guard are human-scale summaries of arrangements people found worth distinguishing, and the summaries are useful. The theory’s claim is narrower and sharper: they are coarse, they are not the foundational description of the state, and the finer description is the one the theory works with. Every experienced grappler already knows this, which is why useful advice arrives as a qualifier: heavy mount, low mount, mount with the arm trapped. Those qualifiers are doing the real work.

Reachable, enforceable, viable

Three questions that get asked as one, and pulled apart here for good. Reachability asks whether a target can occur at all: is there at least one admissible way the exchange could go, given what both people could do, that ends there. Enforceability asks something strictly harder: can this participant force the target across the class of replies the opponent has, a class that has to be stated or at least understood, which is why “against everything” is never a complete claim. Viability asks a different question again, usually from the defender’s side: can they preserve a safe, non-losing relation over the window, which does not require getting anywhere. In the back control above, the choke is reachable and not yet enforceable, and the defender is viable for the window; the hand fight is where the second and third answers are being contested.

Reachability asks whether a target state/action region can be physically reached. Enforceability asks whether the participant can reliably cause or preserve the desired progression against relevant opponent responses. Viability asks whether an acceptable continuation set can be preserved while avoiding unacceptable or terminal outcomes. Evidence for one must not be silently counted as evidence for another.

The Control Grammar keeps the same three as one-line questions, with the response class made explicit. The first, in its own words:

Can the target occur under at least one admissible joint interaction?

The second asks whether this participant can force the declared target across the relevant opponent response and disturbance class, and the third whether they can preserve the required safe or non-losing relation. Reachable never implies enforceable, and viability never requires positional improvement.

Anyone who has taken a position and immediately lost it has met the gap between the first and the second. Anyone who has held a position while slowly running out of good options has met the gap between the second and the third. The theory’s insistence is that evidence for one must never be quietly counted as evidence for another. “I can get to the back” is not evidence that you can keep it, and a highlight reel of finishes is not evidence that the finishes were enforceable rather than gifted.

What control changes

Think of a state as a place on a map and of the transitions out of it as the roads. A position name tells you the place. Control is the shape of the roads: which ones are open to you, which of them you can force rather than merely attempt, which are open to the opponent, and which of theirs you have closed, made expensive or made predictable. Two athletes can stand on the same place with entirely different road systems around them.

That picture is the content of several of the theory’s claims at once. Control is relational: it belongs to the pair, and the same top-player shape is worth more against one opponent’s road system than another’s. The opponent’s residual response structure is part of the state, not a footnote to it. And exchanges are temporally competitive: when two roads are open at once, who arrives first can decide the outcome, so a transition that is forceable in principle can still be lost to a faster interruption. One consequence is that much of the work of control is closing their roads, opening yours, and winning the races between them.

The fifteen claims

The theory states itself in fifteen claims. Each is listed exactly as the theory names it, with what it means for someone standing on a mat.

1. C01 coupled adversarial system Two grapplers are one system. Neither state can be described without the other.
2. C02 causal feedback policy What a grappler does next is chosen on the state and on what they can see, over a horizon longer than the current grip.
3. C03 bilateral state evolution Both states change together. Your posture change is their option change.
4. C04 reachability != enforceability Being able to get somewhere is different from being able to make it happen against resistance.
5. C05 viability != reachability Surviving is different from getting somewhere. A non-losing relation can be kept without improving anything.
6. C06 control is relational Control is a property of the pair, never of one body: a relation between the two of you, not a property of your shape.
7. C07 residual response structure matters The options an opponent has left are part of the state, and they shape what can happen next.
8. C08 grappling temporally competitive Who becomes effective first can decide an exchange. Grappling is a race as well as a shape.
9. C09 terminality mechanical and temporal A finish is a mechanical and temporal state, reached before the tap and independent of it.
10. C10 technique feedback-policy implementation A technique is a policy: entry, branches on response, and exit. It is never one fixed sequence.
11. C11 positions coarse-grained surface labels Position names are human-scale summaries of something finer, and the finer description is what the theory works with.
12. C12 partial observation native Nobody sees everything. Reading, guessing, feinting and scouting are inside the theory.
13. C13 rules/value distinct from mechanics Rules and scoring change what a state is worth without changing what it is.
14. C14 hybrid transitions legitimate A grip made, a post lost, a boundary touched: some changes are discrete events, and the theory keeps them.
15. C15 non-scalar structure permitted There is no single number along which grappling situations can be honestly ranked.

Three are worth reading twice. Control is relational: it is not a property of your body but of the pair of you. Technique is a feedback policy: a way of choosing what to do next given what you can see, rather than a fixed sequence. And the structure is non-scalar: there is no single number along which grappling situations can be honestly ranked, which is why nothing on this site will offer you one.

Techniques are policies

Two guard passers run the same knee cut. The first runs it the same way every time and succeeds when the guard player fails to answer. The second runs it as a branch structure: a different continuation for each of the guard player’s answers, so that a frame, a turn and a retracted knee each meet a prepared response rather than a repeat of the same drive. Only the second has a technique in the theory’s sense, because a technique is the implementation of a policy conditioned on the opponent’s response, over a horizon longer than the current grip.

The evidence rule that goes with this is strict on purpose. Seeing an athlete choose a move, succeed or fail, and switch to another later is compatible with a policy and is not evidence of one. Positive evidence is visible branching, continuation, troubleshooting and selection that tracks the opponent. For a coach, that decides what to watch for: not whether the pass looked smooth, but whether the passer answered the frame.

C02 and C10 concern state- and/or information-conditioned action over an appropriate horizon. Positive instantiation requires source-visible opponent/state-conditioned selection, continuation, branching, troubleshooting or equivalent policy structure. Mere technique choice, success/failure and later switching supplies compatibility or pressure, not automatically strong positive evidence.

Terminality is a state, and the tap is an event

A rear naked choke reaches a point at which no timely defence remains: the hand-fight is lost, the chin is gone, the second arm is locking. That point is mechanical and temporal, and the theory calls it terminality. The tap comes after it, sometimes well after it, and sometimes before it, when a defender concedes a choke that was not yet finished. Keeping the state and the event apart is what lets the theory talk about a finish precisely, and it is also a safety fact: an early tap concedes to the threat before the state arrives, and waiting for unconsciousness lets the event do what the state had already settled.

Evidence about terminality is graded accordingly. Naming a submission, showing how it is finished or defended, and showing enough of the mechanical and temporal structure to say when no timely prevention remains are three different levels, and only the last is strong evidence for the claim.

A source naming a submission, a source exposing finishing mechanics or defensive interruption, and a source exposing enough mechanical-temporal structure to support C09 are different evidential levels. Only the last is strong positive C09 evidence.

Reading the opponent is inside the theory

A wrestler feints the single leg, the opponent sprawls early, and the snapdown that the sprawl exposed is the real attack. Nothing mechanical changed until the sprawl; what changed was the opponent’s picture of the situation, and they acted on the picture. The theory treats partial observation as native to grappling rather than as noise: incomplete recognition, opponent reading, uncertainty, prediction, deception, scouting, delayed information and learned expectation are all part of how the coupled system evolves.

That has a coaching consequence. An athlete who is never deceived and never deceives is not demonstrating the whole game, and evidence that someone reads an opponent counts as evidence in the theory’s own terms, without anyone having to write down a formal model of what the athlete sees.

C12 does not require a source-published formal observer model. Positive evidence can include incomplete recognition, opponent reading, uncertainty, prediction, deception, scouting, delayed information or learned expectations. Mere absence of explicit omniscience is not strong positive evidence.

Where rules come in

The theory describes a grappling interaction under a declared rules and value context. Under one ruleset the reap that sets up a heel hook is a foul; under another it is a common entry. The mechanics of the position are identical in both rooms; what the rules change is which of the available roads is worth taking, which are forbidden, and what the clock makes urgent. The same locality lets the theory describe the grappling phases inside a mixed contest without claiming to be a theory of striking.

AVCT applies to a grappling interaction under a declared rules/value context. It may describe local grappling phases inside MMA or another mixed domain without claiming to be a complete theory of striking, weapons or all mixed-combat victory mechanisms.

How the theory could be shown wrong

A theory that cannot fail is not saying much. This one names seven ways it could, and keeps them where they can be used. The sharpest is position superiority: if named positions turned out to predict what happens next better than the option-structure account does, the theory is in trouble. Response-restriction redundancy is another: if closing the opponent’s options turned out to add nothing beyond reachability and viability, a central claim is empty.

1. F1 coupled-game insufficiency
2. F2 reach-viability insufficiency
3. F3 position superiority
4. F4 response-restriction redundancy
5. F5 hybrid irrelevance
6. F6 feedback-policy failure
7. F7 terminality mismatch

The rule for reading comparative evidence is equally deliberate. Material that merely does not contradict the theory counts as compatibility, which is weaker than positive instantiation; a gap in the sources counts for nothing; and no grade is ever converted into a percentage, because a number would let a pile of weak compatibility look like support.

What it hands to the Grammar

This theory says what control is. It does not, on its own, give a vocabulary precise enough to describe a particular grip, a particular frame, or the moment a position starts to leak. That is the job of the Control Grammar, which takes reachability, enforceability, viability and the opponent’s residual options and turns them into named relations and questions you can ask of a specific exchange. The capability architecture then asks what a grappler has to become able to do to create, keep and break control so described.

One thing this theory is careful not to be is a coaching method. It describes what is happening between two grapplers. The layers that say how to teach it are kept separate, and are tested separately.

Standing

The account has been tested comparatively against a wide range of submission-grappling material and has strong bounded support across that domain: real evidence, and not deductive proof, and not evidence that the theory is itself a way to coach. What stays open is mostly measurement: how to measure enforceability and viability in a live round, how to measure what an athlete perceived, how finely a policy should be described, and whether independent observers agree.