MMA submissions explained: chokes, joint locks and the tap

A submission ends a fight the instant the losing fighter taps, and the tap is the only part of the whole subject that is a rule. Everything else is mechanics.

That distinction is worth holding onto, because it explains why submissions look so strange to a new viewer. A fight stops, both men get up, nobody appears to have been hurt, and one of them has lost. What happened is that a position became unwinnable, the fighter in it recognised that continuing would cost him a joint or his consciousness, and he used the one mechanism the sport provides for getting out of it.

Here is how each family of submission works, what makes one quick and another slow, what the rules forbid, and why an escape has to happen earlier than instinct suggests.

The tap is the rule, not the technique

Under the unified rules a fighter concedes by tapping clearly and repeatedly, with a hand or a foot, against the opponent or the canvas. A verbal concession counts. A referee who sees it stops the contest at once, and the result is recorded as a submission.

Three features of that make it the most elegant rule in combat sport.

It is instant and unappealable. There is no count, no recovery period, no doctor's assessment. The fight is over at the moment of recognition.

It is available to the losing fighter alone. No corner can tap for a fighter, and no opponent can claim one that was not given. The person whose joint is at risk is the only person who can end it.

It transfers the decision to the person with the best information. Nobody in the building knows how close an elbow is to failing except the man whose elbow it is.

The rule works because the alternative to tapping is not survival but injury. A fully applied joint lock has already won. What remains is a choice about what the loss costs, and the sport has arranged things so that a fighter can take the cheaper option without being penalised for it beyond the loss itself.

• 3Accepted ways to signal a submission
• 0Seconds of grace after a clear tap
• 0Fingers and toes a fighter may attack
• 0People other than the fighter who may tap for him

Values as the unified rules and standard refereeing practice set them. The zeroes are deliberate rather than missing.

Two things a choke can do, and only one of them is quick

Every choke in the sport belongs to one of two mechanical categories, and the difference between them is the single most useful thing a viewer can know.

A blood choke compresses the arteries in the neck. It interrupts the flow of blood to the brain. The effect arrives quickly, the fighter who takes it loses consciousness without much warning, and he recovers quickly once the pressure comes off. Grapplers call these strangles, and almost every named choke in MMA is designed to be one.

An air choke compresses the windpipe. It interferes with breathing instead. It takes much longer, it hurts a great deal more, and it is far more likely to leave the throat damaged afterwards.

The counterintuitive part is that the quick, clean mechanism is the one grapplers want. An air choke is not a more brutal version of a blood choke; it is a worse one. It is slower to finish, it gives the defender more time, and the damage it leaves behind is to a structure that does not need to be damaged for the fight to end.

So when a commentator says a choke is "under the chin" or "across the throat", that is not a compliment. A well-applied strangle sits on the sides of the neck, and a choke that has slipped onto the front of it has become harder to finish rather than easier.

This is also why a fighter who is being strangled properly often looks calm. There is no suffocation, no struggle for air, and very little to feel until there is nothing to feel at all. A fighter grimacing and gasping is usually in an air choke and is usually going to get out.

The chokes that come from behind

The back is the best place in grappling to be, and the reason is that the opponent cannot see, cannot strike effectively and cannot defend his neck with his own structure.

The rear naked choke is the canonical submission of the sport. One arm goes around the neck with the forearm under the chin, the hand grips the opposite bicep or the fighter's own shoulder, and the other arm goes behind the opponent's head to close the space. The compression comes from the two sides of the neck, so it is a blood choke, and the name refers to its lack of any clothing grip rather than to anything else.

Its reputation rests on one property: it is extremely difficult to defend once the arm is under the chin. The entire defensive battle therefore happens before that, in hand fighting, and a fighter who has lost the hand fight has usually lost the fight.

The body triangle is not a choke at all but it belongs in the same discussion, because it is what makes the choke available. Locking the legs around the torso from behind removes the opponent's ability to turn, which removes his ability to defend his neck with a shoulder. A fighter who takes the back and secures the body first has converted a transient position into a permanent one.

The defensive principle from the back is worth stating because it is so consistently ignored. A fighter being attacked from behind must fight the hands, keep his chin tucked to deny the forearm a path, and turn into the opponent rather than away from him. Turning away exposes the neck further, which is exactly what panic produces.

The chokes that come from the front

Front headlock positions generate a family of their own, and the key variable is whether the opponent's arm is trapped alongside his head.

The guillotine is applied from the front with the arm wrapped around the neck and the opponent's head under the armpit. The variants differ in where the forearm sits and how the finishing pressure is generated. The high-elbow versions place the forearm across the side of the neck and function as blood chokes. Flatter versions drag the forearm across the throat and behave more like air chokes, which is why they are slower and why defenders survive them more often.

The guillotine is the submission most available in open fighting, because it is the natural punishment for a badly executed takedown attempt. A fighter who dives at the legs with his head down has offered a neck.

The arm-in and arm-out distinction runs through this whole family. Trapping the opponent's arm inside the choke changes the geometry: the arm occupies space the neck would otherwise have, and in some variants the opponent's own shoulder becomes part of the compression. An arm-out guillotine is a tighter squeeze on the neck alone. An arm-in version is a different mechanism using the fighter's own body against him.

The anaconda and the darce are front headlock chokes that thread the arm under the opponent's near arm and across the neck, closing the circuit against the fighter's own bicep and the opponent's trapped shoulder. They look similar from the outside and the thread runs in opposite directions, which is the kind of detail that matters enormously to the person applying it and not at all to the result.

The chokes that use the opponent's own arm

A distinct group of submissions works by trapping the opponent's arm across his own neck and closing the other side with the fighter's body. These are the most mechanically interesting chokes in grappling because the defender supplies half the apparatus.

The triangle choke is applied with the legs. One of the opponent's arms is inside, the head is inside, and the legs close in a figure configuration around both. The compression comes from the fighter's thigh on one side of the neck and the opponent's own trapped shoulder on the other. A fighter caught in it is being strangled partly by himself, which is why pulling the trapped arm out is the first thing every defender is taught.

The arm triangle inverts the geometry. The fighter uses his own arms and chest rather than his legs, pinning the opponent's arm across his neck and closing the far side with his shoulder. It is a top-position submission where the triangle is a bottom or transitional one, and the mechanism is identical.

The omoplata is the shoulder lock built from the same leg configuration, and it demonstrates how thin the line is between a choke family and a joint lock family. Change the angle of the trapped arm and a strangle becomes an attack on a shoulder.

47%53%

• Chokes and strangles8
• Joint locks9

A count of the submissions described in this article, grouped by mechanism. It is a classification of this article's contents rather than a measure of how often each family appears in competition.

Show the numbers

The two families, counted across the attacks named here

Item	Value
Chokes and strangles	8
Joint locks	9

The near-even split is not an accident of this article's structure. The two families are roughly equally developed because they solve different problems. A choke requires access to the neck, which means winning a dominant position. A joint lock requires access to a limb, which an opponent offers far more readily, often while defending something else.

Joint locks start with a hinge and a direction

Every joint lock does one of two things to a joint: it extends it past its range, or it rotates it past its range. Knowing which tells you everything about how the submission behaves.

An extension attack, such as a straight armbar, is felt early. The joint approaches its limit, the sensation is unmistakable, and the fighter has a clear warning that the position is lost.

A rotation attack is felt late, and in the wrong place. The force is applied at a distance from the joint and travels along the limb, so the pain can arrive after the damage has begun. Every submission with a dangerous reputation in this sport is a rotational one, and that is the reason.

Elbow2

Shoulder3

Knee2

Ankle and foot2

A count of the named attacks described below, by the joint the force arrives at. The heel hook is counted at the knee rather than the ankle because the knee is where the rotation is resolved.

Show the numbers

The joint locks named in this article, grouped by target

Item	Named attacks described
Elbow	2
Shoulder	3
Knee	2
Ankle and foot	2

The elbow: the armbar and its relatives

The armbar is the archetype. The opponent's arm is isolated with the elbow pinned against the fighter's body or hips, and the fighter arches to extend the joint beyond straight. It is an extension attack, it is felt early, and a fighter who does not tap will have the elbow hyperextended.

Its defensive principles are unusually clean. Keep the elbow bent, keep the hands connected, and get the trapped arm's thumb pointing in a direction that resists the extension. Once the arm is straight the defence is over, which is why the whole fight happens in the fraction of a second before that.

The straight arm lock from the top, applied with the opponent's arm pinned to the canvas or across the fighter's shoulder, is the same mechanism arriving from a different position. The rule holds across the family: the elbow is a hinge, hinges fail in one direction, and the attack is always about getting the limb straight.

The shoulder: kimura, americana, omoplata

The shoulder is a ball joint rather than a hinge, so it cannot be hyperextended in the way an elbow can. It is attacked by rotation, and the direction of rotation separates the techniques.

The kimura bends the arm at the elbow and rotates the shoulder internally, with the fighter gripping his own wrist in a figure configuration to generate leverage. The elbow feels pressure first, the shoulder takes the damage, and a defender who focuses on the wrong joint will tap too late.

The americana, also called a keylock, rotates the shoulder in the opposite direction with the arm bent and the hand pointing up. It is the top-position mirror of the kimura and is most available when a defender leaves an arm away from his body while pinned.

The omoplata achieves the same internal rotation using the legs rather than the arms, from a bottom position. Its practical value in MMA is complicated by the fact that it requires the fighter to put himself in a position where a standing opponent can disengage, which is a trade the grappling ruleset does not demand and the fighting one does.

The shoulder family teaches the general lesson about rotation. A fighter can be in serious trouble in a kimura while the arm looks merely bent, and the warning signal arrives in a joint other than the one that is going to fail.

The legs, and why the heel hook attacks the knee

Leg locks are the fastest-moving area of grappling and the most misunderstood.

The heel hook uses the heel as a handle to rotate the lower leg. The force travels up the shin and arrives at the knee, where the ligaments resolve it. The name is misleading and the misleading name is dangerous: a fighter who expects to feel his heel hurting is waiting for a warning that is not coming. Rotational knee attacks are the reason leg locking is treated with more caution in training than arm locking, and the reason several grappling rulesets restrict them by experience level.

The kneebar is the knee's equivalent of the armbar: an extension attack, with the leg isolated and the joint pushed past straight. It is felt early and defended by keeping the leg bent, exactly as the armbar is defended by keeping the arm bent.

The straight ankle lock attacks the ankle by extending it, with the fighter's forearm as the fulcrum against the Achilles tendon. It is the most common leg entanglement in MMA because the entry is available from the simplest leg positions.

The toe hold rotates the foot with a grip across the toes and the instep. It is a rotational attack arriving at the ankle, and it falls on the legal side of the line because it grips the foot as a unit rather than manipulating individual digits.

Leg locks spent decades with a poor reputation in grappling circles, dismissed as low-percentage, unsporting or a shortcut taken by people who could not pass a guard. The reputation collapsed when specialists demonstrated that leg entanglements were a complete positional system with their own hierarchy of control rather than a collection of opportunistic grabs. The reversal is the clearest case in combat sport of a technical prejudice being dismantled by results, and it happened within a single generation of competitors.

Neck cranks, spine locks and the line the rules draw

A neck crank attacks the cervical spine by bending or twisting it rather than by compressing the arteries or the windpipe. A spine lock does the same to the thoracic or lumbar spine.

The regulatory position is counterintuitive. MMA under the unified rules generally permits neck cranks, while numerous sport grappling rulesets ban them outright, as they ban several spine attacks. A fighter moving from a grappling competition into an MMA contest is therefore moving towards a more permissive ruleset in this particular respect, which is the opposite of what most people assume.

The reasoning behind the difference is about the surrounding environment rather than the technique. In a grappling match the only way to end a contest is a submission or points, so a ruleset that permits dangerous attacks on the spine creates a strong incentive to use them. In an MMA contest there are many ways to win, strikes are available, and attacks on the neck are one option among several rather than the shortest path to victory. The incentive is weaker, and the rules have been written accordingly.

The practical distinction a referee has to make is between a choke that has slipped into a crank and a crank applied deliberately. Both are legal, so the question is not legality but recognition, and it affects the stoppage: a fighter in a crank will not go unconscious, so the referee is waiting for a tap rather than watching for a fighter going limp. The broader framework the referee is working within is set out in the explanation of the rules a UFC contest is fought under.

What the rules actually forbid

The foul list touching grappling is shorter than people expect and very specific.

Small joint manipulation is a foul. Fingers and toes may not be grabbed and twisted individually. The joints are small, the leverage available against them is enormous, and no amount of conditioning protects them. The practical test is whether the fighter is gripping a limb as a unit or isolating a digit, which is why a toe hold across the whole foot is legal and pulling on one toe is not.

Spiking an opponent onto the head or neck is a foul. This matters in grappling because a fighter defending a submission is sometimes lifted, and the way he is returned to the canvas is regulated.

Grabbing the fence, the shorts or the gloves is a foul. All three are grappling fouls in practice, because all three are used to stabilise a position or prevent one.

Everything else in the submission catalogue is permitted, which is a wider licence than sport grappling allows and a narrower one than the sport's reputation suggests. There is no rule against a submission because it is dangerous. The dangerous ones are legal, and the responsibility for recognising them in time sits with the fighter.

Position before submission, and why it is more than a slogan

The oldest principle in grappling instruction is that position comes before submission, and in MMA it is close to a law of physics.

A submission requires the opponent to be unable to move in the direction that would defeat it. Securing that immobility is what position means. A fighter who attacks a limb without it is attacking a limb that can be withdrawn, and the attempt costs him the control he had.

1. 1The position changesA takedown, a sweep, a scramble or a reversal puts one fighter in a controlling position. Nothing is threatened yet, and the submission is not available because the opponent can still move.
2. 2Control is establishedThe controlling fighter removes the opponent's ability to move in the directions that would escape. Hips are pinned, the body is secured, posture is broken. This is the step that is invisible on television and decides everything.
3. 3A limb or the neck is isolatedWith movement removed, one target becomes available. The opponent cannot withdraw it, because withdrawing it requires the movement that has already been taken away.
4. 4The lock or choke is appliedThe mechanism goes on. For an extension attack the defender feels it immediately. For a rotational attack he may feel very little until the damage is already in progress.
5. 5The defender decidesHe escapes, if the escape was started early enough, or he taps. A fully applied submission offers no third option, and waiting is not a defence because the position has already resolved the fight.
6. 6The referee intervenes if neededA fighter who goes unconscious in a choke, or whose joint has clearly failed, is protected by the referee without a tap. The result is still recorded as a submission.
7. 7The bout is overA tap or a referee's stoppage ends the contest immediately. There is no count, no recovery period and no appeal.

The generic structure of a finish, from the change of position to the stoppage. Every step is a control problem rather than a technique problem.

The reason this ordering is non-negotiable in MMA specifically, rather than in grappling generally, is that a failed submission attempt in a fight gives up more than it does in a grappling match. A fighter who abandons a dominant position to chase an armbar and misses is not back at neutral. He is underneath an opponent who is now free to strike, and the cost of the gamble is measured in damage rather than in points.

How an escape actually works

The single most important principle is that escapes happen early. A submission has stages, and the number of available defences falls sharply at each one.

Hand fighting comes before everything. Most chokes require a specific grip. Denying the grip is easier than breaking it, and breaking it is easier than escaping the finished choke. A fighter whose hands are busy is defending the stage he is actually in.

Posture and alignment are the structural defence. A tucked chin denies a forearm the path under it. A bent elbow denies an extension attack its geometry. Connected hands keep an isolated limb from being isolated. None of these are techniques so much as refusals to supply the position the attacker needs.

Do not trade position for a defence you did not need. The classic error is giving up the back to escape a choke that was never going to finish, which converts a survivable position into a worse one.

Recognise rotational attacks by position rather than by feel. Since rotation is felt late, waiting for pain is not a plan. A fighter whose leg is entangled with an opponent's hips is in danger whether or not anything hurts yet, and the appropriate response is to resolve the entanglement rather than to wait for information.

The general shape of all of this is that an escape is a decision made several seconds before the escape itself. By the time a submission looks finished on a broadcast, the decision that mattered was two positions ago.

What a failed submission is worth

A submission attempt that does not finish is not nothing. Under the unified rules the judges assess effective grappling, and a genuine attempt at a finish is evidence within that assessment.

What it is worth is a matter of judgement rather than arithmetic, and the honest answer is that it depends on how close it came and what it cost. A deep, properly applied choke that the opponent survives is strong evidence of control. A speculative grab at a leg that is immediately shrugged off is not, and may count against the fighter who attempted it if it surrendered a position. How judges weigh those competing signals is set out in the explanation of how MMA judging criteria are applied.

There is also a cumulative effect no round score captures. A fighter who has defended three serious submission attempts has spent energy that does not come back, and he has spent it in positions where he could not do anything else. A round can be scored for the defender and still have cost him the fight two rounds later.

The submission as a threat rather than a finish

A large part of what submissions do in MMA never appears in a result at all, and it is the part a results column cannot show.

A credible submission threat changes what the opponent is willing to do. A striker who knows he can be caught in a guillotine throws fewer looping punches that put his head in front of his hips. A fighter who knows he can be heel hooked declines leg entanglements he would otherwise accept, which means he declines takedown defences that would have worked. The threat is doing work in every exchange whether or not it is ever applied.

This runs in both directions, and the symmetry is the interesting part. A grappler facing a dangerous striker cannot shoot a takedown from distance, because the entry is the moment he is most exposed. A striker facing a dangerous grappler cannot commit to a combination, because the end of a combination is when a takedown lands. Both fighters are being constrained by a capability the other has not yet used.

The practical consequence is that a fight between two complete fighters often contains very few submission attempts and is still being shaped by submissions throughout. A card with no finishes is not evidence that the grappling was irrelevant. It is frequently evidence that it was respected.

Why gi chokes mostly do not survive the transition

Grappling with a jacket produces an entire family of chokes built on collar and sleeve grips. A gi gives a grappler handles that do not exist on a human body: a stiff lapel that can be fed across a neck, a sleeve that can be gripped to pin an arm, a belt to anchor against.

Almost none of it transfers. An MMA fighter wears shorts and gloves, grabbing either is a foul, and there is no collar to feed. The chokes that survive the transition are the ones whose mechanism never depended on the fabric, which is why the rear naked choke, the triangle and the guillotine dominate MMA while large parts of the jacketed catalogue simply disappear.

One partial exception is worth noting, because it demonstrates the principle. A choke built on feeding a sleeve across the neck can sometimes be reproduced using the fighter's own bare forearm in place of the fabric. The grip changes completely, the mechanism is unchanged, and the result is still a blood choke. Where the fabric was structural the technique is gone; where the fabric was only a convenient handle, the technique adapts.

How to watch a submission sequence

Four things, and all of them happen before the part the replay shows.

Watch the hands, not the limbs. Almost every choke is decided in a grip battle that precedes the choke by several seconds. If one fighter's hands are controlling the other's wrists, the submission that follows was settled there.

Ask whether the position is secure. An attack launched from a position the opponent can still move out of is an attempt. An attack launched from one he cannot is a finish. The difference is visible in the hips.

Identify the mechanism before the name. Is the force extending a joint or rotating it? Extension means the defender has a warning and a clear decision. Rotation means he may not, and that the referee needs to be close.

Notice where the pressure sits on the neck. Across the front is slower and more survivable. On the sides is quicker and usually final. A commentator's phrase about a choke being "right on the throat" is describing a choke that is going the wrong way for the man applying it.

Do that and a submission stops being a sudden ending and becomes the visible last step of a sequence that was decided earlier. The rest of the combat sports explainers here take the same approach, starting with the mechanism before the vocabulary, and the companion piece on how contenders are matched against each other covers the other half of why a grappler ever gets to meet a striker in the first place.