The kimura as a control before it is a finish

The grip is simple to describe and awkward to do well. Your hand takes the opponent's wrist, your other arm threads under their upper arm and takes your own wrist, and the resulting figure encloses their arm at two points. The shoulder is the joint it acts on. The elbow is where it is held.

It has more names than most techniques. Judo's nomenclature places it in the ude-garami family, catch wrestling calls it a double wristlock, and most jiu-jitsu rooms call it a kimura. Where the last name came from is a story told confidently and told in several versions, and this article does not adjudicate between them. What matters here is that three traditions arrived at the same grip and used it differently, which is a hint about what it is for.

What follows treats the grip as an object in its own right, appearing from half guard, from side control and from a front headlock. Those three positions share almost nothing. The grip is the same in all of them and does the same job in all of them, which is why it is worth learning as a unit rather than as three techniques that happen to look alike.

Start with the claim, because it is the whole article. A figure-four on a wrist removes one of your opponent's arms from every other exchange happening at the same time. That is worth more, more often, than the submission is.

Consider what an arm does in a scramble. It posts, it frames, it underhooks, it grips a wrist of yours, it catches the mat when the base goes. An arm inside a figure-four does none of those things, and it cannot be withdrawn without dealing with the grip first, which takes both attention and a hand. You have not only taken an arm; you have committed a second one to getting it back.

This is why the grip travels between positions that otherwise have nothing in common. It is not a finish that happens to be available from several places. It is a control whose value does not depend on where you are standing, and the finish is one of the things you can do with it once you have it.

It is also why so many people learn it badly. Taught as a submission, it produces the pattern where somebody grabs the grip, immediately tries to rotate the arm, loses it, and concludes the technique does not work on stronger opponents. Taught as a control, the first question is not how to finish but what the grip prevents and what that makes available next.

From the bottom of half guard the grip solves a specific problem. The top player's near arm is either underhooking you or posting to stop the sweep, and both of those are arms you can reach.

Take the figure-four on that arm and the top player's options collapse. They cannot post on it to stop being rolled, they cannot use it to flatten you, and if they try to pull it free they generally have to give up the pressure holding you down. In practice the grip is a sweep more often than it is a finish, because the moment they defend the arm, their base changes and the sweep is there.

There is a second, quieter effect. Half guard is a position where the top player's chest and your chest are fighting over an inch or two of space, and an arm trapped in a figure-four is an arm that is not contributing to that fight. Reclaiming the underhook while somebody's arm is tied is a different task from reclaiming it while both of their arms work.

The risk in this set-up sits with the sweep rather than with the lock. Rolling somebody while you hold their arm in a figure-four means their arm arrives at the mat in a position they did not choose, and their body follows. Go slowly, particularly on the rotation as they turn, and let go of the grip if the roll is happening faster than you can control the arm through it.

From side control the grip is usually taken on the far arm, and the sequence people learn first is to step over the head to finish it. That step is the part worth being careful about, and it is worth understanding why the grip is useful even when you never take it.

Held from side control, the figure-four stops the bottom player from framing with the far arm, from turning away, and from getting the far shoulder off the mat. Those are the three components of most escapes from underneath. A player who holds the grip and does nothing else has made the position substantially harder to leave, and has done it with a structure that does not fatigue in the way that shoulder pressure does.

The finish is a rotation of the shoulder against a held elbow. This is where the specific injury risk of the technique lives, and it is worth naming precisely: the rotation has a short range and the attacker's body weight can move through it faster than a person can decide to tap, particularly when the finish is applied by stepping over and dropping rather than by turning the wrist gradually. Rotational loads are also easy to apply from a position where the person receiving them cannot see what is happening. Apply it slowly, in stages, and release completely on any tap without a final adjustment.

The corresponding rule for the person underneath is to tap early rather than to test it. A shoulder held in a figure-four with somebody's weight above it has already made the point. This is a technique to be introduced by a coach and drilled with a cooperative partner before it appears in a live round, and the round should be one where the terms were agreed before it started.

The third set-up is the one that most clearly shows the grip working as a control. When somebody shoots and you meet them with a front headlock, one of their arms is often inside, near your hip, and reachable.

Taken there, the figure-four is rarely a finish and is frequently everything else. It stops them building back up to their feet on that side, because the arm they would post with is committed. It gives you a handle to rotate them with, which is how the grip becomes a route to the back or to a change of angle. And it is a stalling point that costs them more than it costs you: your two hands against their one arm, in a position where they need both.

That is the pattern to notice across all three set-ups. The grip's value is not that it threatens a shoulder. It is that it converts a symmetric exchange into an asymmetric one, and then lets you choose what to do with the asymmetry.

It also generalises the safety rule. Anything that rotates a shoulder from a position where the other person cannot see it needs to be applied with less force than feels necessary, and abandoned rather than forced when the position moves. That applies more here than anywhere, because a front headlock is a scramble position and scrambles move.

Shoulder locks are broadly legal in adult submission grappling, and the shape of the exceptions is instructive. The ADCC rules list any arm bar, shoulder lock or wrist lock among legal techniques. The Unified Rules of MMA, defining small joint manipulation as a foul, state that fingers and toes are small joints while wrists, ankles, knees, shoulders and elbows are all large joints. United World Wrestling's grappling rules restrict guillotines, foot locks, calf and bicep slicers and knee bars in the under-13 and under-15 categories, and do not restrict shoulder locks there.

The IBJJF's rule book takes the opposite approach to the same problem, listing prohibited holds in a table organised by age and belt division. The numbered items in that table include the omoplata and the wrist lock, which are the shoulder lock's near neighbours, along with a set of leg attacks and spinal locks. The table is a grid, the restrictions differ by division, and the only sensible instruction is to read the version that applies to you.

Judo restricts joint attacks more narrowly than any of them. The IJF's rules describe the sport as one where, apart from the elbow joint, techniques are executed in the sense of the articulation and never in hyperextension — a constraint on the direction a joint is loaded rather than a list of joints that may be attacked. The lock itself sits in judo's ude-garami family, which is a recognised kansetsu-waza. What changes between documents is how a technique may be applied, not whether the shape exists.

Nothing in the research literature is about this grip. There is no measurement of how often a kimura control leads to a sweep, a back take or a finish, and no comparison between rooms that teach it as a control and rooms that teach it as a submission.

The nearest relevant finding is indirect. Spanias, Kirk and Ovretveit, analysing 26 no-gi submission-only matches, found that time spent in dominant positions correlated with winning by an upper-body submission and did not correlate with winning by a lower-body one. Upper-body attacks in that sample arrived through position. A grip that improves position and threatens a shoulder sits on the same side of that line, which is consistent with the argument here and is not evidence for it.

So the case for treating the kimura as a control is mechanical and it is testable in your own training rather than in a citation. Take the grip in a positional round and forbid yourself the finish. Count what becomes available. If the answer is nothing, the argument is wrong for the way you play, and that is a more useful result than a statistic somebody assembled from a different sport under a different ruleset.