Thoughts On Unarmed Defense Against A Knife

There has been a recent spate of slashing attacks in New York City which piqued my interest in strengthening my knowledge of unarmed defenses effective against a knife attack. There are various techniques that are taught in martial arts or self-defense classes, or that can be found in books, articles or videos.  I decided to test/practice some techniques or common defenses to determine their effectiveness. Now, still with very sore wrists and forearms, I have some thoughts about unarmed defense against a knife attack.

The effectiveness of a knife is often dependent on surprise and/or being able to confine the victim. Since it is a contact weapon, simply keeping one's distance seems the best defense. However, backing away or retreating ("run away!" as King Arthur yelled in a Monte Python movie) may not be possible if the attack is sudden or if, for some reason, you are constrained against leaving. (It may not be a physical barrier that keeps you from leaving, but the need to protect a loved one. I would refer you to this article of an attack and attempted rape of an example of the latter situation).

For our tests we used training knives manufactured by Cold Steel, one being a facsimile of Cold Steel's Recon Tanto and the other was of a K-Bar style knife (the Leatherneck S/F trainer).

Initially, we practiced with basic stabbing attacks, holding the knife in a saber grip. The intent of the attacker was to deliver a stab to the midriff at approximately waist level (i.e., between the pelvis and the sternum), although it varied up to stabs at the area of the heart.

The first defensive technique we attempted was one that I had been taught many years ago. In that case, to defend against a stab with a forward grip, the instructor had taught a technique that required you to slap the back of the knife hand with one hand (your left hand if the attacker was using his right hand), while simultaneously slapping the attacker's wrist from the other side. The theory was combination of impacts would flex the hand, and cause the knife to fly out of the hand. Although not perfect, it worked well enough in the dojo with students using Sharpies and highlighters to stand in for knives. However, as expected, it did not translate well to the training knives and an aggressive opponent. Not only is the technique somewhat complex to successfully accomplish, but it requires you to place your body in line to the advancing knife. So a failed defense leads to a disastrous outcome. I would recommend against using this technique.

Next we tried standard blocks. The problem with these, we discovered, is that it is simple to disengage from the block, dip the knife hand under the blocking arm, and continue the thrust or initiate another attack. Not too surprising since this is a basic maneuver in fencing.

Next we tried various techniques for grasping or trapping the attacking arm. These showed more promise, but almost always degenerated to a "catch-as-catch-can" wrestle for the knife. Moreover, a grip on the knife arm using but a single hand was easily broken. Even if an arm was trapped, it was generally possible for the attacker to transfer the knife to his other (free) hand.

What seemed to consistently work the best was to use the hand opposite the attacker's hand (e.g., if the attacker was using his right hand, the defender would use his left hand) to grab the wrist of the attacker (grab slightly underneath so you can more easily twist the attacker's arm). This does two things: it requires the defender to move his body out of line of the stab (so even if the defender failed in the defensive maneuver, the knife would probably still miss the defender), and initiates control of the knife hand. Simultaneously, the defender grabs the attacker's arm with his other hand about the wrist of the knife hand. The attacker and defender are essentially, at this point, standing almost side by side. During the same maneuver, the defender's arm first used to grab the knife hand (in this example, it would be the defender's left arm) goes over the knife arm of the attacker (in this example, the right arm), with elbow into the side of the attacker's chest. The defender drops down which should drive the knife straight into the ground. At that point, the defender has a good lock on the attacker's right arm, the attacker probably cannot get his left hand over to the right hand, and the defender can force the attacker fully to the ground, remove the knife from the attacker's hand, or jump back up and run away.

So, in summary, assuming an attacker using his right hand as his knife hand:

1.  Attacker stabs with right hand;
2. Defender uses left hand to grasp the attacker's right wrist, sidestepping;
3. Defender's left arm goes over attacker's right arm, with the Defender's elbow going into the side of the attacker's chest (very painful if you happen to strike the right spot); 
4. Defender's right hand is also used to grab the attacker's right wrist;
5. Defender drops, diverting the attacker's forward momentum, and driving the knife into the ground. 

With an attacker using the left hand, the sides are simply reversed (i.e., the defender uses his right hand and arm).

We tried this technique against other types of attacks, such as slashing attacks where the knife was held in a reverse grip (edge out), and a reverse grip (edge in). In these, we tested it against a horizontal slashing attack, such as might be made against the torso, neck or face. The defense described above seemed to work well against the attack where the hold was a reverse grip (edge in), but appeared to be less successful against one where the hold was a reverse grip (edge out). In the latter case, we could successfully drive the attacker to the ground, but it was apparent that were a real knife used, the defender would likely suffer cuts to his arm or hand.

This technique was not well suited to an overhand, reverse grip stab straight downward. In that case, the classic technique of using both hands to grab and redirect the thrust was effective--actually, very much so since the natural arc described by the attacker's arm means that the trajectory of the knife will generally be into the attacker's own leg, groin, or abdomen.

I have not had an opportunity to test the technique against a diagonal slashing attack.

What I liked about this technique was that it was actually fairly forgiving of errors. As noted, even if you have failed to grasp the attacker's wrist, you have stepped out of line of the attack. If you don't get your arm over the attacker's arm, it is not as strong a lock but will still work. If you don't get your second hand around the wrist, you still very well may have enough momentum to drive the knife down into the ground or throw the attacker off balance.

I know that I've seen this technique illustrated before in books or in a video, but I cannot remember the source. Anyway, I welcome any comments, criticisms or suggestions.