Self-defence against knife attacks: a full review

We'll all agree that in order to adopt realistic methods of defence against knife attacks, we first need to understand how an attacker is most likely to use the weapon.

But how do we get that understanding?

The truth is, we mostly rely on other people (the 'experts') to give us the answers.

The problem is that the Martial Arts / Self-Defence industry is plagued by misconceptions and fallacies.

Knife Attacks and Self-Defence

That's why I decided to do a research. I wanted to understand the dynamics of real knife attacks and I wanted it to be based on evidence.

Luckily, in recent years CCTV and phone footages have provided us with abundance of real-life examples we can learn from.

Data were readily available for anyone who was willing to carry out the (tedious) analysis.

And today I'm going to share with you the results of my investigation where some interesting findings were uncovered.

This article is structured in three parts:
  • Part 1 presents the keys points of the analysis of 150+ knife attacks (video material)
  • Part 2 addresses commonly-debated issues such as awareness, avoidance ("running away"), compliance, the use of weapons (incl. improvised weapons) in the context of knife attacks
  • Part 3 covers empty-hand techniques in view of what's discussed in the first two sections
This a substantial article which took months of painstaking and meticulous work, so I hope you'll enjoy reading it .


The following article contains graphic content that could be disturbing to some.
Viewer discretion is strongly advised



Defence against knife attacks remains one of the most contentious parts of Martial Arts and Reality Based Self-Defence (RBSD).

The reason for this, in my view, is that very few people actually have any substantial experience of this type of violence. And among those who have that experience, even fewer are actually keen to talk about it.

So we're left with people who have been "twice attacked" by a knife-wielding nutcase. Although, I'll listen to these people -any experience is worth hearing- two incidents don't make you an expert on a question.

Think about it.

Would a major car-maker hire you as a car-safety expert if you just added "Have been twice in a car crash" to your CV/resume?

It takes more than that.

The issue of expertise, with regard to Martial Arts, was addressed in a very discerning way by Wim Demeere in his 2014 article "Are you really an expert?".

In the Martial Arts / Self-Defence industry, unfortunately, we too often hear "appeals to self-authority" where the speaker expect you to believe him based uniquely on his (often unverifiable) pedigree.

In other words, "Trust me, I'm the expert!"

What is sorely lacking is comprehensive, analytical and thoroughly researched studies. Facts and Statistics.

One notable example of this type of evidence-based approach is the work of John Correia from Active Self Protection. As of today, Correia has carefully analysed over 175 incidents (robberies, muggings, attacks, etc) caught on camera (see his youtube channel).

To a lesser extent (26 incidents), Rener and Ryron Gracie have done the same but with a focus on grappling situations. You'll find the videos (along with the analysis of 62 UFC fights!) on their youtube channel GracieBreakdown.

The detail examination of CCTV and surveillance footages of knife attacks offers invaluable lessons.

And here is what I've learned:



Data for this research has been gathered and compiled from 150+ knife incidents caught on surveillance and phone cameras.

All the videos are accessible on my youtube channel in the following playlist: Knife Attacks (CCTV).

More videos will be added through time and statistics will be updated.

Let's get started,

71.1% of knife attacks are led with the empty hand

The fact that, during a knife attack, aggressors usually lead with their free hand, while keeping the knife close to their side, was first pointed out in 1988 by Don Pentecost in his contentious book Put'em down, take'em out! knife fighting from Folsom prison.

Back in the 80ies, Pentecost simply and brutally shattered a number of popular myths and preconceived ideas about the use of knives in real life situations (see a review of the book on Wim Demeere's blog).

Most commonly (71.1% of the time) aggressors will lead the attack with their empty hand, effectively shielding the knife, such as in the following video (segment starts at 0:21):

As Pentecost also clearly stated, the empty hand is not a dead, "paralyzed", hand and attackers will use it to strike or, more commonly, to grab the victim.

Indeed, when aggressors lead with their free hand, 80% of the time they will also use it to latch on the victim:

This use of the empty hand (also known as 'leveraging arm') greatly changes the dynamic of the fight. Particularly because your first reaction, as a victim, will be conditioned by the movement of the aggressor's empty hand (more on that in Part 3).

But the prime lesson though is that most of what is taught in the Martial Arts and RBSD industry (i.e. attacker leads with the knife, no grabbing, no forward pressure) doesn't apply to the vast majority (70% +) of real life knife attacks.

Most knife attacks are ambushes, not duels

For the same reasons bad guys will use a force multiplier such as a knife -i.e. they don't want a fair fight but an easy prey- they'll make a surprise attack on you from a 'concealed' position or with 'concealed' intentions.

And you probably won't see it coming.

As you can observe in this CCTV footage, the saleswoman was ambushed and cornered. She stood absolutely no chance.

The aggressor didn't threaten the victim. Instead, he kept his weapon concealed until he struck the woman. Clearly, his intention was not to try to get what he wanted through intimidation/coercion but to eliminate the saleswoman.

“Victims who survived a violent confrontation against a knife-wielding assailant consistently reported that they were completely unaware of the existence of the weapon until after they had suffered stab or slash wounds. In essence, these survivors of edged weapon attacks state that they believed they were engaged in some sort of fist fight; only later, after sustaining injuries, did they realize that the assailant was armed.” Imi Lichtenfeld (Krav Maga founder)

In 80% of the cases I've analysed, the knife is kept hidden until the very last moment. That is until the attack is launched.

Aggressors will try to distract the victim and wait for a good window of opportunity to strike and won't hesitate to attack the victim from the back.

Situations with multiple attackers seem less common with only 11.4% of the incidents I've analysed.

The following video (segment starts at 0:31), shows the ambush of a store clerk:

In a typical way, the victim is 1) distracted, 2) cornered, and 3) the attack is launched at close quarters.

70.6% of knife attacks are launched within 3 feet of the victim

Knives are short range weapons, so it doesn't come as a surprise that 70.6% of knife attacks start at conversation range.

It's important, though, to stress that "within 3 feet" really means "at arm-length or less":

This leaves you with very little space (and time) to react!

Such a small 'reactionary gap' means that it's almost impossible to stop the first stab if you're not expecting it.

This is why it is so important to maintain distance and keep control of the space when you're in a confrontation (see here an interesting article on Space Management in Self Defence)

As aggressors so often lead with their free hand (usually the left one), victims end up being stabbed in the neck, like in the video above, or to the chest on their left side (where the heart is located!).

Knife attacks are fast and furious

Another consequence of such close-range attacks is that victims tend to fall as they move backwards trying to escape their aggressor.

This is the case in 55% of the incidents I have analysed:

This way of charging a victim, as seen in the video above, is known as a "prison yard rush" (or "prison knife rush"). It was made notorious in the Martial Arts industry by Don Pentecost in 1988.

In his then-contentious book, Pentecost further points out that someone who is attacking you with a knife is trying to kill you.

They won't hold back, they won't hesitate.

They will go after you like mad dogs.

"Research tells us someone willing and ready to carve you up like a thanksgiving turkey is far different than someone with a gun [...]" Hank Hayes (Knife Defence 101)

Here is an illustration of these points with an interesting analysis by John Correia from Active Self Protection:

As you can see in this video, the aggressor goes full on with ruthless determination in an attempt to overwhelm his victim and cause maximum damages as quickly as possible by whatever means necessary.

Knife attacks don't last long

Indeed, the average incident time (i.e. arithmetic mean) for knife attacks, from the moment the attack is launched to the moment it stops, is 23 seconds.

"The time frame of a knife attack is usually very short - it is often over in a matter of seconds" Don Pentecost

The median time is 14 seconds which means that half the attacks last 14 seconds or less.

But 80% of all attacks last less than 32 seconds:

The graph shows a sharp increase in numbers up to 23 seconds indicating that most attacks (70%) last 23 seconds or less.

After that point, the curve starts to flatten and reaches 90% at 59 seconds.

It's interesting to note that, although it takes 9 seconds for a 10% increase to 80%, it then takes 27 seconds to complete another 10% and reach 90%.

What this means is that if a knife attack lasts 23 seconds, you have a fair bit of chance that it will go on only 9 more seconds. But once a knife attack reaches the 32 seconds mark, the same "fair bit of chance" may mean an extra 27 seconds.

In other words, if a knife attack goes on for longer than 32 seconds, it is more likely to last a lot longer.

Did you notice how the curve seems to go up again around 45 seconds instead of plateauing as expected?

Let's visualise the data differently,

The following graph shows the number of knife attacks plotted against time. To make it clearer, I just show the trendline:

There's a peak in the number of attacks around 7 seconds with 25.2% of all attacks lasting between 5 and 10 seconds, and half of all attacks lasting 14 seconds or less.

The fear of being caught is likely to be an important factor that keeps knife attacks very short.

Obviously, the longer an attack is carried, the more likely someone -including the police- will intervene.

In that regard, it is interesting to note that 55.9% of all attacks are stopped by the intervention of a third party.

But here is the interesting part,

From the 7 seconds mark, the number of attacks goes down sharply to reach a minimum at about 45 seconds and then goes up again.

What the graph suggests is that there's a tipping point around 45 seconds after which a knife attack will tend to go on much longer. Most commonly an extra 14 seconds. More than half of the "over-45-seconds" attacks last between 53 and 66 seconds.

Although it is hard to infer any solid evidence from such a small number of  cases, it seems that third party intervention occurrences are much lower in the 45-seconds + group than in general.

In other words, these attacks lasted longer because no one intervened.

Unsurprisingly, attacks by 'psycho lovers' -who don't care much about being caught- seem to be concentrated in this group as well.

Although the average duration of a knife attack is 23 seconds, it is important to keep things in perspective.

In that short amount time, the average attacker will stab you at a rate of 5 to 7 times every 5 seconds!

Knife attacks are more often performed with quick, short, repetitive stabs at different angles

Commonly, there will be a first wave of stabs during which the attacker, taking advantage of the surprise, will land between 5 and 10 stabs. Then, as the victim fights back trying to escape, the stabs will be spaced-out.

The more you get cut or stab, the more chances a vital organ, such as the heart, or a major blood vessel, such as the carotid artery (neck), will be hit which would result in a quick death.

As you can witness in the following surveillance video, things can go really fast. The first victim is stabbed twice between 0:03 and 0:06, and totally collapses less than 40 seconds later!

Sure, many people have survived a greater number of cuts and stabs. But the unavoidable truth is that you only need one to die!

This is an important point because most attacks are not a 'single straight thrust' (i.e. bayonet thrust type of motion) or a 'wide sweeping stroke' (i.e. slash).

Knife attacks are predominantly executed with quick, short, repetitive stabs at different angles (e.g. switching from low, upward, stabs to the chest, to high, downward, stabs to the neck).

Also known as the "sewing machine" and "prison shanking", short rapid stabs are very difficult to stop because there is not much time and not much space to 'deflect-and-redirect' the attack or 'block-and-strike'.

The majority of attacks are carried out using a regular grip (58.8%) but the reverse grip (also known as "icepick grip") is, at 29.9%, more frequent than commonly thought.

In a small number of cases, around 6%, the aggressor changes his grip during the assault, moving from regular to reverse or vice versa.

It's worth noting that I've not seen any case of hand-swap (i.e. knife being moved from one hand to the other).

What have we learned

In this first part, we've seen that knife attacks are most commonly ambushes, launched at short range (within 3 feet of the victim).

They are extremely violent; carried with speed and determination within a very short span of time (i.e. they don't last long).

Attackers will lead with their free hand, effectively shielding the knife, and will stab the victim repetitively with quick short thrusts at different angles.

Attackers frequently grab and push the victim who quite often will fall to the ground.

In other words:

  1. you will be taken by surprise and you will be overwhelmed by fear and aggression
  2. you won't see the blade before the attack is launched
  3. you very likely won't be able to run away and avoid the attack
  4. you will have very little time and space to react and deploy a counter-attack
  5. you probably won't be able to stop the first stab(s) so, yes, you will be you will be cut and stabbed a number of times but you might not even realise you've been stabbed (which is why you should always check yourself for wounds after a physical confrontation)
  6. you probably won't have time to draw your own weapon (gun, knife, kubotan keychain, pepper spray, etc); at least not before being stabbed a couple of times
  7. you will be likely moving backwards, your balance will be compromised, and you'll probably fall to the ground
  8. your movements will be restricted, your fine motor skills will be gone, you won't able to access the knife bearing arm easily
  9. any technique that is based on smooth arm deflection and manipulation has very little chance to work
  10. any technique that rests on the assumption of a single straight thrust ('full-stop one-step zombie attack') or a wide sweeping slash has very little chance to work

These results should definitely inform our training methods and I will expand on that point in Part 3.

But the nagging question at this point is "how do people survive knife attacks then?"



Some of the most common pieces of advice you can get from self-defence experts and martial artists with regard to knife defence revolve around 'avoidance' ("run!", "Don't be there in the first place"), 'awareness' ("look for the dodgy guy"), the need for a force multiplier (i.e. 'weapon') and 'compliance' in case of mugging situations.

Although these are definitely good general pointers, they also tend to over-simplify the reality of knife attacks.

As an instructor, I'm not satisfied with guidelines. I want to understand the full dynamic of situations. The ins and outs. So, I'll always question the status quo.

And there's a lot to learn this way.

Let's start with 'avoidance'

Avoidance & Situational Awareness

"Don't be there in the first place", "avoid dodgy places and dodgy people", these are commonly heard pieces of advice.

The mother of 'avoidance' is 'situational awareness'. Clearly, being able to anticipate danger can go a long way in keeping you safe (see how to improve your situational awareness with your brain natural capabilities).

But as I pointed out in another article, troubles may find you despite your best efforts to avoid them.

    "No matter how good your awareness, you can’t remain switched on all the time, you can’t always be at the top of your game. There will be times when you are pre-occupied or ill or injured or whatever, and these are the times when you are usually attacked. Attacks always come when you least expect them and usually when you are not ready for them". Neal Martin (Combative Mind)

You can avoid dodgy places and dodgy people, but not all attacks occur in contexts that could be deemed dangerous or at least more prone to incidents (e.g. night clubs, security work, police work, etc).

Actually, an awful lot of knife attacks take place in shops where the unlucky victim happened to be working.

Attackers don't necessarily come for the victims; they might come for the money or anything else, and the victim just happen to be on their way. Bad luck (see below at 0:20).

You could also cross the path of a deranged person like below in the streets of New Orleans:

Although it's not always possible to avoid situations and attacks, violence is a long road and in a given situation you might be able to read the signs of an impending attack.

Knife Awareness

Being able to quickly recognise an incoming attack before it actually happens, can make a big difference.

Look at the following video, can you see the main cue (attack occurs at 1:36):

Yes, the aggressor checks the entrance door a lot but the main thing is that he kept his right hand in his pocket and his body was always bladed like he wanted to keep the knife away.

Now, look at the following footage, can you see the cues?

See how the guy with the brown jacket angled himself and kept his hand down along his leg when he moved (at 0:20) towards the young lad in a dark sweater just before striking him?

His attack was somehow disrupted a bit when his younger opponent weakly punched him in the face. Brown jacket also briefly struggled with his weapon (a pair of scissors according to a report). But all this didn't stop the attack and he managed to stab the teenager in the neck.

Had the kid recognised the cues, he could have kept his distance and avoided a very serious injury.

Can you now recognise the signs in the following footage (starting at 0:41)?

Two guys can be seen arguing. The man with the dark t-shirt keeps his right hand down and behind his leg. Because of the darkness we can't see if he's holding anything in his hand but his distinctive position strongly suggests he does.

As the argument goes on, he starts angling his body (0:45) -shifting his right side back- with his hand still down and hidden behind his leg.

At this point (0:47), his body language says "I'm going to hit you":

He used a hammer but it could have been a bottle or a knife.

It's important to stress that if, during a verbal confrontation, you don't see your aggressor's hand(s), you should assume he has a weapon (knife, screwdriver, stick, bottle, hammer, etc).

The question, now, is: at such as short distance can you run away?


Given a good head-start (see more about that below, in the section "Weapons") and a quick reaction, running away can be a successful tactic as seen in the following surveillance video:

Note that the guy who got away had a decent head-start and a bit of time to assess the situation.

He was also lucky to have an exit route readily available. He managed not to fall (unlike what happens in half the incidents) and the bad guy chasing him could not catch-up.

Now, remember that most knife attacks are ambushes;

They're usually launched in a semi-closed environment, within 3 feet of the victim who is cornered.

This means that there will be few exit routes and they won't be easily accessible. On top of that, you won't have a great head-start and the bad guy will be on your heels right away.

Victims are often distracted and don't see it coming. So you will be taken by surprise and won't have much time to figure out your options.

But can you still escape once you've been engaged?

In an interesting vlog, Scott Babb from Libre Fighting describes some of the experiments he and his team had been conducting back in 2012.

On a number of points his research on knife attacks confirms my own conclusions:
  • attacks most commonly happen in a semi-closed environment, 
  • within 3 feet of the victim, 
  • aggressors don't brandish the knife ahead of time
  • aggressors latch on to the victim with their free hand
  • the stabbing rate is 1 to 2 thrusts per seconds ("5 to 7 times per 5 seconds")
  • most wounds are inflicted to the left side of the victims (abdomen, lungs, throat)
What Babb found out during his experiment is that against a knife attack, the most successful strategy is to prioritise escaping and running away.

In other words, people ended up being stabbed fewer times when they focused on escaping than when they were trying to disarm/stop the aggressor.

This is a significant result which should inform every knife defence training programme.

It's a shame, though, that Babb doesn't show more of the experiment. I would love to know which techniques they've tested, how many times they pressure tested each technique ('sample materials') and how many people participated ('subject pool') other words, what was the size of their sample?

For statistical analysis, the size of the sample is important as small samples are not always reliable and the results might not be statistically significant.

Particularly, I understand that the people who participated in the experiment were all people from Libre Fighting. So they were all trained in Filipino Martial Arts (FMA) which means that they were all familiar with the use of a blade.

In other words, the 'subject pool' of aggressors was flawed to a certain extent. This is probably why, on the latter phase of the experiment, Babb measured a stabbing rate of 16 to 26 thrusts every 5 seconds (3 to 5 times higher that what can be seen on videos of real knife attacks).

I would also like to know how 'lethality' was measured/decided (the devil is in the detail).

Did the experiment stop when victims manage to disengage or did the aggressors try to run after them very much like what might happen in real life?

It's an important question because you can get lots of punishment during the 'escape phase' (i.e. when you turn your back to run away).

Most of all, I would love to see what people actually did when they put the priority on escaping.

How did they escape? Did they punch? Did they push? Did they wrestle? What did they do? Chances are, they used some fighting techniques.

Sadly, we have to admit that, in most cases, victims have to fight their way out and it's not a straightforward business as we can see in this CCTV footage.

Once engaged, you would have to fight too and the difficulty is to recognise when you have a good window of opportunity to escape. Even though you might not have a choice because the bad guy might just not be too keen on letting you go and/or the circumstances might not be favourable.

To his credit, Scott Babb highlighted a major point: escaping should be the priority. Knife defence training should emphasise that aspect and techniques should be used to facilitate the escape.


When threatened by a knife-wielding mugger who's asking for your belongings ('territorial violence'), compliance is probably the safest course of action (note that the opposite is true in cases of sexual assaults).

    "a smartphone can be replaced, your life cannot"

If they use the knife as an intimidation tool, they probably don't intend to hurt you if they don't have to; otherwise, as we've seen, they would've already stabbed you.

But not always.

In the surveillance video above, the victim was stabbed several times despite being quite compliant.

Although this is quite "evil" from a moral standpoint, accepting this argument as the only explanation for what happened seriously limits our understanding of criminal behaviour.

As you probably noticed in the video, the aggressor's accomplice who drives the car didn't totally stopped the vehicle and the door was kept open.

Obviously, the pair was in a hurry.

And that's a very important point.

Time is always against the bad guys. Cops could be around the corner, someone could intervene, the victim could pull a weapon, etc...

In other words, bad guys are always in a rush and their patience is limited.

They won't take a chance to let the situation get out of their control and they'll stab you if they have to in order to get things going faster because they want to be out as fast as possible (and this, by the way, also means that if you can defend yourself, they probably won't stick around too long).

So, whatever you decide to do (i.e. to be compliant or to fight back), I would advise, in view of what we just saw, to show committed compliance right away. Don't be as hesitant as the victim in the video.

If you decide to fight back it will give you some cover and surprise will be on your side.

Whether you decide to fight back or not, you should always expect the worse and be ready to block a stab.

Again, it's important to repeat that if you don't see your aggressor's hand(s) during a confrontation, you should assume he has a knife (or any other blunt/edge weapon for that matter).

Knowing basic self-defence against knife attacks is important because you can't always run away and compliance doesn't always guarantee your safety.


The "Glock" answer is a very common reply to the question "What do you do if someone attacks you with a knife".

One problem is that weapons such as knives and guns are not legal everywhere. In most of western Europe, for example, you can't carry a gun and there are strong restrictions on knives.

Another problem is deployment.

If someone ambushes you and launches an attack within 3 feet of you, would you have time to draw your weapon?

Here comes the "21 feet" rule

Research carried out in the early 80ies by police sergeant Dennis Tueller (Salt Lake City PD), showed that in the time it takes a trained officer to recognize a threat, draw his sidearm and fire 2 rounds at center mass (i.e. around 1.7 seconds), an average subject charging at the officer, from dead stop, can easily cover a distance of 21 feet.

This means that you need enough time to draw your weapon and operate it properly before your aggressor reaches you. But time is related to space ('distance travelled' in this case):

Time to reach target = Distance travelled / travelling Speed

In other words, you need enough space/distance to react to an attack (i.e. draw your weapon and operate it).

This space is your 'safety distance'. That's the minimum distance from your aggressor you will be able to react properly to his attack. This is why it's also called 'reactionary gap'.

If the attack is launched within your safety distance -i.e. a distance that is shorter than your safety distance- you won't have enough time to draw and operate your weapon which means that you will need to know empty-handed defence.

In the following video, Eskrima martial artist Dan Inosanto demonstrates how fast, and thus dangerous, an attacker with a knife can be:

The safety distance a trained police officer needs, according to Tueller's research, is 21 feet.

First published in SWAT magazine in 1983, the results of this experiment have since been known as the "The 21ft rule".

Note that in this type of situations, running away ('avoidance') is not an option because, as you can infer from the video, the aggressor would be on you before you can create enough space between you and him.

It's also important to keep in mind that speed can be increased by 'rage' and 'stimulants' such as certain drugs and alcohol.

Similarly, your reaction time can be increased by a number of factors (e.g. fatigue, fear). This is why the 'safety distance' has been extended to 30ft by a number of specialists.

Weapons of fortune

Weapons of opportunity can be great equalisers in a fight.

Any object that would extend your range, thus keeping the knife at a distance, can be used (see here for some interesting suggestions and discussion).

Beware that, against a knife-wielding attacker, car keys and pens are not as good weapons as commonly held.

The biggest issue with improvised weapons though is "deployment" (check the other issues in The Truth About Improvised Weapons For Self Defence by Neal Martin).

As we've seen, most attacks are ambushes so you might not have much time to look for a weapon of fortune and grab it.

This is why a little preparation, training and awareness can go a long way.

Here are some suggestions:
  • a metal pipe,
  • a lamp,
  • a chair,
  • your backpack,
  • a roll-up magazine
  • your belt
  • bicycle chainlock

Your belt, in particular, is a great improvised weapon against an attacker armed with a knife. Readily available, a belt can be used as a slungshot, particularly efficient if it has a strong buckle.

Here is a good tutorial video by Nick Drossos on how to use a belt as an improvised weapon:

Let me insist on one point, though: you must know how to use your weapon of choice.

It doesn't always work as well as expected as you can see in the following surveillance footage:

At 0:13 the guy in red pulls his belt and attacks the knife-wielding man (in white). White-shirt uses his free (left) arm as a shield and counters with an over-hand stab to the side of the neck (0:17). That's a very bad injury and you can see red-shirt bleeding profusely in the rest of the video.

There are important lessons here:
  • Don't jump into an armed attacker
  • Use the weapon to keep him at a distance (which red-shirt does properly after being stabbed!)
  • Don't overlook training! Training and preparation are key

Now, let's see what are your options if you can't run, can't deploy a weapon to keep them at a distance and have to fight bare handed.



We've reviewed lots of information regarding knife attacks so far and it's fair to say that someone armed with a knife is at a huge advantage so it's best to avoid physical confrontation if possible.

This being said, we've also seen that a knife attack is decided by the aggressor and it will happen on his terms.

    "If someone with a knife tries to attack you, then guess what? You're in a knife fight whether you like it or not."

    At such a short range it's not possible to avoid the fight. Once you're engaged you'll need to defend yourself and you might not have the time and opportunity to deploy a weapon.

    Although priority should be put on escaping, you'll need to fight your way out while minimising damage as much as possible because each new cut or stab could be fatal.

    At this point, it's important to say that there's no silver-bullet empty-hand technique to thwart a knife attack. No technique is fool-proof.

    Due to the variety of possible situations, my opinion is that you'll need to know a range of empty-hand techniques that will allow you to improvise and adapt to different situations.

    Techniques should therefore be seen more like tools/skills that you can use to adjust to various situations.

    Let's now see some techniques.

    The 'free hand' issue

    In his 1988-book, Don Pentecost insisted that aggressors won't lead with the knife when they're within striking range (i.e. within 3 feet) and they will use their free hand.

    "The fact is, an experienced fighter will have his knife hand held close to his side until the empty hand has done its job and created an opening" Don Pentecost

    This is something we can clearly see in CCTV/surveillance videos. More than 70% of knife attacks happen that way.

    The study of these footages reveals that even inexperienced aggressors tend to use their free hand instinctively to latch on their victims.

    They won't necessarily hit you with their free hand as Pentecost contends in the case of experienced fighters, but the use of the 'leveraging arm' makes the whole situation a lot worse for the victim.

    A leading free hand will change the dynamic of an attack in several ways:

    1- it will condition your initial reaction

    You will react to what you see first, and that will be the free hand coming towards your face; not to the knife that will still be concealed.

    2- it will allow the aggressor to keep you at a distance

    Leading with the free hand will allow him to keep you at the right distance to prevent you from (i) reaching and controlling the knife, (ii) reaching him (e.g. punches).

    The blade will give him additional reach that he will use to stab and cut you, inflicting maximum damage.

    3- it will give an aggressor some control over you

    Your aggressor will try to grab you which will make it harder for you to escape.

    4- it will allow him to apply forward pressure

    The aggressor will be steaming in. Leading with his free hand will allow him to apply forward pressure (i.e. to push the victim). The victim will be going backwards which is why over 50% of victims fall on the ground while trying to escape.

    To be able to deal with a knife attack, you need to find a response to these issues.

    There's very little material online with regard to this specific situation.

    Actually, the most surprising thing for me during this research on knife attacks was that this type of attacks -which are the most common- is almost never addressed by martial artists and self-defence instructors.

    So I went to see a couple of people I knew in London who had both extensive knowledge of martial arts and combatives, and also first-hand experience of knife attacks.

    The first person I met to talk about this situation was Stewart McGill.

    Founder and Chief-instructor of Urban Krav Maga, Stewart has his martial art background primarily in Judo and Goju Ryu karate (3rd Dan) and was a Civilian/Law Enforcement Instructor with several Krav Maga organisations (including IKMF under Eyal Yanilov). He's also a senior instructor with the British Combat Association.

    Here is one option we came up with:

    Here are some other options for a slightly different situation (but see at 4:38), by Stewart McGill and MMA/Vale Tudo fighter and BJJ world Champion Leo Negao:

    Then, I went to see my friend and fellow instructor David Kyriacou who grew up in one of East London roughest boroughs and experienced several violent situations involving knives.

    David is an accomplished martial artist (primarily TKD, Muay Thai and more recently BJJ and wrestling) and a Krav Maga Instructor. He's worked at the door of some of the most unpleasant clubs in the capital for more than 10 years. So it's an understatement to say that David knows violence (see his story here).

    Here is his take on the situation:

    As I said earlier, very few instructors have actually addressed this particular situation, so it's worth having a look at their solutions.

    I like Nick Drossos' genuine approach to self-defence and I follow him with interest but I'm not sold on this one:

    Drossos raises good points regarding the necessity to trap and isolate the knife and his move is functionally simple but I can see a number of issues:

    Firstly, it might be hard to apply a strong bearhug to a big guy. It's also relatively easy to get out of such a bearhug, so it's fair to say that many things could go wrong for you. Thirdly, you need to know how to finish it.

    Let's be honest, once you 'bearhug' the aggressor, you're not going stay there to cuddle and hope that calms him down, are you?

    You'd need something like this, or a more violent version of it rather:

    Not an easy feat against a much larger, adrenalized aggressor who's holding a knife with the intention to carve you like a Thanksgiving turkey.

    Besides, you have to do the move before the guy secures a firm grip on your shirt. In a typical attack (no warning) the knife gets into action shortly after the initial move, so you need to react fast.

    Renowned martial artist Ron Balicki, who extensively trained under Dan Inosanto, offers some interesting insights on the issue here:

    Whether you like his concept or not, whether you agree with him or not, we have at least to acknowledge Balicki's genuine effort to figure out a solution to a very difficult situation.

    I've personally tried a variety of things and I can't see this option working for several of reasons:

    The first one is that the impact of the "shoulder grab" is very limited particularly if the aggressor changes angles and goes overhand to strike the neck which is quite common.

    The second reason is that the technique relies on fine motor skills and complex moves that will be hard to pull off in real conditions.

    The third reason is that your position is not improved so you'll be going backwards which makes the whole punching thing a lot less efficient/disruptive than shown in the video.

    "You may get lucky and deliver a knockout blow as you jump in, but this would be unlikely. It is very difficult to knock someone out when they are completely adrenalized and riled up. They probably won’t even feel the hit and they will carry on attacking like nothing has happened." Neal Martin (Combative Mind)

    Besides, the aggressor's arm is in the way and might shield his face. Also, if he has longer arms than you, you might simply not be able to reach him.

    That's also why I don't see the classic Krav Maga 360 defence working in this situation either.

    Knife attacks are extremely violent and the efficiency of punching and kicking in these situations should not be overestimated as can be seen in the following video (punches at 2:20; kicks at 3:20 and 5:14):

    Finally, a number of martial artists advocate elbow/shoulder manipulation and armbars (see here for example). Pressure testing, though, shows that joint-locks of this type are really hard to pull off against a non-compliant, aggressive attacker, so I'm not a big fan.

    We've covered the most common type of knife attacks (70% of cases): no warning (knife concealed), attack launched within 3 ft of the victim and led with the free hand.

    If you have any links or videos addressing the issue please feel free to post in the comments. All suggestions are welcome.

    Now, let's see what can be done for the other 30% of knife attacks where the aggressor leads with the knife.

    The good news is, there's lots of material for this situation.

    Parrying the knife

    In discussions about how to defend against a knife attack (see here or there for example), one common piece of advice is "keep them at a distance".

    It's a very tricky and contentious point.

    It sounds like common sense to stay far from a knife but, remember what we've seen, it's not always so easy to achieve once the fight is on at close quarters. And most knife attacks occur at conversation range (less than 3 feet of the victim).

    Truth is, if you can't put significant room between you and the knife (i.e. escape), distance might not be such a good friend.

    How's that?

    Look at the following video. Paul Vunak shows the problem with parrying a knife bare-handed:

    The point is: if you can't escape or if you don't have a weapon to safely keep your aggressor at a distance, moving out and creating space will just give him more opportunities to slash/stab you.

    And he'll keep coming at you, adding more cuts every time.

     "You can’t defend against a determined and/or frenzied knife attack. You’ll just end up cut to ribbons..." Neal Martin (Combative Mind)

    Let's repeat it here: the more times you get stabbed or cut, the more chances a vital organ, such as the heart, or a major blood vessel, such as the jugular vein, will get punctured or slit.

    The same reservations can be applied to kicks.

    Kicking the attacker

    In cases of knife attacks, 'groin-kicks' and 'push-kicks' are often presented as a viable options.

    Remember, though, most attacks are launched within 3 feet of the victim so you probably won't have the space for a kick.

    Even if you had the space, kicks are actually much harder to pull off in a fast moving, dynamic, real-life encounter, than a drill in a controlled training environment would suggest.

    Below is a video by Neal Martin during seminar with IDF Yamam operative and Krav Maga instructor Itay Gil, that shows the limited effectiveness of kicks (at 0:05):

    The limitations of groin kicks are very well detailed in "The Myth of the Krav Maga Groin Strike" by Gershon Ben Keren (Krav Maga Yashir Boston).

    But the main point is that a knife-wielding aggressor who is charging and reaching, will probably stab or cut you before your kick lands.

    In that regard, it's important to keep in mind that a serious cut to any of the major blood vessels -such as the femoral artery (inner thighs), the brachial artery (arm) or the carotid artery (neck)- could result in your death in less than 5 min.

    Note: If your femoral artery was severed, bleeding would be heavy at first. Then, after losing 1/3 of your blood volume -which would take 30 seconds to 1 minute depending on a number of factors such as your built, the depth of the cut and your heart rate- your blood pressure would drop and you would pass out. At that point the flow of blood would decrease significantly but, after another 1 to 3 minutes, you would go into an irreversible shock due the loss of so much blood (i.e. exsanguination) and eventually die.

    This leads us to the conclusion that you need to control the knife-bearing limb.

    By "controlling", I mean 'trapping'/'capturing'/'grabbing' and 'holding'/'locking' the knife-bearing arm so your aggressor can't use it.

    Trapping the knife

    Before you can properly trap the knife" -more precisely the knife-bearing arm- you need to block it.

    "...when I say stop the knife attack, I mean stop the knife from entering your body by whatever means necessary..." Hank Hayes (Knife Defence 101)

    Obviously, you want to block it in such a way that allows you to get control and capture the arm in order to stop the attacker recoiling and thrusting.

    For this reason, you will often see double-hand grabs with C-grips / V-grips in knife defence training videos such as the one below by Jeremy Pollack:

    It's been demonstrated on a number of occasions that grabbing your aggressor's arm (ideally the wrist) with only one hand is not the safest option because it's easy to break free from that type of grip.

    It's better than nothing, sure, but the moment you are dealing with a concerted stabbing effort, 'one-on-on' might not be enough to mitigate the attack. '2-on-1' is always safer.

    If you end up with only one hand on your aggressor's wrist -which is by no mean the safest situation- you'll need to seriously disrupt his attack by hitting him (namely in the face).

    Double-hand grabs give a strong control but there a small catch.

    Trying to block someone's arm at full speed just with your hands is very difficult.

    Additionally, your outstretched and already hyper-extended arms can become a target for some kind of armbar: all the attacker has to do is spin clockwise and smash his left forearm into your elbow.

    It may also result in your thumbs being broken/dislocated as Richard Dimitri perfectly explains it in the following video (skip to 1:35):

    For these reasons many instructors advocate blocking with the forearms as demonstrated by Michael Janich in this video from Black Belt Magazine:

    Or Ken Brayman in the following tutorial:

    The problem with this type of blocks is that the space between the arms, along with the underhook, leave too much room for the knife to slide up your arm and directly into your neck.

    This is why some instructors advocate the cross-block -using both forearms- as shown by London-based Darren Selley in the following video:

    The main criticism of these types of blocks is that the attacker will retract/recoil (e.g. if you get the timing wrong) and your arms will get slashed, resulting in severe limb mutilation.

    Of course, that's what will happen if you just stand static after you blocked the knife.

    As you block the knife, you need to explode forward and take control.

    Close in & Take Control of the Knife

    In order to attack properly, your aggressor needs space and movement.

    By closing in after you've blocked the knife, you'll deprive him of both as demonstrated by Stewart McGill (Urban Krav Maga):

    Moving into the attack is critical to stop the stabbing (by reducing space and movement) and to prevent counter-attacks (e.g. punch with free hand, head-butt, etc).

    Now, you may still get stabbed or cut, but it is better to get stabbed once than multiple times, as you would if you backed away from the attack.

    Here's the important part:

    You have to shut your aggressor down right away.

    The solution is to cut the attack short and close in with as much kinetic energy as possible.

    The movement is driven by your legs which hurl you forward into the attack as explained by Itay Gil in the following instructional video:

    It's critical to lock your aggressor's arm so he can't use it. As soon as you give up that control, the knife will be back in action.

    "Your first objective should be to control that knife before anything else. If you don’t control the knife you don’t get to do anything else without getting stabbed repeatedly." Neal Martin (Combative Mind)

    The whole issue is brilliantly summed up by Master Wong here:



    Knife attacks are a very difficult and complex topic.

    Most of what we know looks more like inherited wisdom, passed on from and by instructors, than primary information source.

    Hence the importance of evidence-based approaches.

    For that reason, it would also be amazing if readers who experienced knife attacks first-hand could tell us their story. That would make a great contribution to the topic.

    I would love to know your thoughts particularly about training methods, but also about techniques.

    Or maybe you have a question or just want to share a link.

    Either way, don't hesitate to leave a comment below.

    Thanks for reading this article, I hope you enjoyed it and found it informative.

    Remember to share it!


    In Part 1, we saw that:
    • 71.1% of knife attacks are led with the free hand and this dramatically alters the dynamic of the attack along with your initial reaction.
    • Most knife attacks are ambushes, not duels. Attackers rarely brandish the knife; they keep it concealed until the very last moment.
    • 70.6% of knife attacks are launched within 3 feet of the victim
    • Knife attacks are fast and furious, often resulting in the victim falling on the ground.
    • Knife attacks don't last long, avg time is 23 seconds but 50% of attacks last 14 seconds or less.
    • Knife attacks are more often performed with quick, short repetitive stabs at different angles. There's usually a first, very aggressive, wave of stabs (5 to 10).

    In Part 2, we covered:
    • Avoidance and Situational Awareness
    • Knife awareness is key
    • Escaping should be the priority
    • Compliance and its limitation
    • Weapons and the issue of deployment

    In Part 3, we highlighted a number of principles to remember:
    • you need to get control over the knife-wielding hand (or arm)
    • grabbing the knife wielding hand is always preferable to parrying/blocking
    • close the distance quickly and smoothly and stay as close to your aggressor as possible
    • maintain forward pressure and
    • take control of the knife-bearing arm as soon as possible 
    • shut down your attacker aggressively


    1. Brilliant stuff. I was involved in an attempted murder case with a knife a few years ago and since then i have been searching for the holy grail of technuiques. I dont belive one exsisits but this was very well thought out and informative,Thankyou

      1. You're right, no holy grail unfortunately.

        The only thing that can be done it to mitigate the risks . Or, to put it differently, limit damage but avoiding a number of mistakes.

        Thanks for sharing.

      2. PAtrice, really excellent article. Many thanks. I struggle with this issue looking at UKM ,IKMF, Senshido, Fight, Jim Wagner, TargetFocused. etc...this evidence based work reminds me of Fight like a Physicist..good work certainly explore in drills.

    2. Great article sir, and I really appreciate the time you must have taken to put it together.

      I always use CCTV videos and compare it to the training that is available. In most classes, "Grab and Stab" is not practiced. Nor is being attacked from different angles. Also, there is a disproportionate amount of time spent doing knife v knife training - as pointed out in your article.

      Where it is knife v empty-hands, it's usually an instructor showing off with a training knife (in full view before the first cut). Little emphasis is given to making space and running, or dealing with knife wounds - despite the oft repeated adage "Expect to get cut".

      I wrote about some of this bullshit here:

      Thanks again for your article and the great info / videos within!


      1. Hi Tony,

        Thank you.

        Great article of yours; Thanks for drawing my attention to it. Very well written and insightful.

    3. The Teuller, "21 feet" rule, rule is just and approximation. Different people have different reflexes, speeds, ability to accesses weapons, etc.
      I would NOT rely on that rule as an absolute for gun, knife, club, whatever. And I say this being in Texas, were lots of us folks have CCWs and IDPA/IPSC is popular.

      But I do say, your advice on this page is excellent.

      1. Thanks Paul.

        That's correct. The "21 ft rule" as it's known is more a guideline.

        I wouldn't rely on that either but it's useful in that it helps people to understand the concept of "reactionary gap" and that they should integrate these parameters into their training.

    4. At the age of 8 I was stabbed 27 times by my mother. It was a hand to hand fight until I could get out of the room. I then had to conduct first aid to myself until the paramedics arrived. If you would like to talk more on this hit me up on Facebook with a PM. Nate McSweeney

      1. Sorry to hear about that terrible experience. Thanks for sharing.

        I'll get in touch on FB.

    5. Thanks for this. Knife attacks sound bloody terrifying.

    6. Nice study !

      Do you see any methodological flaws your dataset could have ? At first I would have assumed that Youtube would be biased towards CCTVs footage but your sample seems surprinsingly varied.

      Have you though asking your local police department for similar informations ?


      I was also wondering where knife attacks were happening. You state

      > Now, remember that most knife attacks are ambushes;
      > They're usually launched in a semi-closed environment, within 3 feet of the victim who is cornered.

      This seems reasonnable, but I assume this conclusion stems from your dataset of video. Perhaps you could make your case even stronger by looking at various governmental statistics website ? I wouldn't be surprised if there was some informations that would be coherent with your statement. You could try the UK [ ], and US [ ] , Eurostat [ ] and puisque vous êtes francophone l'INSEE [ ], l'OFS [ ], et iln'yapasdejoliaccronymechezlesBelges [ ] devrait avoir quelque chose à vous dire.

      1. Thanks.

        I'm not sure the police in the UK would release much information, but that's worth exploring.

        I'll have a look at the links but from experience official data rarely provide of level of detail needed.

        It's hard to find a perfect data set but the prevalence of CCTV cameras in many (and more) places along with the number of phone footages result in a good sample which is also supported by the variety you mentioned.

        It's true, though, that surveillance cameras being mostly in shops and shops being a places that are targets for robbers, the "semi-closed" factor is probably amplified. But, from a number of footages, it also seems that semi-close environment confers good tactical advantages to an aggressor. So we may not be that far off in terms of representativity.

        Of course we're missing data from domestic contexts. Here, information provided by the police helps a bit as it appears that reverse ("icepick") grip is most most common. But that doesn't make a bit difference with regard to the way you're going to defend the attack.

        In any case, the most objective information we can get are these videos.

    7. One of the best articles on knife defense i have read!

    8. Incredible work. You can download a PDF of the Pentecost book here:

      1. Hi Richard,

        Thanks for the kudos and the link.

    9. Great article. I appreciate the let the data (video of real knife usage) speak for themselves.

      1. Hi Craig,

        Thanks. In this day and age, I think it's massively important to put the facts at the forefront.

        What's your main take away from this article?

      2. So true. Here is my biggest take away. Ever since I read Don's book over 20 years ago, this is how I taught my students to use a knife. I did not realize that it was so common in the real world. I have seen many of the same videos you reviewed so knew it was used but to see 70% of the time was an eye opener for my weapon defense training. I utilize a combination of footwork, parries/blocks and evasive body movements to get out of the knife path followed by arm monitor provided it does not significantly delay my strikes. I am not a fan of submissions or control movements to stop a knife attacker nor someone being able to take their weapon away (disarm). You need to launch power strikes that shock him.

      3. Thanks Craig,

        It seems there will always be discussions around the technical aspects but I suspect that, more than the technique, what makes the most difference is speed of reaction and focused violent intent. So, knowing the situation well enough to understand the dynamic and realising what's happening is critical which is why I presented the article this way.

        Do you have videos of your work?

      4. Yes, here is me tonight at a scenario. They distracted me with the first attacker shoving me to the ground then the second attacker entered with a knife. It happened too fast to parry and step to the outside on the grabbing arm. I blocked the low strike twice before I could finally get my hands to stop the knife enough for me to launch my own strikes. There is lots to learn in this clip if you look closely -

      5. Chaos and surprise! I like to replace knives by markers to see the amount of damage.

        Good work!

    10. Just a great article. I was for years saying "there are a lot of real world fighting videos out there now, unlike twenty years ago we can just go look at them." You did it as a study and survey. You debunk respectfully so much nonsense. Many thanks. RM in Massachusetts.

      1. Hi,

        You're very welcome!

        As you rightly point, lots of data is now freely available and time was long due to take our industry beyond the all too common "personal opinion" quarrels.

        I also believe that for the sake of everybody, egos should be kept in check. Respectful, evidence-based discussions can be very constructive.

        All the best

    11. I was also stabbed by a family member 15 times a few years ago. How it physically feels to be stabbed: it feels like being hit with a really hard, really precise punch. Fine motor skills definitely disappear, as you mentioned. Using my left arm as a shield probably saved my life. The arm was messed up, but you can live without an arm.

      1. Hi,

        Sorry you had to experience that and thank you for sharing here.

        What stopped the assault in the end? Third party intervention? Did you manage to escape?

        Have you fully recovered the use of your arm?

      2. Thanks for the reply. Yes, third party intervention ended the assault. And yes, after multiple surgeries the arm is almost as good as before.
        Thanks for compiling this article. It's a fantastic study. I hope a lot of martial arts and law enforcement people will use it.

      3. Glad to hear your arm recovered.

        Mu hope is that, following this article, people in the industry are going move beyond "opinions dueling" and towards evidence-based research.


    12. What tactic adjustments would be feasible in consideration of age and gender?

      1. Hi,

        Considering how bad knife attacks are, I don't think there is any tactical adjustment to be made based on age or gender.

        Against a knife, everybody is weaker and in grave danger of being killed. So everything (awareness, techniques, principles, etc) applies the same for everybody.

        Whether you're male or female, young or old, if you get stabbed in the heart or cut on the throat, you die.

        So, you'll have to do whatever it takes to survive regardless of your gender, age, size, physical abilities.

        The possible adjustments are on a personal level. That is, what you can personally do or not.

        What I mean is that you, as an individual, need to understand where are your strengths and weaknesses, what your speed allows you to do, what your physical strength allows you to do, etc...

        This understanding is gained through training, and more particularly "pressure testing".

        I hope this answers your question,

    13. Great article. It opened my eyes and made me reconsider all I've been taught. Every serious martial art should study these concepts at length.

      1. Hi Darrin,

        Thanks, glad it had a positive impact on your approach to these issues.


    14. Are you familiar with Tim Larkin and his Target Focus Training? Whether empty hand, impact weapon, knife or gun, they teach you to cause injury in the other guy, period. Real injuries to the body such as strikes to the eyes, throat or neck, groin, solar plexus, collar bone, ear cupping, breaking an ankle, etc., will put the real weapon, the person's brain, out of commission. They say yes, you may be hit, cut or shot, but if you get a good injury, you will win. Stop focusing on the tool and put the weapon (the person's brain) out of commission. Something to think about anyway.

    15. Hi,

      I haven't trained in TFT method but I'm aware of Tim Larkin and his system.

      I'm on the same line as I think it's more important to develop skills than to have a collection of techniques.

      As they say "you're whether at the table or on the menu"

    16. I appreciate the effort and time you put into researching this issue. The fact that you reviewed so many videos, source your material and identify what if any issues you have with the technique being shown speaks to your dedication and thoughtfulness. As someone in the US who carries a gun every single day the belt technique wouldn't work for me. So I'm left with the empty handed skills portion. Thank you for providing such insightful material.

    17. This is perhaps the best article on self-defence against knives I have ever read. Kudos!

      I have training in JKD and Kali under Burton Richardson for more than a decade, and learned through regular sparring that empty hand vs. knife is just a terrible situation to be in. Below is a short clip of Burton demonstrating a disarming technique. I welcome your thoughts and criticisms.

      1. Thank you.

        Burt Richardson is a name I could have included in the article as his work is well known and respected.

        Regarding the technique in the video, I would prefer to be closer to the aggressor with a stronger grip on his arm (you're stronger when you maintain something closer to your chest than at arm length). This would also allow me to use my head and eventually my knees to strike the aggressor while maintaining a firm control. More leverage on the arm too.

        This being said, empty hand against a knife is one of the worst situation ever, so 2 hands on the wrist is better than only 1, and 1 hand is better than no control at all.

        As I say in the article, you need to be able to adapt to a variety of situations.

        What if I totally mess up my move and end up with only one hand on my aggressor's wrist? How do I recover from that and improve my position?

    18. Great breakdown of material. I have trained in FMA for a while, and can see the points you are making and how it relates to training. Still with the tendency to focus on trapping the attacking arm; I still believe this would be tough for a person of small stature to achieve. Can't offer any solutions, need to take this material back and experiment.

      Everyone should be focusing on multiple talents, everything from situation awareness, pre-attack signals/behaviors, through self defense, post-incident response and first aid.

      One thing to add; identifying specific skills that could improve your survival, then developing drills that can instill that skill/reaction. One example that comes to mind is trapping (pick your poison JKD, FMA or Wing Chun) in the development of hand speed/reaction.

      Great work, thank you for taking the time to put this together.

      1. Hi Junshi,

        You're welcome.

        I'm totally with you on this. Multiple talents/skills including post-incident response and first aid.

        I always encourage my students to cross train and try other styles to develop or improve their skills. I believe that's the essence of learning various systems/styles. It's not about creating a collection of techniques but about developing skills.

    19. If I were an employee in a high risk crime area it might make sense to wear certain items that can degrade a knife attack. Usually they are made of a Kevlar material for the throat or forearms. Others options include abdomen protection. They are not obvious but might make the difference between survival or crippling wounds / death. Searches on Amazon will reveal various anti-blade items that are not terribly expensive and still allow an employee to function normally. IIRC, back in the late 80's, some felons recommended thick bound copies of National Geographic magazine under the shirt to protect the kidneys. And as Pentecost also wrote: "in a life-threatening situation, get whatever you can, when you can, as many times as you can!"

      1. I would do the same. As you say this stuff is really not expensive and could make a big difference. I'm still amazed how few security workers actually wear these

    20. Looks like it is very hard to even know that you are being stabbed let alone try and fight with a disadvantage. Seems to me that armor is the best if you can wear it all the time, passive protection. It would be nice to have a shield as well but you cant carry one around and ,like a gun, wont work if the attacker is on top of you before you know it.

      1. Yes, that's what most victims report. Not knowing they had been stabbed. In that regard, one commentator here said "How it physically feels to be stabbed: it feels like being hit with a really hard, really precise punch."

        Anti-stab vests are not the only grail but that would definitely improve you chances.

    21. Absolutely fabulous collection of material. Thank you sir. I look forward to more.

      1. Thank you Daryl!

        I'll try to keep up with the good work ;)

    22. Patrice - thank you so much for this. What I appreciate the most is your evidence-based reality-focused approach. You've opened up the conversation for experts to start engaging together and create workable solutions. I am grateful for the time and effort you put into this.

      1. Hi James,

        Thanks, means a lot coming from you!


    23. Article has good and bad stuff in it. Good is that you analize many real situations, bring statistics so we can see what is most likely scenario, tell on what we need do pay atention, what are options and some Stuart technique (going behind opponent's back) and some different aproach from common. Bad is technique especially in part 3, you say that you put some video's for us to see grip or arm position you prefered and that's ok, but technique people show up is not real. In some cases defender see the knife, in real attack most likely not. There is no surprise element in any of gym attack, also according to you and I agree with it, is big part of potential defence.
      Not mentioned in the article, but I saw many people practice to defend with knife from knife attack. Yes is good if you have that skill, but what are the chances for civilian to have knife close to himself when attack with knife occurred, if we exclude kitchen fights :D . People who have weapons most of the day (police, military, security officers ...) should train to use them, others no because in attack situations there are no weapons in his arms, so it just waste of time to train knife ws knife. Instead, we should train how to be awere stuff around us and what we can use if attack occurs, it could be stick, bottle, belt and bagpack as you said and many more things.
      To end with great stuff of the article. Weapon part is something that in my opinion is need to focus on and develop skill for recognize potential one around us and use them for defence. Also, great is that someone invested time to see and collect all the videos, experience to separate useful one and share with the world this article, so we can discuss and make some progress on field of defence from knife attack.
      Congratulation on that.

      1. Hi Vladimir,

        Thank you for taking the time to leave a comment and being part of the discussion.

        Fair point regarding the techniques. I guess we'll always be frustrated with this technical videos which have to conciliate clear explanations and realism. Not the easiest thing.

        I could have gone to the extent of just including videos of me focusing on the techniques I consider the best with regard to the situations at hand. Quite like what I've done with Stewart and David. But I also felt it was important to show and acknowledge the work of other people (however imperfect their videos or techniques were!).

        There are three reasons behind this:

        (i) credit should be given where it's due. A large part of our knowledge and work is often the sum of a number of other people's work and knowledge. This has to be honestly recognised.

        (ii) Similarly, solutions are found because we share knowledge and experiences. We confront points of view and ideas. Therefore, it's important to open up the conversation and engage with fellow "martial artists".

        (iii) I wanted people (the readers) to get a sense of the variety of techniques and approaches. And I wanted people to see for themselves and make their own mind.

        Thanks again for reading and commenting.

    24. Very well written article. Some of the best information I've seen regarding knife defense. So many people continue to believe "I'll just shoot the guy", and then we put them in a real time situation. Eye opener. Yes, best advice is avoid and run away. But, practice with as many people as you can to get a feel for different situations. Big, small, fast, strong, as many touches you can make helps the motor skills. Be prepared and hope you never have to use it. Great job.

    25. Thanks Ernie.

      Good points re. the importance of training with a variety of people!

    26. Well written and excellent research in this article. I have a question regarding a comment, you say you don't think the classic krav maga 360 defence wouldn't work in a scenario, could you possibly expand on why you don't believe it would? I know it's not a perfect solution, but the principle of blocking while striking the attacker seems a sound solution. It's simplicity helps as no fine motor skills are required in a high stress situation.

      1. Hi Ciaran,

        Thank you. There are two things here regarding your question about the Krav Maga 360:

        The first is that I mentioned this point in the context of the "free hand" lead scenario which is the most common.

        The best way to understand my point is to recreate the scenario: ask someone to grab you by the lapel, then push you backward as hard as they can, arm extended, while stabbing you with a training knife.

        In that situation, (i) you're moving backward or at the very least you're on your back foot which is not a position from where you can punch efficiently, (ii) the aggressor's arm will likely be in the way, offering him some protection and impairing your aim and strike, (iii) you also have to deal with the knife, which leads me to my second point regarding the "360 + punch" defence:

        From a neurological perspective, the fact is that our brain can't multi-task. That is, it can't actively do two things simultaneously. What's happening when we do what we call multi-tasking, is that in fact we are shifting our attention sequentially; not doing two things simultaneously. That is, our attention will go from one thing to the other alternatively. Really, we're juggling tasks.

        This is fine with a number of tasks like keeping an eye on the food which is cooking on the stove while also keeping an eye on the news on TV (even though, food gets burned on a regularly basis!)

        But for a number of other tasks it can't work because attention can't be divided. You need to have your full attention on only one thing. Particularly when things go very fast.

        Ask a baseball player where his attention is when he's in the batter box and the pitcher is winding up. He'll tell you that all his focus is on the ball. Ask a tennis player where his attention is when his opponent is releasing the ball to serve. He'll tell you that all his focus is on the ball. Ask a goalie the same question and you'll get the same answer. Over and over again across all sports.

        That's also why so many countries have made illegal to use a hand-held phone or similar device while driving.

        "Multitasking" doesn’t just slow you down and increase the number of mistakes you make; it temporarily changes the way your brain works.

        Similarly, when someone attack you with a knife at full speed, you can't divide or shift your attention between (i) blocking the blade and (ii) precisely and properly punching the aggressor in the face.

        You need your full attention on blocking the blade, because if you mess it up you're dead!

        For further reading (with references to scientific research):
        -The Multitasking Mind

        - The Myth of Multitasking

    27. My name is Hank Hayes I'm the CEO of NLB Tactical and No Lie Blades. The author of this article was nice enough to mention (quote) our work a few time. Thank you. We provide tactical combative training for the Military and Law Enforcement communities offering edged weapons, impact weapons and firearms training mainly to the special teams communities. I though the article was well written and covered a wide variety of important items and did a good job at that!

      Edged weapons skills is a very sensitive topic subject filled with trainers that don't know what their talking about and instructing bad training. We've been able to remove the guess work by inventing the marking feedback training knife. It provides immediate, accurate real-time feedback letting everybody on the training deck know whats fight strategies are working and which are not.

      I have found that very few people can speak intelligently when it comes to the truth in knife combative. This article has achieved that goal.

      Please seek out our company's services as we are the only United States GSA contract provide that has this training technology and accountability training platform.
      Hank Hayes / NLB CEO

    28. Excellent article, great information. I was attacked with a knife in the Philippines back in the '60s. It was a robbery attempt, and in reality, the guys intent was to get my watch an wallet, not my life. (Thank goodness.) I used the belt technique.. flipped it just like the old locker room towel snap, and when the buckle hit him over the right eye, it was over. He took off to find a more compliant victim. Had he intended to kill me, I'd likely not be sitting here writing this.

    29. Hi Ed,

      Thanks for leaving a comment and sharing your story.

      Good reaction though. I truly believe that the belt technique would work with most mugging attempts. No one is too keen on being slashed in the face by a belt buckle.

    30. Don Pentecost book will be available at for $16.95 beginning 30JAN2017 at this link:

    31. Thanks for the heads-up. Is that an updated version?

      1. No it is not, just a different publisher and cover art etc.

    32. Patrice,
      What is your opinion of this? I currently have no training, only an idea of what I might try, if ever in a knife attack. (something I hope to fix) AFTER attempting to escape, I would try one of two things, based on if I am carrying. I just recently got a Conceal Carry License, but do not carry yet (training first).
      Plan 1, if I were carrying, would be to escape or put enough distance, to draw and fire. If the attacker was too close, or got close after I started to draw, I would sit/roll to my back, and use my feet to keep him at bay long enough to fire. this idea comes from seeing training of LEOs who fire from lying on their back, and also the evidence that I may end up on my back from the attacker's momentum and trying to escape. I also think that my protected feet would be best way to keep him away from my head and organs. Again, if escape failed, or he was on me with no chance to draw.
      Plan 2, if I was not carrying, would be to grab the knife arm, as discussed above, and roll backwards, using his momentum to flip him over me. In a perfect world, he would hit his head on something, or otherwise suffer damage. If not, then hopefully I would be quick to my feet, and flee. This seems good to me, since grabbing and flipping him seems way less dangerous to me, than grappling, or trying to reverse his momentum and push back. Again, I have no fighting skills/training, so the Judo like move, and fleeing, seem good to me. In many of the videos shown above, the attacker is blocking the only exit, by his design. So, I also think this type of move would get me on the "exit side" of the area.
      do these have any deadly flaws/consequences, in your opinion?

    33. Hi Jason,

      [I'm not well versed in firearms so make sure you cross check my opinion with an experienced firearm instructor]

      Theoretically, rolling on your back with your legs towards the aggressor would indeed give you (relative) protection and time to draw your weapon and fire.

      This being said you need to be sure you'll be able to draw your firearm because once you're on the ground -struggling to deploy your firearm- you're in a very bad position (little protection, no mobility, etc).

      I personally would still try to stay on my feet (you move faster on your feet) and create space laterally.

      As for Plan 2, I would advise against going to the ground intentionally. Particularly if you put yourself in bottom position.

      As a general rule for self-defence, don't put yourself at a disadvantage with the idea that you're going to pull off that amazing technique which will turn the table.

      It rarely works as intended and if you mess it up, well, you're dead!

      I've train a lot with people who had background in judo and BJJ and they tend to try to take the fight to the ground and pull you in their guard and in my experience it doesn't work well in the context of self-defence simply because I can do a number of things that are forbidden in these sports (like punching) so it changes the dynamic of the fight.

      If you're taken to the ground and there's no coming back, in that case sure, you can try to flip him over.

      Otherwise, I'd try to stay up.

      The best way to figure this out is by pressure testing during training. Try it and see how it works for you.

    34. It's why I don't devote much time to knife specific defenses like a lot of my fellow Krav instructors do. Empirical evidence teaches us (as you've mentioned here) there's almost never any advance warning and even when it's on most people don't know there's a knife involved until later. (I think it's fair to say the Filipinos can make that assumption ;) )

      If we don't know there's a knife involved why are we practicing different techniques for knives than for empty hands?

      What would you do if a guy ran at you with his lead hand attempting to drive you backwards and was punching you with his right? Personally, I'd throw an overhand right and knock him out because the fastest way to win altercations is to hit the other guy as hard and as fast as you can...and that works whether he's unarmed, tooled up or has friends with him.

      Ergo...spend most of your time working on really hard fast accurate striking techniques - predominantly with your hands - and that will give you the greatest chance of survival.

      PS: Not a huge fan of the prison study because in prison their weapons are improvised and usually only capable of stabbing, therefore they don't deal with any type of slashing attack like we say on the planes on 9/11 with the box cutters they deployed.

      PPS: Good work doing the research and attempting to tackle what is a tough subject.

      1. Hi Nicholas,

        Very good points, indeed.

        We could also take the problem the other way around. If we do know whether a knife is involved or not, maybe we should act with the assumption that a weapon is in play.

        The only reason we would work on specific knife defence techniques is to develop specific skills. But, then, everything has to be bound together (pressure testing is a good thing for that).

        Thanks for your thoughtful input!

      2. I totally agree with your comment - if you do not know they have a knife then your defense should assume it which means everything needs to be synced up. Totally agree on the importance of power strikes.

    35. Very well researched, thank you Patrice!BTW have you watched the wonderful Jim Carey sketch on knife defence on youtube?
      Great lesson in how not to teach!

      1. Hi Sue,

        Thank you.

        Yes, I know that Jim Carey sketch. Hilarious!

    36. Erwin Kastl

      Thank you very much for this great article. I was doing a lot of research and training on knife defense. If the attacker leads with the knife than marc danny shows on his great DVd´s " die less often" a good solution. I mean the best solution I have seen so far.
      If you don`t see the knife, i think it is always the best to work on a strong simultaniosly conterattack like Tim Larkin or Wing Chun.

      1. You're welcome.

        Do you have any reference for Marc Danny's work?

    37. There are some clips from Marc Danny on youtube:
      This is my work:
      If both arms are used to deflect the knife, the head is free for counterattack. Use it.

    38. Awareness is key... good article. Unless you are a military personnel or someone training with knife attacks consistently, chances are you will get cut and killed. It is best to carry a good brand of pepper spray or a stun mechanism and keep it ready. Along with awareness you can survive a knife situation. A gun is totally a different story.

      1. Thanks for dropping a comment. A weapon could help indeed but deployment remains an issue.

        Pepper spray is a very contentious one as the effects are not necessarily immediate.

    39. The first step is to do research--which you have done some work on. You forgot to look at the Dog Brothers and you obviously forgot my work from 1995 (more on that later).
      The salient area that you neglected to research that is crucial is the blade LENGTH: that determines how you will be attacked and what sort of unarmed answer you will have.
      I might also draw your attention to the fact that I put out a knife defense tape in 1995 that used real knives that were sharp at real speeds--the "gurus" in the knife fighting industry called me irresponsible. Really? They are charlatans--you cannot learn how to land a jet if you do not practice landing a real jet--markers, rubber knives compliance are all thins that are not ISOMORPHIC. We have always used this powerful technology as have NASA and militaries around the World.

      1. Hi Chris,

        Many thanks for taking the time to comment here.

        The main objective of this research was to see what actually happens in real life (not among trained people which it's quite different) which is why I started with the analysis of CCTV and phone footages of real life situations.

        Then, I tried to see what where the potential solutions to these situations.

        As far as I can tell from the study of CCTV videos, blade size doesn't make any difference in the way people are attacked in real life (except, of course, when it's a machete).

        This being said, I'm sure FMA people have a lot of interesting things to say about these aspects but it was beyond the scope of this article.

        Do you have any link to your 1995 knife defence?

    40. I've been threatened with a knife twice: Once in Amsterdam, a guy wanted to sell me drugs and wouldn't take "no" for an answer (kind of deranged). Fortunately, he didn't have the knife ready and sort of turned to get it out. I noticed, we were a little hemmed in so I quickly said "OK, I'll buy", which distracted him and then I employed the sudden "run away" move which worked. The second time was in Old Jerusalem in an alley -- I saw the knife he was sort of concealing and something paranoid in me that day made me think about routes of escape. Again, to distract, I yelled "get him!" and then ran, I remembered seeing a stair case nearby to the roof level (which was open and you could run anywhere) and run up there where I could openly run. If you can (situational awareness), running is a great defense. I'll also say: in both situation, my brain was raising hairs on other cues that I wasn't consciously aware of, but listening to your brain is good advice.

      1. Thanks for sharing your experiences. Glad you were able to escape unharmed.

    41. Hi, your article is the best one I have read with regards to knife defense.
      In fact, you should make it into a mini booklet for sale.
      To me, knife defense is the most important aspect of self defense. The reasons being it is a weapon that is easily available and be improvised as well (screwdriver, ice pick, broken bottle, etc).
      It is silent and allows the assailant to stab several people in a crowded place without being detected till it is too late.
      Unlike a gun, you need no technical skills to kill with it, just need pure bad intent.
      Hence, I actually evaluate whether a self defense system is worthy by looking at their knife defenses. If the concepts and approaches are unrealistic, everything else is a joke and complete waste of time.
      When dealing with a knife, personally, I recommend the "Dog Catcher", a cross block done vertically to block both high line and low line stabs, as taught by the Dog Brothers (see their dvd - Die Less Often Vol. 1).
      The Dog Catcher is zone blocking which works well when you have no idea (since it's an ambush most of the time) the next stab is going to be a high (chest and above) or low line (midsection or below) attack.
      It also allows you to block the sewing-machine stabbing motion, long enough for you to close in and trap the knife-bearing arm (using what I learnt from Itay Gil during his seminar which I have attended)
      I will then attack the head (especially the eyes) and groin. Most importantly the throat (in Asia where I come from, many robbers come on bikes and they will rob with their crash helmet on, so that you can't identify their face later on and also for their protection in case you fight back. As a result striking the head will be useless, thus the only lethal targets left are the throat and sides of the neck).
      Another option I used is from Lee Morrison, where I will deflect hard the assailant’s leading left hand (push anywhere along the elbow to the shoulder) to my left with both of my hands, assuming his knife is in his right hand, to close his centerline as well as to change the direction of his forward pressure. This will bring me slightly behind him and I will slam both hands on the sides of his head to grab it and yank him backwards, following with knee to the back of his head as he goes to the ground.
      The above are just some of my thoughts on knife defense.

      1. Hi Alvin,

        Thank you; it's a good idea indeed.

        I personally like the cross-block too. Your point about going for the throat to avoid punching the helmet is a good one.

        I don't think I've ever seen that Lee Morrison technique you mention. Do you have a link by any chance?