I don’t teach the double knee bend


Credit Wonderlifter

Many things in life are done out of tradition. What used to work in the past should work in the present. Only a change technical rules of the sport could influence majorly how you would want to lift a barbell overhead .  In other words, although the general concepts of weightlifting don’t change much, the best technique will change – or evolve – according to the rules of competition. Over the years, weightlifting has evolved in many ways and the best methods are the one that match the current situation and rules.

Don’t miss out on Updates : Make sure you follow First Pull on Facebook and Instagram for daily pictures and advice.

In the older days, competitions involved many more events that no longer exist – such as single arm movements. This meant that you needed to (1) actively train those movements and consequently (2) the tonnage ratios were different than they are now. Fast forward a few years, only three movements were performed in competition : Snatch, clean and jerk and clean and press. For a long time, the bar did not touch the legs and many people lifted in split. This meant that training had to include a lot more pressing movements and unilateral movements than current methods often include. Currently, only two movements remain (Snatch and clean and jerk) and the bar is allowed to brush the thighs and contact the hips. Nobody lifts in a split anymore – except may be a few crossfiters and/or injured athletes and/or master lifters.

It is fair to assume that these changes in rules and the removal of some competitive events have led to changes in training methods and technique. The tonnage and assistance exercise selection have to change to best prepare the lifter to the current state of competition. There is a Russian text that has actually looked at the impact of the abolition of the clean and press on the performance of the jerk. They were interested by how much pressing tonnage is required to keep progressing on the jerk nowadays versus pre-abolition of the press. I believe they concluded that only 10-20% of the conventional tonnage was now required.


Picture by Hookgrip. I drew on it. You can see how Apti’s knees are never shoved forward – this is illustrated by the blue line (his back angle does not change much up until the last moment).


In the older days, teaching and drilling an active double knee bend – which is to deliberately push the knees far forward to rebend them while the torso opens once the bar is at knee height – made a lot of sense. You could not touch the thighs – so you had to use a lot more strength and power to get the bar moving up. A lot of people caught bars rather high too. Now the focus is to pull in a way that allows you to move down as fast as possible. Ever since the thigh brush has been allowed, the result in the snatch has moved up.

I have many friends – including some well known and very good coaches – who use this method in their coaching with success. Many international lifters use this technique as well – again with success. This is why I am not knocking it out by suggesting that it is out of date. I believe you should try it for yourself and see if it works for you. My personal experience as a coach is that it brings more mistakes in the pull than it corrects.

I personally do not coach this pulling technique and I also disagree with some of the reason it is coached as a deliberate/conscious action . I have often actually coached the opposite of this technique by limiting the forward motion of the knees at knee height in athletes that push them too forward. I believe that the DKB (double knee bend) should happen naturally – if it has to happen. A natural double knee bend is a double knee bend that happens without the lifter thinking about shoving the knees forward as they pass the knees – thus as a result of a good pull and going under the bar. That is, it will happen around the power position (upper thigh and not at knee height).

I don’t believe in teaching it in a way where the lifter is conscious about introducing horizontal movement within the pull or conscious about bringing his/her knees forward as the bar pass the knees.  I also think that it does not set well the lifter for going under the bar. I thought I would go quickly over some of the reasons that led me to not coaching the DKB.


Credit Wonderlifter


1. An active DKB creates a vertical torso too early in the pull – reducing the contribution of the back in the pull

Let’s be clear that weightlifting is not the same as jumping for height. However, contribution of the legs, hips and back seem to be fairly similar between the two during the extension phase. Moolyk et al. (2013) writes : ”Further, the correspondence in kinematics between impact phases of jumping and weightlifting tasks suggests that similar muscular strategies are used to perform both types of activities”. We can derive concepts from looking at the similarities between a high jump and weightlifting movements.

That is, to achieve height in your pull or height in your jump, you need to extend forcefully the legs, hips, and back – not just the legs. If you look at the video above, what you can observe is the best strategy in jumping. As the jumper dips, you can see that he pushes his hip back, bends over  and bends his knees. His ascent is him extending his legs, hips and back. He doe not bend, extend then rebend his knees. This position resemble greatly what happens at knee height in the pull or pretty much identical to what you would do in a hang snatch.

The premise of the DKB is that it supposedly allows you to use more legs in your pull by putting back the load on your quads via a change in knee extension angle. However, if a DKB is not required for a height jump, why would it be required for a snatch? It is not required because jumping and weightlifting performance is not solely dependant on knee extension. It is also extremely dependant on back and hip extension.

No body would jump high if they only bent their knees and kept their torso vertical. You can try this at home. Go against the wall, put your back flat against it, bend at the knee and try to jump for height. Compare the height you got versus the height you got from jumping using a technique like shown in the video. Which technique made you jump higher?

While a weightlifter is not trying to jump for height when performing a snatch or clean and jerk, a height jump is indicative of the ability of the body to produce power. The technique required to jump high is the technique in which you can produce the most power – and it does not involve a vertical torso early in the push.  The higher you can jump, the more powerful you are. Power is good for weightlifting.

Credit Hookgrip

Credit Hookgrip. Lydia does bring her knees slighty forward (frame 3 vs frame 4) but notice how she is still directly over the bar (frame 4)

2. Weightlifting success is not just dependent on leg strength and drive.

If this was the case, then nobody would be doing posterior chain work or upper body work. You cannot lift heavy weight without back, glute, hamstring and shoulder strength. This is why we need to consider more than knee angle within the pull. We have to consider knee angle, hip angle and back angle.

The deliberate double knee bend will not put the hips and back at mechanical advantage. If you push the knees forward, because of the reality of human anatomy, the hips move forward, and the back opens. A general concept that is time tested is that you need to stay over the bar if you want to be successful. For most, an aggressive and deliberate double knee bend puts you behind the bar.

Most people will either keep the knees bent slightly or rebend them late in the pull naturally (not actively). This is a much better alternative because it allows you to use the legs, hips and back. It makes you faster under the bar as well as keeping the bar much closer to you. Compare the following pictures.

Tatiana Kashirina snatching 155kg. From below the knees to above the knees, the knees have not moved forward. Instead of rebending them actively, she extends them! This makes the biggest difference in bar trajectory and speed under the bar. Notice how she is over the bar - not behind like the active DKB camp teach. Credit All things gym

Tatiana Kashirina snatching 155kg. From below the knees to above the knees, the knees have not moved forward that much. Instead of rebending them actively (shoving her knees as forward as possible), she extends them! This makes the biggest difference in bar trajectory and speed under the bar. Notice how she is over the bar – not behind like the active DKB camp teach. Credit All things gym

155kg snatch by Tatiana Kashirina. She is over the bar, and her torso is not vertical at hip height.

155kg snatch by Tatiana Kashirina. She is over the bar, and her torso is not vertical at hip height. A natural – unforced (not a deliberate action) – DKB can and will happen here. Trying to get to this position when you are just above the knee cap will create problems in the pull. It does not have to be coached. Credit All things gym


Here is somebody off youtube who does a pretty aggressive DKB. Notice how at knee height her knees are straight, and her toes are up. Notice after she passes the knees how forward she brings her knees compared to Tatiana. When she does that, the bar is much more forward and as a result her back is much more vertical compared to Tatiana. The result is the bar going forward and her being slow under. Granted she is not an elite lifter like Tatiana, but this illustrate the potential flaws that come with an aggressive DKB.

3. You need back action to make the bar travel back behind your head.

A technique in which the lifter pulls with a vertical torso is a technique that is not based in reality. The bar starts in front of the body and ends up behind the head. There are only two ways the bar will end up being there. Smacking the bar forward which happens if you rebend the knees so much that the back is vertical. At the moment of hip extension, the bar will be sent forward because you cannot compensate with the action of the back. The other way is to use the hips and counter the forward movement with the back while acting on the leg drive.


That being said, an active double knee bend could, in theory, work well with a with lifter that has a longer torso and shorter legs. That is, if you have a long torso and you push the knees forward, you will still remain over the bar. Even though the knees are moving forward, the longer torso can still cover the bar.

Now, let’s be clear. Yes some athletes use it. Yes some coaches coach it. I do not have a negative opinion about their success and/or their expertise. These are some of the reason why I don’t teach my athletes to push the knees forward at knee height. I would rather cue them to keep the bar close to the thighs, and bring it to hips while being over the bar. Introducing thinking in the movement slows the movement down and thinking of that movements does not allow the lifter to use all of their muscle groups correctly.

Here is a video of Game athlete Michele Letendre – who I have worked a lot with. We removed the notion of the DKB from her technique and here is what it looks like now (the rebending is now not conscious).

A natural DKB will happen much higher and not at knee height and this makes the biggest difference. There is nothing wrong with a bit of a double knee bend activity – but to cue a lifter to push the knees forward does not make sense to me. Again, many were successful with such a technique – so it can work if you have the right body type for it.

Don’t miss out on Updates : Make sure you follow First Pull on Facebook and Instagram for daily pictures and advice.



4 thoughts on “I don’t teach the double knee bend”

  1. Well; the examples you gave from people that did the DKB were just a guy in youtube; would be good that you give some examples of weightlifters doing the dkb. Thanks for your posts.

    1. those that do so successfully tend to have a very long torso. Shorter torso people are more often than not, not successful with this technique.

  2. Got to disagree with you on this one Jean-Patrick. I think your premise about why coaches talk about the DKB is wrong and your comparison of lifting and jumping is wrong. “However, if a DKB is not required for a height jump, why would it be required for a snatch?” The answer is simple; because a snatch isn’t a jump for height! We aren’t extending to produce height but to best accelerate a bar upwards over the base AND be able to instantly transition to driving ourselves under it to a receiving position. So, the idea of lifting is to accelerate the bar upwards, while we go no further up than we need to because our final positional goal is down in the squat for most efficient lifting of big weight. I will agree with you on the point that if anyone is teaching people to FORCE their knees forward to achieve the readjustment(DKB and torso lifting away) then they are wrong. I teach, by drills before lifting instruction, that the knees have to be loose and able to react quickly to rebend a bit as the bar moves up the legs. They are ALLOWED TO REBEND,not forced forward. Even with the very different Chinese pull, there is a bit of rebending of the knees, although not as much as in the more orthodox European pull but it’s there. And, many lifters these days do not follow the old standard of driving with their legs in the 1st pull until their shins are vertical. The so-called DKB occurs as part of the readjustment to help keep the bar more centered over the base, so that the acceleration in the 2nd pull will occur over the base as well. A jumping motion w/o this reorienting of the torso/leg positions and angles will create a track that is too far forward of the base, creating more unsustainable positions at receiving. I find I have more of a problem with lifters who have been taught, once the knee ext. is initiated, don’t stop it, sort of like you might on a dead lift. I have to take time to show them that their knees can work in both directions during the pull and not be frozen, which puts more work on the back and hip ext. muscles than their should be. As usual, stimulating article

Leave a Reply