Program Shop Apex Athlete Training Team Free Training Templates Login

Why Your Coach Might Be Wrong About Your Squat: Knees In For The Win

Jun 27, 2017

Whether you're an elite powerlifter, a strength and conditioning coach, a personal trainer, or a physical therapist, you've probably been versed in the concept that proper cueing for the squat with the lower extremities is, spread the floor with the feet, and push the knees out.

Perhaps you've even gotten the tid bit about screwing the floor with your feet in the direction of external rotation as well. If you've learned that these particular cues are the way to go, then you've probably also learned that knees caving in towards midline, or valgus is the devil.

You've probably seen the technique involving putting a band around the knees so that you reflexively push the knees outwards (varus) against the input of the band. The rationale for squatting this way usually involves the concept that you're going to utilize more gluteal tissues since the actions of the femur will feature external rotation, via the feet screwing, and abduction with the feet spreading the floor and the knees pushing laterally.

This is the accepted way to teach the squat, and seems to be agreed upon by the people who are super strong with all the in the trenches experience in the world, as well as the people who understand how to get people out of pain and how to perform prehab to prevent people from doing things like tearing their ACL in the first place.

Despite all these recommendations and agreed-upon findings, you'll still see elite weightlifters feature the rapid action of knees moving towards midline while squatting up massive weights that they just dove under and caught in positions that require incredible levels of mobility and stability to get into. Some of the most gifted athletes I've ever worked with who demonstrate great sports biomechanics and produce incredible amounts of force production also seem to show this action of knees going towards midline during the upward portion of heavy squatting.

I'm here to say that I don't think I have any problem with highly athletic individuals who possess great joint biomechanics demonstrating knees towards midline during the upward portion of a squat. I'm also here to say that I think I have the answer as to why the strongest power lifters and strength coaches amongst us have fallen so deeply in love with the spread the floor, knees out approach. To explain my argument I'm going to walk you through concepts relating to natural femoral biomechanics, and what joint force moments have to do with this approach. I'll also give some recommendations on what to do with this thought process, and what measurements might be helpful with making objective determinations of what's optimal for people regarding squat biomechanics.

What Walking and Squatting Have in Common

During movement, the human body has to absorb forces as well as create propulsive forces. The gait cycle is the most fundamental and stereotypical movement pattern for humans. Gait is comprised of stance and swing, each featuring early, mid, and terminal phases.

Absorption of forces, as well as the creation of propulsive forces, are primarily rooted in the stance phase of gait. On a very basic level we could say that early stance is the absorption dominant part of stance, terminal stance is the propulsion dominant phase of gait, and mid-stance is the transitional period. Gait and squatting are certainly not the same movement pattern, therefore generalizations between the two movements should be looked at with skepticism; however, examining gait will help with the formation of the primary reasoning piece helping us understand the biomechanical argument being presented here.

While going through the gait cycle, the femur typically features alternating triads of movement. There are times during the cycle where the femur will group flexion with abduction and external rotation (mostly swing), and there are times when the femur will match extension with adduction and internal rotation (mostly stance).

The first triad of flexion, abduction, and external rotation is the strategy the femur uses to prepare for force absorption and to absorb force, while the second triad of extension, adduction, and internal rotation is the strategy the femur uses to prepare for and execute propulsion. Ultimately I see the same sort of strategy used by the femur during the squat, and that is the central premise of this argument. The common link between gait and the squat in this context is that both rely on the femur and the femur ultimately has its preferred strategy for absorption of force and creation of propulsive force.

The descent into the deep squat position would be the force absorption portion, and pushing back up to standing would be the propulsive component of the squat pattern. The natural tendency of the femur during the descent would be to group flexion with abduction and external rotation.

Conversely, during the concentric portion of the squat pattern, the natural tendency of the femur would be to group extension with adduction and internal rotation. This would be an intelligent strategy to use as an organism from an energy conservation standpoint because the actions of flexion, abduction, and external rotation would lead to lengthening the extensors, adductors, and internal rotators, which may both maximize the length-tension relationship of those tissues, and provide a stretch reflex to assist the forthcoming concentric phase of the movement.

Why Do We Cue Knees Out?

So why do we cue knees out so hard? My first thought is that the coaches who have figured out that this helps in the ability to squat heavy weights are people who have done a significant amount of training. People with extensive resistance training backgrounds are usually fairly easy to spot. Their bodies have undergone fairly stereotypical adaptations. They've clearly undergone extensive amounts of tissue remodeling, and show demonstrable hypertrophy of skeletal muscle. Heavily resistance-trained individuals also seem to move can spot them walking. These individuals seem to present with an anteriorly tilted pelvis more than the general population.

To understand the essence of what kinds of bodily relationships happen with an anterior tilted pelvis, let's use a thought experiment. If you were to prop up someone who was unconscious and you were to anteriorly tilt their pelvis, that individual's femurs would passively follow the anterior tilt of the pelvis, and those femurs would orient internally and towards midline. Ultimately, if you excessively anteriorly tilted an unconscious person's pelvis, they would simply fall on their face (aka, go into prone collapse).

With a conscious human being interacting with the world, things will change, particularly in regards to battling gravity and preventing prone collapse. The typical response of an individual's femurs to a pelvis that is going into anterior tilt, while standing, is that the femurs will reflexively start externally rotating and abducting to prevent prone collapse. This is the typical strategy that will be used by a heavily resistance trained individual, who has an anterior tilted pelvis, to deal with gravity and prevent prone collapse while standing.

Where things get interesting is when we lay this person on a table and measure femoral ranges of motion. When the person lies down on the table, we are removing the need for them to fight prone collapse, and the femurs will stop their reflexive behavior of externally rotating and abducting. This individual's femurs will now simply exist in the passive state of adduction and internal rotation because they're simply in line with the orientation of the anteriorly tilted pelvis.

When I test the person for femoral range of motion on the table, they will demonstrate reduced adduction and internal rotation...why? Because the femur is already internally rotated and adducted. The femur has less room to further internally rotate and adduct. So, when dealing with a heavily resistance trained individual who has an anteriorly tilted pelvis, what you're really dealing with is a situation where the more the pelvis anteriorly tilts, the more the femurs are internally oriented and adducted PASSIVELY, and the more the behavior of the femurs in a weight bearing situation will be one where they move by externally rotating and abducting ACTIVELY to prevent prone collapse.

Regardless of how you slice it, the individual will demonstrate a lack of adduction and internal rotation from a testing perspective, and from a functional movement perspective. Yet this person needs to utilize the actions of adduction and internal rotation with extension of the femur to be able to create the concentric portion of a squat. How does one obtain that which they do not have?

Joint Actions vs. Joint Moments

Picture the following scenario. I am standing with my weight equally on both feet. Both feet are side by side and pointed straight ahead. I see something move to my left. Without moving my feet, I rotate/orient my entire center of mass to the left to look at it. I then hear something slightly to my right, so I adjust myself to examine what made the noise. To do so, I begin rotating back to the right. Interestingly, the noise that I heard was still to the left of my starting orientation in this scenario. Despite the fact that I was rotating back to the right, my center of mass still remained rotated and oriented to the left. This example illustrates the difference between true joint actions and joint action moments.

The true joint actions that took place were vertebral rotations left. When I heard the noise, I started rotating back to the right, but my vertebrae were still rotated left. I was creating the moment of right rotation, but I never got to the point where you could say I was rotated to the right...I was always rotated left. With the heavily resistance trained individual who is squatting, this person, while standing, has an anteriorly tilted pelvis, and their femurs are reflexively externally rotating and abducting. This person overall cannot reach the true joint actions of adduction and internal rotation, but what this person can do is utilize a strategy of further abducting and externally rotating, so that they can utilize the moment of adduction and internal rotation to group with extension to take advantage of the way the femur naturally creates propulsive force.

How would this person do this? They would spread the floor and externally screw the floor with their feet, and push the knees out during the descent to go further into the actions of abducting, and externally rotating their femurs. To reverse the squat and come back up, the person would then begin creating the moment of extending, adducting, and internally rotating...they may never reach the true joint actions of femoral extension, adduction, and internal rotation (likely because they lack those motions), but they can still utilize the propulsive force coming from the moment of that triad to execute the concentric portion of the squat.

When watching these individuals squat, you may never see the knees crashing towards midline...because they likely can't...but they're still relying on the same muscular strategy as someone who is demonstrating medial translation or true adduction in their squat. This strategy may also get confused with some sort of pathological strategy where the femur is moving medially and rotating internally, but the tibia/calcaneus/foot complex remains lateral and externally rotated, which will be discussed more in a forthcoming paragraph.

When dealing with young athletes, these individuals likely haven't gone through many of the adaptations (or perhaps stress responses) that heavily resistance-trained individuals have, and thus likely do not demonstrate as much anterior tilt of the pelvis. These individuals likely can get into the true joint actions of adduction and internal rotation, so when we see knees moving towards midline, we may just be seeing the manifestation of the true joint actions being driven by the muscular moment of those actions.

With elite weightlifters, and other extremely athletic individuals who produce tremendous force during the squat they may also be showing the true joint actions of internal rotation and adduction, and we may be seeing people who have somehow avoided changes that drive a pelvis into anterior tilt that come along with extensive heavy resistance training. With these individuals we may be seeing the most resilient amongst us to consistent heavy resistance training/those who have survived and thrived in the selection process.

The tricky parts to this concept involve the areas that revolve around injury. The mechanism of injury for the ACL is when the knee moves medially and rapidly internally rotates. The catch though is that the injury to the ACL is based on the relationship of the femur to the tibia. There has to be torque, twist, and a difference in direction between the femur and the tibia for this particular knee injury to happen.

Frequently you'll see a situation where the tibia, relative to the femur is translating laterally and externally rotating at that same time as the femur is rapidly translating medially and internally rotating. If people are squatting and maintaining a flat foot with the weight evenly distributed throughout the foot, and the foot/calcaneus/tibia complex is not externally rotating during the descent, then witnessing the true actions of adduction and internal rotation of a femur is probably not cause for concern. When we do start seeing the foot/calcaneus/tibia complex start to spin out with concomitant medial translation and internal orientation of a knee, then we are probably seeing torque between the femur and the tibia, which could be threatening to ligamentous and other soft tissue structures of the knee, particularly if the force and/or velocity of those divergent joint movements is high.

Fully addressing the entire scope of injury mechanics is beyond the scope of this article; however, supposing that any observable movement of the knee towards midline is threatening, dangerous, likely wearing down soft tissues, and is undesirable is perhaps a rush to judgment on something that may be benign, so long as foot, calcaneus, tibia, and femur are synchronously working together in a fairly natural and stereotypical way for the human squat pattern.


To determine whether seeing a knee translate medially is acceptable or not, testing must be done. I would recommend using an Ober's test to determine if the individual has a femur that can extend and adduct. I would recommend using supine and seated femoral internal rotation to determine whether the individual possesses this joint action. I would also recommend using a Thomas test to confirm femoral extension or determine that the femur is extending through some compensatory/pathological method if it was unable to extend in the Ober's test.

If the person passes all these tests, then they possess the true joint actions of extension, adduction, and internal rotation, and this person would probably be fairly safe with demonstrating medial translation of the knee during the ascent of a squat so long as weight bearing on the foot remained appropriate. If someone is lacking these true joint actions, based on test results, and demonstrates medial knee translation during the squat, it's likely that there are some aberrant joint actions that could be problematic.

There are two other specific topics that should be discussed before wrapping up this article. The first is the topic of seeing medial translation during the eccentric portion of the squat. This, to me, would be indicative of someone who would be existing in the state of having an anteriorly tipped pelvis, but is not kicking their femurs into the actions of externally rotating and abducting as their anti-gravity strategy. This is probably a highly untrained person, who is probably going to be fairly unimpressive and lack most of the desirable force production capabilities you'd look for.

During the descent of the squat when the femurs should be flexing, abducting, and externally rotating, the pelvis should also be using the moment involving posterior rotating (aka, pelvic inlet flexion). If the pelvis is struggling with this moment and stays in the position of excessive anterior tilt, and the individual is not utilizing compensatory actions of the femur, you'll see this incredibly unimpressive looking, melting candle, squat presentation. This is undesirable, but we should begin the coaching process by trying to get the individual to gain better pelvic control first before trying to coach the femur.

Coaching Strategy

This brings us to our last topic in this article, which is coaching strategy. Where I would start giving recommendations is to quote Charlie Francis in regards to advice to coaches...think twice and speak once or not at all. When you start thinking that you're smarter than the millions upon millions of years of evolution that led to a human organism standing in front of you executing patterns that are the result of protein behavior that were coded for by a genome and wired up by a nervous system that has figured out the most effective way to guide someone through the complex and multi-faceted environment that they've lived in for their entire life, you're starting to border on being someone who is either way too ego driven, or ignorant of the depth of reality, that you could be problematic. Observe people for a while.

Check your opinions. Provide the feedback that the person needs, but be careful about what you say and how you say cautious and humble. Then let testing guide you. Find your algorithm that takes decision-making out of your mind. Use the previous tests to determine whether the person has authentic joint actions or if they're resorting to compensatory strategies. Second, if you want to figure out what the most energy/electrically/physiologically efficient and effective squat would be, I would recommend looking at it through EMG, but I wouldn't try to see what the highest EMG would be.

In fact, I would look for how I could get the lowest EMG for a squat. If I keep the load constant, and I see that one strategy uses less electrical energy, that should be the least compensatory strategy possible, and theoretically, that would be the best approach. For more on why that is the appropriate thought process regarding EMG readings, see the explanation given in this article.

My concluding thought on coaching and optimal performance with the squat is that the most likely best way to perform the movement would feature performing the pattern in the mid zone of abduction and adduction, and internal and external rotation. If I want the least wear and tear over time, I would want to try to move through the middle of the pathway rather than forcing my way into the boundaries at each side.

So I personally wouldn't start coaching a beginner who possesses the true joint actions of extension, adduction, and internal rotation with the cues of spread the floor, screw the floor out and push the knees out. I would cue them to find and feel the middle of their heels on the descent. I would cue them to keep their heels at the bottom and to find their big toe. To push back up, I would have them continue to find and feel their heels and big toes, and then find and feel their medial arch, and to push through the medial arch to drive back up. I would observe their femoral behavior with those cues for a while, continue to monitor joint actions through table tests, progress them with load and reps and sets over time, and not think I necessarily had the right answers. Lastly, I wouldn't treat everyone the same, and I'd be very careful about giving cookie cutter coaching cues to groups.

Closing Thoughts

To progress an industry or advance the overall information available to practitioners within a discipline, professionals should avoid dogmatic thoughts and behaviors. Often times, the degree to which a concept is entrenched in the collective working body of accepted beliefs can blind everyone in that field. I'm not saying that I'm absolutely correct with my assertions in this article. I may be damn well wrong, and somebody may come along and prove it to me so bad that I feel like an enormous fool. I don't think I'm wrong though. In fact, I think I'm very right on this one. It may take you a minute to see what I'm saying here, particularly with all this junk about true joint actions, and moments, and the difference between being rotated and rotating, and all the other blah blah blah.

So I ask you to consider what I'm saying here in this article before you blow it off and quickly say that I don't know what I'm talking about and I'm going to get people hurt, and the sun won't come up tomorrow, and stand outside my door with torches and pitchforks, because I'm saying that knees moving towards midline may be no big deal, and perhaps is even natural and appropriate...and if you can't do it, maybe you're stuck and there's a problem. I'm not questioning a long-held assumption because I'm looking to cause a controversy or anything else like that. I like trying to see a puzzle for what it actually is.

Active human anatomy is a beautiful puzzle. It can be an extremely vexing puzzle, and often times, what you think you see isn't what you get...but the logic always works if you follow it long enough with the right starting assumptions. Perhaps my starting assumptions are off, in which case this whole thing is wrong, and you can commence throwing a pie in my face. But if my starting assumptions are correct, I believe my Bayesian reasoning throughout this is strong, and you may have to come around to this way of thinking regarding what's actually going on with the squat. Anyways, here's to questioning authority, thinking for yourself, being unafraid of backlash and criticism, and trying hard in the life you've got.

I'd like to thank the Postural Restoration Institute for providing the theoretical foundations for helping me think through the big ideas of this article. They teach the ideas of the differences between being rotated/oriented and rotating. They also teach the idea that if you can't express a joint action, like adduction, it's probably because you're already existing in that position, and therefore are limited in being able to get further into that position. Without the course work and learning I've done with PRI, I would not have been able to conceive of the ideas for this article.

About the Author:  Dr. Pat Davidson

-Director of Training Methodology and Continuing Education at Peak Performance, NYC.

-Assistant Professor at Brooklyn College, 2009-2011

-Assistant Professor, Springfield College 2011-2014

-Head Coach Springfield College Team Ironsports 2011-2013

-175 pound Strongman competitor. Two time qualifier for world championships at Arnold Classic

-Renaissance Meat Head


50% Complete

One More Step...