The Knees-Out Debate – Crossfit and Squats

Shedding Light on the Knees-Out Debate

“Knees-out” is a coaching cue that is quite popular, due to the fact that novice lifters tend to let their knees cave-in during squats. In recent times, this cue has been used not only to fix bad technique, but to encourage the lifter to push the knees outside the path of the feet. This has actually become a source of controversy known as the “knees-out debate”. Is the “knees-out” squat safe, and where did it originate from?

For anyone who’s familiar with Crossfit, I’m sure you’ve heard of Kelly Starrett, the maniac of a physical therapist who is extremely knowledgeable in his field. He is the “mobility expert” in Crossfit, and his website,, is full of videos that offer a lot of solutions for various injuries and issues.

In one of his earlier videos, he uses a bodyweight squat to test out mobility of the ankle. There’s something in it that confuses the hell out of me. Here’s the video.

What I don’t get is, how is the type of squat used in the video safe or effective under heavy load? This by no means is an attack on Kelly Starrett, but I’m totally perplexed here. The squatting-technique that Mr. Starrett advocates is a feet-forward, externally-rotated-femur, knees-outside-of-the-feet position. Like I said before, this technique is the heart of the knees-out debate.

As I mentioned in my buttwink article, shoving the knees out through hip abduction and external rotation would increase the range-of-motion of the hips, as state by Mark Rippetoe in his book, Starting Strength. This is because the femurs avoid jamming into the ASIS, the anterior-superior iliac-spine, which is basically a bony protrusion of the pelvis. The ASIS normally reduces the amount of hip flexion we have, but the knees-out position forces the femurs to go around the ASIS during hip flexion.

Without shoving the knees out, you will fail to achieve enough depth in your squat.

There’s a commonality between Starrett and Rippetoe with both of them advocating knees-out. However, there is still a major difference between the two. Mark Rippetoe advocates the feet to be pointed out at about 30 degrees, and Kelly Starret says to put the feet forward. This is where I have an issue with Starret’s recommendations.

Knees-Out vs. Knees-Neutral?

Knees-out vs knees-neutral. (Knees-out debate).
Let’s take a look at the diagram I drew. Imagine performing a bodyweight squat right now, and having someone stand above you and take a picture while you’re sitting at the bottom of your squat. The images above are what your legs would look like from the other person’s top-down view. In the picture, the femurs (AKA the thighs) are represented by the blue lines, the knee joints are represented by the green circles, and the feet are represented by the red lines. From a top-down view, the shins would not be visible, because the femurs would be covering them.

Anyway… In a squat that Mark Rippetoe advocates, with feet turned out and knees shoved out, your legs would look like the drawing on the left. Notice how everything is in line with each other. In a squat that Kelly Starrett advocates, with the feet pointed forward and the knees shoved out, your legs would look like the drawing on the right. Notice how the feet and femurs aren’t in line.

When the feet and femurs are in line, the only rotation that occurs happens in the hips, which the hips are highly capable of doing that. When the feet and femurs are NOT in line, some sort of rotation happens elsewhere down the chain. Either in the knees or ankles, some twisting must occur to point the feet forward but keep the knees shoved out.

The knees and ankles aren’t built for this kind of movement. They both are designed to flex and extend (and the ankle can perform eversion and inversion), and that’s it. They do have ligaments that help PREVENT this kind of twisting and rotation. But then, if you force those joints to twist while under load, it isn’t your muscles that bear that load during the twisting motion, but it is your ligaments that take it. This is how we run into injuries.

Isn’t this oddly similar to the dreaded “knees-in” squat position that we’re all familiar with? If the knees can get injured when rotated in one direction while squatting, can’t they also get injured when rotated in the other direction while squatting? Twisting is still twisting, right?

I’ve been told that the Starrett-squat technique creates a “more stable base”. The only way I can see it making a more stable base is by displacing some of the load off of the muscles and on to the ligaments. This is the same concept of using athletic tape or powerlifting-suits to provide additional support, except you can easily replace the tape or suit when it breaks down. You can’t readily get new ligaments for your knees when they break down.


In my opinion, there is no knees-out debate. There is simply only safe squatting technique and dangerous squatting technique. The idea of debating the efficacy of a safer squat technique versus a more dangerous squat technique is silly. Thus, the knees-out debate itself is silly as well.

The knee-joint being twisted, whether it’s externally or internally, is bad. One is not better than the other. The hips are made for rotation. The knees and ankles aren’t made for rotation. Keep them in line when you squat. You’ll thank me later when you aren’t getting surgery on your knee ligaments.

Don’t squat with your knees out. Squat with your knees neutral. If your knees cave in, then use the cue “knees-out”.

Bookmark the permalink.


  1. Photo Credits
    Strongman Squat. Author: Artur Andrzej.

  2. Here is Kelly’s response to this.

    It would appear that the feat straint position creates more toque and stability through the hips helping the knees track the same on the way up as down.
    I don’t have my degree in physiotherapy yet but to me Kelly’s idea makes more sense…

    • Just saw your comment, as it was mistakenly marked as spam. I will check out KStarr’s videos later and let you know what I think.

    • Okay, so I watched parts of his response. He and his two comrades more or less advocate knees tracking in line with the feet during a squat, and that the knees-out cue is simply a cue, to prevent lifters from letting their knees cave-in during squats. He talks a lot about externally-rotating and abducting the femurs in relation to hip torque and stability, but while still in the knees-in-line-with-feet position.

      I think the video on MWod that I posted on this blog post is the source of all of this miscommunication. I think a lot of the community has understood this diagnostic test performed in the video (feet straight, knees out while squatting) to be what Dr. Starrett advocates while actually training. I mean, I’ve seen Crossfitters doing this style of squat with a barbell on their back. But, from his response, it appears that he does not advocated the knees-out, feet-straight position while actually training.

  3. It’s a pleasure to post here…I too, like Starrett, have a DPT degree and just like any person concerned about biomechanical efficiency, look at how peeps’ bones stack up during exercise. If they don’t, there’s an unawareness, weakness or imbalance (because bones don’t move themselves right?). Kelly seems like a brilliant guy, as I’ve watched a lot of his videos and don’t have anything to disagree with. One thing I have learned over the years is that if you speak with a lot of academic jargon, the point of your message is lost to many. The benefit of using massive amounts of academic jargon is that you can amaze people into believing you, impressing them with your pool of relevant vocabulary. The drawback is that with this tendency, you become less than able to giving simple, concise instructions that people can really understand and take to heart. That being said, I like Starrett’s inclusion of evidence based practice in his reasonings and only can say that I get exhausted trying to validate every word he speaks in his videos- because there are many. My two cents on the whole squat thing is that any squatting motion will cause shearing forces on the meniscus as the femur translates forward over the tibia (during flexion/extension in CKC). Simple rehab protocols for meniscus repairs (no i’m not addressing the type or direction of tear) call for very little knee flexion or weight bearing the first 4 weeks post-op. Also, MCL sprain rehab calls for very little knee flexion, time depending on the severity of the sprain. This occurs during knee flexion even without valgus collapse. So you can see that knee flexion in squats causes stress on those two passive support structures alone, and when damaged, are precursors to osteoarthritis. If you sit on a couch and do nothing, you’re at risk for osteoarthritis. If you are out pounding concrete and doing deep squats, you’re at risk for osteoarthritis- so have fun and try to achieve your goals as safely you can. Keep your bones stacked up correctly and stay safe. I’m glad this stuff is being talked about. Keep it simple yo! Thanks.

    • Thank you so much for your comment.

      Yes, Starrett’s explanation are “otherworldly”, if there’s even a word for it. I prefer Rippetoe’s explanation – concise and simple (and makes some effing sense).

      As for the meniscus and shear force, wouldn’t the ACL be affected too by this shear force, as it also prevents forward movement of the tibia against the femur? I believe Rippetoe brings this up in his book, Starting Strength.

      But you’re right, at the end of it all, it’s about learning how to do things in a correct manner, which requires simple coaching, and is about making careful progressions – all to avoid injury and increase athletic ability.

Leave a Reply

Your email address will not be published. Required fields are marked *

Please solve the question below before continuing. * Time limit is exhausted. Please reload CAPTCHA.