Diagnosing Your Own Squat: What Knees Caving, Butt Wink and a Falling Chest Actually Mean

Squat coaching has a structural problem: the advice is delivered as cues — 'knees out,' 'chest up,' 'sit back' — when the underlying issue is a diagnosis. Knees caving inward has at least four distinct causes, and the cue that fixes one does nothing for the others. Somebody with genuinely weak glutes and somebody with restricted ankles produce a similar-looking squat and need opposite interventions.

This guide is organized the other way around. Each section starts with what you can see happening, lists what could be causing it, and gives a self-test that separates them. You need a phone, a wall, and somewhere to squat.

Fault 1: knees caving in (valgus)

The knees track inward, usually on the way up and usually worse on the heavier sets. Worth stating up front: a small amount of inward travel under maximal load is common even in strong lifters and is not automatically an injury. What matters is when it is large, asymmetric, or appears at light loads.

The stance row deserves emphasis because it is the one nobody checks. Hip anatomy varies enormously between people — the angle and depth of the hip socket differ, which means the stance that lets one person squat deep with vertical shins physically jams another. There is no universal correct stance. Spend one session cycling through widths and toe angles at light load and find the one where depth comes easily.

Fault 2: “butt wink” (posterior pelvic tilt at the bottom)

The pelvis rolls under and the lower back rounds at the deepest point. This is the most over-panicked fault in lifting, so calibrate first: a small amount of pelvic tilt at the very bottom of a deep bodyweight squat is normal and not evidence of damage. Loaded, pronounced rounding under a heavy bar is worth addressing.

• Hamstring tension is the popular explanation and is usually wrong — hamstrings cross both the hip and the knee, so in a deep squat (knee flexed) they are slack, not taut. Stretching them rarely fixes butt wink.
• Hip anatomy is frequently the real answer: the femur runs out of room in the acetabulum, the pelvis tilts to make space. This is structural. The fix is not to force depth — it is to squat to the depth where your hips actually work.
• Ankle restriction again: limited dorsiflexion forces the hips to go deeper relative to the knees to reach depth, dragging the pelvis into tilt.
• Core stability matters at the margin — bracing well delays the point at which the pelvis gives way.

Fault 3: heels lifting / chest falling forward

These two are usually one problem. To squat with an upright torso, the knee must travel forward over the foot, which requires ankle dorsiflexion. If the ankle cannot supply it, the body finds the range elsewhere: the heel lifts, or the hips shoot back and the chest drops to keep the center of mass over the foot.

Fault 4: the hips rising faster than the shoulders

The hips shoot up out of the bottom and the squat turns into a good-morning. This is almost always one of two things and they are easy to separate:

• The weight is too heavy for your quads, so you unconsciously shift it to the stronger hip extensors. Test: does it happen only on the top sets? Then it is a load problem, and the fix is to drop weight until the pattern holds, then rebuild — plus direct quad work (front squats, leg press, split squats) to close the gap.
• A bracing or timing failure, where the torso collapses forward before the legs begin extending. Test: does it happen even at light load? Then it is technique — think about driving the *chest* up rather than the hips, and pause squats at the bottom will teach the position better than anything else.

Bar position changes the whole geometry

Before diagnosing a fault, check that you are not comparing yourself to a different lift. High-bar and low-bar squats are mechanically distinct, and half the arguments about correct squat form are two people describing different exercises.

The practical consequence for someone with limited ankle mobility: switching to low-bar is a legitimate answer, not an admission of defeat. It shifts demand from the ankle to the hip, which is why lifters with restricted ankles often find it immediately more comfortable. Heel elevation solves the same problem in the other direction. Both are valid; picking one deliberately beats fighting a high-bar squat your ankles cannot support.

Bracing, briefly

Several faults above list bracing as a contributor, so it is worth defining. Bracing is not sucking the stomach in — that reduces stability. It is taking a breath into the belly and pressurizing the abdominal cavity outward against the torso wall, creating a rigid cylinder that resists the spine folding forward. The cue that works for most people: breathe in as if filling the lower ribs and stomach, then tense as though about to be punched, and hold it through the rep. Exhale at the top, not at the bottom.

What actually matters, ranked

Having gone through four faults in detail, the honest hierarchy is worth stating, because it is easy to over-index on form and never train:

1. 1Consistency. A mediocre squat performed for two years beats a perfect squat performed for three weeks. Nothing on this page outranks showing up.
2. 2Load progression. The progressive overload rules are what produce adaptation. Form is the container; overload is the content.
3. 3Repeatable depth and control. A squat you can hit the same way every rep — whatever depth that is — is worth more than chasing a range you cannot own.
4. 4Cosmetic form details. Bar path aesthetics, exact stance width, whether you look like a powerlifter. Real, and last.

Common mistakes

• Applying a cue without a diagnosis. 'Knees out' fixes one of four causes and wastes the other three.
• Forcing depth your hips do not have. Anatomy varies; load the range you control.
• Stretching hamstrings for butt wink. Wrong muscle for the joint position.
• Never filming. The gap between how a squat feels and how it looks is the single largest source of self-deception in lifting.
• Treating heel elevation as cheating. It is equipment that solves a real constraint.
• Diagnosing pain. Everything here is about movement quality. Actual pain — sharp, joint-line, or persistent — belongs with a physical therapist, not an article.

Who should adjust

Anyone with existing knee, hip or lower-back injury should get a movement screen from a physical therapist before self-diagnosing from video — the faults above have overlapping presentations with several pathologies. People with hip impingement (FAI) frequently have a genuinely limited depth that no mobility work changes, and forcing it makes things worse. And for complete beginners, the highest-value intervention is not any of this: it is a handful of sessions with a competent coach who can see what a camera cannot.