If you filmed your squat from the side last week and saw your pelvis tuck at the bottom, you are probably reading the third article today and they keep contradicting each other. One says butt wink will herniate your discs. One says it is a non-issue. One sells you a $200 mobility course to fix it. None of them tell you whether your squat — the one in the video on your phone right now — has a real problem.
Butt wink, technically, is posterior pelvic tilt with mild lumbar flexion at the bottom of a squat. In plainer terms, the spine doing what the hip cannot — like a tape measure that curls at the end of its reach.
This article does not pick a side in the harmful-or-not debate. It gives you the rules to pick yours. Four things follow: what butt wink really is, why it happens, what the research shows about the harm narrative, and a 60-second side-view self-check that puts your specific squat into Green, Yellow, or Red. There is also a contrarian claim no other top-ranking article makes — chasing the fix can hurt you in its own way.
What butt wink really is (and what it isn't)
Butt wink is the posterior tilt of the pelvis combined with slight lumbar flexion at the deepest point of a squat. It is not a global "rounding" of the back. It is not the same thing as losing your brace. It is also not what is happening when a lifter's whole upper back collapses forward — that is thoracic flexion, a different problem with a different cause.
In a textbook squat, the pelvis stays relatively neutral all the way down. In a real squat, that does not happen for very long, because the hip joint runs out of clean flexion range and the body looks for motion elsewhere. The pelvis tucks. The lumbar curve shallows. This is part of normal lumbo-pelvic rhythm — the same mechanism that lets you touch your toes when your hamstrings are tight. The spine takes over where the hip stops.
A useful reframing from Austin Baraki, MD, of Barbell Medicine: "Everyone experiences a degree of lumbar flexion during the squat, whether it's visible or not." The question is not whether your spine flexes. It does. The question is whether the amount, the timing, and the load combine into something worth caring about.
One more clarification, because this is where lifters confuse themselves: butt wink is a side-view phenomenon. You cannot reliably feel it from the inside. Your proprioception at end-range hip flexion under load is not good enough to tell a 5-degree tilt from a 15-degree tilt. If you have never filmed your own squat from the side, you do not actually know what your hips are doing.
Why it happens — three real causes, not seven
Most articles list six or eight causes. Three matter.
Hip socket anatomy. Where your acetabulum sits in the pelvis and the angle of your femoral neck determine how deep you can flex the hip before the femur runs into the rim of the socket. Catelli et al. (2018) compared three groups — symptomatic FAI, asymptomatic femoroacetabular deformity (FAD), and controls — and found maximum squat depths of 39.4%, 30.0%, and 27.1% of leg length respectively (lower numbers = deeper squats). The FAD group also showed greater posterior pelvic tilt at the bottom and significantly stronger hip extensors than the other two groups. Translation: they "winked more" and squatted deeper, with no symptoms. Their anatomy got out of the way of impingement by tilting the pelvis.
So if a lifter on Instagram with a different hip anatomy squats butter-smooth at depth and you do not, the difference may not be effort or mobility. It may be the shape of the socket itself, which no amount of stretching changes.
Ankle dorsiflexion. Hemmerich et al. (2006), studying populations habituated to deep squatting, measured the mean maximum ankle dorsiflexion required for a heels-down deep squat at 35.4° ± 5.5°. Clinical norms for passive dorsiflexion in untrained adults sit at 15–20°. That gap matters. If the ankle bottoms out at 20° and the squat needs 35°, something else has to give — usually the pelvis tucking forward as the torso falls and the lumbar spine takes the slack. This is the diagnostic Aaron Horschig, DPT, of Squat University runs first: "If an athlete has very stiff ankles and they go into a deep squat, they're going to hit their end range of ankle mobility early on in the descent and, in order to continue squatting deeper, their pelvis will have to move excessively as a compensation." (For a related side-view issue that often co-occurs with ankle limits, see our knee valgus guide.)
Asking the hip for more depth than it has. This one is relative, not absolute. A squat to parallel may be well within your hip range. A squat to ass-to-grass may be well past it. The wink shows up at the depth where you have run out of clean hip flexion. The "cause" here is not weakness — it is geometry.
Two causes you will see in other articles that we are not putting on the list: "weak core" and "tight hamstrings." Neither Marc Surdyka nor Aaron Horschig — the two most-referenced DPTs in this conversation — names either as the lead driver. Hamstrings do not stretch across the lumbar spine. Bracing tighter does not change hip socket geometry. They are convenient explanations, not accurate ones.
Is butt wink actually harmful?
This is the section every article should have and most do not, so we will sit in it for a while.
The "butt wink wrecks your discs" narrative traces, almost entirely, to one line of research: Callaghan & McGill (2001), "Intervertebral disc herniation: studies on a porcine model exposed to highly repetitive flexion/extension motion with compressive force." They took porcine cervical spine motion segments — dead pig necks, essentially — mounted them in a custom jig, and ran them through axial compression with pure flexion/extension moments at 1 Hz. At modest compressive loads, herniation showed up after 22,000–28,000 cycles. At higher loads, 5,000–9,500 cycles. That is the finding the harm narrative is built on.
It is a real finding. The mechanism — repeated flexion under compression damaging the annulus — is genuinely demonstrated in that model. But look at the dose.
A lifter doing 5x5 squats twice a week is doing about 50 cycles. Per week. In living, adapting tissue. The pig spines were dead, refrigerated cervical segments cycled the equivalent of an entire training career, in hours, in a jig. The mechanism is real; the dose is not relevant to any recreational lifter on any reasonable program. Marc Surdyka, DPT, of E3 Rehab puts the point bluntly: "The idea that it is harmful actually comes from research performed on cadaveric pig spines, in which the researchers flexed and extended the spines thousands of times under load until damage occurred. This has never been reproduced in humans, especially in a clinical setting."
So the question becomes: when, if ever, is butt wink in a living lifter actually a problem? Working from the lumbo-pelvic rhythm framework McGill (2007) lays out in Low Back Disorders, the risk window is the combination of four conditions: high compressive load, high lumbar flexion angle, repeated cycles, and no adaptation. A lifter at moderate loads, end-range pelvic tilt, and a sane training frequency does not hit those four together. A lifter doing heavy singles at 90%+ 1RM with early-descent flexion, pain history, and progressive worsening rep-to-rep is a different conversation.
Even research that looks like it should support the panic narrative quietly does not. McGill & Marshall (2012) documented meaningful lumbar motion in kettlebell swings and other lifts that looked outwardly neutral to the eye — meaning some lumbar flexion happens whether you can see it or not. (We did not find a verified word-for-word "26° lumbar flexion" figure in the abstract, despite that number floating around the internet; treat it as a directional finding, not a quoted number.) The eye, in other words, is not the right instrument for deciding whether a squat is safe.
In early testing of our AI squat form check tool, the pattern we see most often is butt wink that appears only in the final 5–10 degrees of descent and does not deepen rep-to-rep across a working set. The pattern that warrants a closer look is rarer: tilt that begins well above parallel, deepens with each rep, or shows up alongside pain. Two different phenomena, often called by the same name.
So is butt wink harmful? In the form 90% of lifters are worried about — end-range, painless, non-progressive, moderate load — the honest answer from the available evidence is no.
Forcing the cure has its own cost
Here is the part no other top-ranking article on this keyword says out loud.
When a lifter discovers they have butt wink and decides to eliminate it, the standard advice is some version of "brace harder, drive your chest up, stay tall." The unstated assumption is that the spine should be rigid through the full squat and that the cure has no side effects. Both assumptions are wrong.
Lumbo-pelvic rhythm is not a bug. It is how the system works. The pelvis tilts because the hip has run out of flexion, and forcing the pelvis to stay neutral does not give the hip more range — it just makes something else take the load. Lifters who over-cue thoracic extension to "stay tall" through the bottom of a heavy squat tend to migrate the rounding upward, generating rib flare, lumbar hyperextension under brace, and increased reliance on passive structures (ligaments, the anterior longitudinal ligament, the rib cage itself) to hold position. None of those tissues love being load-bearing.
Is this a documented clinical claim from a named DPT in those exact words? We did not find that quote. The closest steelman is Surdyka's general anti-pathologizing position and the implicit logic of the lumbo-pelvic rhythm literature: if motion at one segment is normal, suppressing it forces compensation elsewhere. Treat this section as our reasoning from the framework, not a quoted clinician position.
The practical version is this. If your butt wink is end-range, painless, and non-progressive, the smart move is often to leave it alone. The opportunity cost of chasing rigidity — months of mobility work that does not change your hip socket, cues that drive other dysfunctions, training time spent below your real working depth — is real even when nobody adds it to the ledger.
The 60-second self-check
Pull up the side-view video. Watch it twice. Run these four checks.
1. Timing. When in the descent does the pelvis first tuck? If it tucks well above parallel — meaning the tilt begins early in the descent — that is a red flag. The hip should have plenty of range left to give at that depth. If the tilt only appears in the last 5–10 degrees of descent, that is end-range tilt, and almost certainly fine.
2. Degree. Is the lumbar curve disappearing entirely, or is the pelvis tucking slightly? A rough cue: if the pelvis rotates more than a hand-width of visible change, that is significant tilt. Subtle tucking that you only see on slow-motion replay is the normal range.
3. Symptoms. Any sharp pain during the lift, lingering ache afterwards, or stiffness the next morning? If yes, escalate regardless of how mild the tilt looks on video. Pain is information; tilt is geometry. They are different signals.
4. Load. At what percentage of your 1RM does the tilt appear? End-range tilt at 60% 1RM is normal mechanics. The same tilt at 90% under fatigue, high-frequency programming, or progressive worsening through the set is a different conversation — not a panic, but a closer look.
Take the four answers together:
- Green — late timing, mild degree, no symptoms, moderate load: do nothing. Film monthly to confirm it is not progressing.
- Yellow — late timing, moderate degree, no symptoms, but loads above 80%: trim depth to where the tilt disappears, widen stance and turn the toes out 5–10°, or try a 5 mm heel lift. Do not "brace harder."
- Red — early-descent timing, large degree, any symptoms, or progressive worsening: stop loading heavy. See a physical therapist. This is not a cueing problem.
If you would rather have a tool measure these dimensions automatically from a side-view phone video, that is what our squat form check does — it scores the same four dimensions we just walked through, so you do not have to eyeball the timing.
What to do, by tier (not generic fixes)
The standard advice list — "do ankle mobility, do couch stretch, brace harder" — treats every butt wink as the same problem. It is not. What you should actually do depends on which tier the self-check put you in.
Green tier: keep training. Do not chase the wink. Film monthly. If nothing changes, nothing has changed.
Yellow tier: change a variable, not your bracing strategy. Three options in order of effort:
- Reduce depth to just above where the tilt first appears. Train there for 4–6 weeks. The wink is the spine doing what the hip cannot — squatting just above that point removes the mechanical cause.
- Widen stance and turn toes out 5–10°. This reduces the hip flexion angle required for any absolute depth, the same accommodation the asymptomatic FAD group in Catelli et al. (2018) used.
- Try a 5 mm heel lift (a small lifting shoe or olympic shoe). If the tilt disappears at depth, that is information — your butt wink was ankle-driven, and Hemmerich's 35° threshold was the actual limiter.
Red tier: stop and get evaluated. Pain or early-descent flexion under load is not a problem you cue your way out of. See a PT before you change your squat. This is one of those situations where the smartest move is to let someone with hands on the joint look at the joint.
FAQ
Does butt wink cause disc herniation?
No human evidence supports that claim. The original source of the disc-damage narrative is Callaghan & McGill (2001), which used porcine cervical spine segments cycled 22,000+ times in vitro. The mechanism (repeated flexion under compression damaging the annulus) is real in that model, but the dose (tens of thousands of cycles in dead tissue) does not translate to recreational lifters. Marc Surdyka, DPT, has been explicit: this has never been reproduced in humans clinically.
Is butt wink worse on low-bar vs. high-bar squats?
Greg Nuckols, MA, has noted in his Stronger by Science squat coverage that butt wink is more associated with low-bar squats than high-bar, mechanically because the low-bar position pushes the torso forward and the lifter ends up at end-range hip flexion at a different depth than they would with a high-bar squat. It is not that low-bar is dangerous — it is that the geometry exposes hip-range limits at a different point.
Do weightlifting shoes fix butt wink?
Sometimes. A raised heel (typically 12–22 mm in olympic shoes) reduces the dorsiflexion demand at the ankle, which removes ankle limitation as a cause. If your butt wink was ankle-driven, the shoes will help. If your butt wink was hip-anatomy-driven, the shoes will not change much. The self-check tells you which one you are.
Should I stop squatting if I have butt wink?
Almost certainly not. Stop squatting heavy if you have pain, early-descent flexion, or progressive worsening across a set. Otherwise, you are looking at normal lumbo-pelvic rhythm and you would be removing one of the highest-return exercises in your program for a cosmetic issue.
Can you have butt wink with perfect form?
Yes. Form is not a binary. The Catelli et al. (2018) data is a clean example — the asymptomatic FAD group winked more than the control group while squatting deeper, with stronger hip extensors and no symptoms. Their "perfect" squat included tilt. If your definition of perfect form excludes any pelvic motion, your definition does not match the kinematics literature.
What this means
Butt wink is a side-view observation, not a diagnosis. The four-point self-check — timing, degree, symptoms, load — sorts your particular squat into green, yellow, or red, and that sorting is what tells you what to do. End-range, painless, non-progressive tilt at moderate load is normal mechanics. Early-descent tilt, pain, or progressive worsening across a set deserves attention. Most lifters reading this fall into the first group and have been worrying about a non-problem.
This week: film one working set from the side, run the 4-point check, and if you would rather not eyeball the timing and degree yourself, drop the clip into our squat form check. And if there is pain — sharp, lingering, or worsening — book a PT visit before changing anything about your squat. The tool measures geometry, not tissue. A clinician does the latter.
Sources
- Callaghan, J. P., & McGill, S. M. (2001). "Intervertebral disc herniation: studies on a porcine model exposed to highly repetitive flexion/extension motion with compressive force." Clinical Biomechanics, 16(1), 28–37.
- McGill, S. M., & Marshall, L. W. (2012). "Kettlebell swing, snatch, and bottoms-up carry: back and hip muscle activation, motion, and low back loads." Journal of Strength and Conditioning Research, 26(1), 16–27.
- Catelli, D. S., Kowalski, E., Beaulé, P. E., Smit, K., & Lamontagne, M. (2018). "Asymptomatic Participants With a Femoroacetabular Deformity Demonstrate Stronger Hip Extensors and Greater Pelvis Mobility During the Deep Squat Task." Orthopaedic Journal of Sports Medicine, 6(7).
- Hemmerich, A., Brown, H., Smith, S., Marthandam, S. S. K., & Wyss, U. P. (2006). "Hip, knee, and ankle kinematics of high range of motion activities of daily living." Journal of Orthopaedic Research, 24(4), 770–781.
- McGill, S. M. (2007). Low Back Disorders: Evidence-Based Prevention and Rehabilitation (2nd ed.). Human Kinetics.
- Surdyka, M. "The Truth About Butt Wink." E3 Rehab.
- Horschig, A. Commentary in "Squat University founder Aaron Horschig on butt wink." BarBend.
- Baraki, A. "Butt wink and injury risk." Barbell Medicine forum.
- Nuckols, G. "How to Squat: The Definitive Guide." Stronger by Science.
We are experienced lifters who built an AI form-check tool, not licensed clinicians. The biomechanics claims in this article trace to the researchers and clinicians cited in Sources. If you have pain during squatting, see a physical therapist before changing your technique.