Squat Technique, Stance and Depth

Published October 23, 2026 · Last reviewed October 23, 2026

The one adjustment: let your hips pick the stance

Hip sockets and thigh bones differ in shape from person to person, so no single stance fits everyone. NSCA Performance Training Journal 5.4 says so directly and recommends a trial-and-error process: start from a baseline stance, widen it slightly with the same toe angle, then try other toe angles, and keep the combination that gives the best depth with the fewest compensations. Comfort decides the final choice, provided the known faults (knees collapsing in, lower-back rounding, heels lifting) are absent. The same article adds that pain is never a normal response to squatting.

The squat is also a skill that responds to load. NSCA Performance Training Journal 10.2 (2023) notes that adding appropriate load early in teaching tends to promote proper technique, and that a touch-and-go box squat and a goblet squat are suitable starting points before a free squat.

Action

Test your stance. Do three controlled squats at about shoulder width with toes turned out slightly. Widen by one small step, keep the toe angle, repeat. Then change the toe angle at each width. Keep the setup that reaches your depth with heels flat, knees tracking over the feet and lower back still.

Why it helps

Hip joint shape varies, so the stance that suits one lifter can pinch or round another (NSCA, 5.4). In one review, a stance of 150–200% of shoulder width was linked to higher gluteus maximus activity in some studies and to higher knee valgus loading than narrower stances; turning the toes out 30° lowered the knee valgus moment by about half, with a rise in the varus moment, in the work it summarized (Straub and Powers 2024). Those are group findings, not a rule for you.

Verification

Film three reps from the front at knee height and three from the side at hip height. From the front the knees follow the line of the feet; from the side the heels stay on the floor.

Action

Set your depth at the deepest point where the heels stay down and the pelvis does not visibly tuck under. Place a box or bench at that height as a target for working sets.

Why it helps

The limit on deep squatting is the hip flexion your joints allow. Past that point the pelvis tilts backward and the lower back flexes, which the review links to added compressive and shear loading on the spine (Straub and Powers 2024). A scoping review of 15 studies concluded that deep squats appear safe for healthy knees when technique is kept and load and volume are not excessive; 13 of the 15 found no added injury risk, though most studies were short (6–10 weeks) and small (Rojas-Jaramillo et al. 2024). In one small study of 18 experienced lifters, more ankle dorsiflexion range was the only measured factor linked to more lower-back flexion; hip range, body proportions and movement-control tests were not (Berglund et al. 2024). So the cause differs between people, and video is the way to find yours.

Verification

On side-on video the lower back keeps the same shape from the top to the bottom, and the bar path stays over the middle of the foot.

Action

Lower for about 3 seconds, pause for 1 second in control, then stand. Use this tempo on warm-up and technique sets.

Why it helps

A controlled descent gives you time to notice a change in position. Repetition durations of 0.5 to 8 seconds produced similar muscle growth in a meta-analysis of eight trials, so a 3-second lowering sits inside a range that does not cost results (Schoenfeld et al. 2015). Not reported: none of the sources reviewed shows that a slower descent lowers the risk of knee injury.

Verification

You can name your position at the bottom, and the video shows no sudden drop or bounce out of the hole.

Common faults and what to try

• Heels lift. Listed as a common error in the NSCA technique deck. Try a slightly wider stance or more toe-out, a box target, or flat shoes; a raised heel moves the shin forward and raises the demand on the quadriceps (Straub and Powers 2024). See equipment and gym setup.
• Knees fall inward. Check stance width and toe angle first, then reduce the load. The NSCA deck lists knee varus and valgus as a common error.
• Lower back tucks under at the bottom. Raise the depth target by one box height and film again. The cause may be hip range, ankle range or something else (Berglund et al. 2024).
• Hips rise faster than the chest. Drop to a goblet squat and rebuild the pattern before adding load again.
• Knees pass the toes. The NSCA deck lists it as an error to watch, while the squat review explains that a more forward shin shifts demand toward the quadriceps. Treat it as a prompt to check heel and back position, not as a law.

By reader profile

For A Start with a touch-and-go box squat or goblet squat, add load early in small steps, and work on one cue at a time.

For B Film one working set a week from the side. Keep the tempo on the first two sets and note any week where the depth marker moves.

For C In a small observational study of 23 competitive lifters (9 women) tested at 70% of one-rep max, all of them moved the upper lumbar spine to a less lordotic position at the squat start compared with standing, while the lower lumbar change appeared in the men only (Bengtsson et al. 2023). The studies reviewed here tested around 70% of one-rep max, so what happens at heavier loads is Not reported; film your heaviest sets separately.

For D Offer one or two cues at a time, and demonstrate three to five reps from the front and the side before the lifter starts (NSCA 10.2). See coaching cues.

For E Clinicians adjust stance and depth for conditions such as patellofemoral pain and hip impingement, and the squat review advises restricting depth to the hip range available for people with impingement (Straub and Powers 2024). Ask your clinician which stance and depth are cleared for you, and do not self-prescribe.

Stop and get it checked

Stop for sharp pain, a pop, swelling, a joint that gives way, numbness or tingling, or hip pain at the bottom of the squat. See a qualified clinician if soreness lasts beyond a few days or rises from one session to the next.

Most of the laboratory studies cited here tested few or no women, and none of them measured injury rates. Where a finding comes from men only, a small group or a single measurement session, the text says so. Any claim that technique differs by sex is Not reported unless a source below states it.

How it works

Three things you can change shape the load on the knees, hips and spine: stance, how far the shin and trunk lean, and depth. The review in the sources describes how each shifts the turning force (moment) at the joints. A forward shin raises the knee moment and quadriceps demand, a more upright shin lowers it, and a more forward trunk lean shifts demand to the hips. The review notes that research findings conflict mainly because many studies do not control the other variables, so a result for “wide stance” can change if trunk lean changes with it.

That is why a stance test with video beats a fixed rule. You hold depth and load steady, change one thing, and see which setup keeps your heels, knees and lower back in position.

Worked example

All numbers are hypothetical and show how to run the test, not a target.

Say your shoulder width measures 16 in (about 41 cm). The squat review defines a narrow stance as 75–100% of shoulder width (12–16 in), medium as 100–150% (16–24 in) and wide as 150–200% (24–32 in). You decide to test 16, 18 and 20 in at toe-outs of 15° and 30°, which is 3 widths × 2 angles = 6 trials of 3 reps each, or 18 reps in all.

After each trial, score three checks from 0 to 1: heels stayed down, knees followed the feet, lower back stayed still. Trial scores might read 16 in/15°: 2, 16 in/30°: 3, 18 in/15°: 2, 18 in/30°: 3, 20 in/15°: 1, 20 in/30°: 2. Two trials tie at 3, so choose the one that felt easier on the hips, and re-test next week.

Tempo check: 5 reps at a 3-second lowering, 1-second pause and 1.5-second rise take 5 × (3 + 1 + 1.5) = 5 × 5.5 = 27.5 seconds. Time a set once so the tempo is real, not assumed.