How to Fix a Rounded-Back Deadlift: Setup, Pattern, and Measurement-Based Correction

The rounded-back deadlift is the single most common form fault in resistance training and is far more layered than 'just bad form'. McGill (2007) and Cholewicki et al. (1991) showed in vivo that lifting 200 kg with roughly 15° or more of thoracic flexion exposes the posterior annulus to 4–6× the shear stress of an identical lift performed neutral. That said, not every rounded-back lift becomes an injury: some elite deadlifters intentionally adopt a stiff, controlled thoracic flexion to lift near a thousand pounds (Vigotsky & Bryanton, 2016). The decisive distinction is between stiff flexion and passive collapse. This guide is a staged correction protocol for converting the latter into the former, grounded in 800Hz IMU spinal segment angle data. We cover setup reconstruction, mobility recovery, pattern relearning, and measurement-driven progressive loading.

The cause of a rounded back is almost always a combination of three axes. First, setups that begin with thoracic flexion already present. Second, insufficient thoracic extension or hamstring/hip mobility that prevents holding neutral. Third, missing co-contraction patterns of the lats and anterior core, so the spine collapses the moment load is recognised. Each axis must be evaluated separately, and the largest contributor must be addressed first.

Most cases are mixed. But by attaching 800Hz IMUs at T7 and L3 and recording segmental angle change throughout the lift, the exact moment and segment of collapse becomes visible. Correction without diagnosis is gambling.

The most common true cause of rounded back is setup. If the thoracic spine is already flexed at the bar, that flexion almost never disappears under load. The PoinT GO lab's 5-point setup reconstruction is as follows.

1. Foot position: bar over mid-foot, heels at jump-stance width.
2. Hip height: above the knee, below the shoulder. Too low becomes a squat; too high rounds the back.
3. Lat tension: cue 'pull the bar into your body' to engage lats; the rib cage stacks over the pelvis.
4. Rib cage position: not 'chest up' but 'don't flare; stack ribs over pelvis' is the more accurate cue.
5. Breath pressure: at the final setup beat, Valsalva to pressurise abdomen and rib cage simultaneously.

Apply these in order, not all at once. Each point should be automated under light loads (<50% 1RM) before adding intensity. The perfect deadlift setup guide has detailed visual cues.

Even with a sound setup, mobility deficits will collapse it. The following 4-week program addresses thoracic extension, posterior chain mobility, lat-core co-contraction, and progressive loading.

Week 3 pattern relearning is critical: athletes whose back rounds as load increases usually never built a precise segmental pattern at light loads. Cues from the Romanian deadlift guide apply directly here. For mobility evaluation see the hip mobility assessment.

The final correction stage is measurement-based progression. Do not rely on 'feels okay'. The PoinT GO lab tracking protocol uses two IMUs (T7 and L3), capturing four-phase segmental angles (setup, lift initiation, knee passage, lockout) on every warm-up and working set. Pass criteria are:

• Absolute T-spine flexion at setup: < 10°
• T-segment angle change throughout the lift: < 5° green, 5–8° amber, > 8° red
• L-segment angle change: hard ceiling at 5°

Across the 4-week program, expect mean segmental change to drop 1–2° per week. If load increases without the change exceeding 5°, advance. If any single set crosses 8°, drop load 10–15% immediately. This protocol shares the same philosophy as why your squat is not getting stronger: objective data catches form faults faster than subjective sensation. A rounded back is not a willpower problem; it is a measurement and system problem.