How to Improve Sled Push Power: Load Curves, Posture Angles, and IMU Prescription

The sled push is the most direct stimulus available for an athlete's horizontal force production, mirroring the mechanics of the first ten metres of sprint acceleration. Morin et al. (2017, Int J Sports Physiol Perform) showed that heavy sled push (load ≥ 80% body mass) develops horizontal force (F0) significantly more than light sled work, and Cross et al. (2017) reported that power output peaks at the load that drops the athlete to roughly 50% of unloaded velocity. This guide is not about running or cardiovascular measurement; it is about power production from the push posture and how a coach can use sled push to develop explosive first-step power, lineman contact power, and rugby scrum drive. With an 800Hz IMU at the pelvis, propulsion angle, per-step horizontal impulse, and L/R symmetry are all trackable. We cover load-velocity-based prescription, angle measurement and correction, and an 8-week progression, all data-first.

The single decisive variable in sled push is propulsion angle. The angle of the trunk to the ground sets the direction of the propulsion vector, which sets per-step horizontal impulse. Effective angles run 25–35° under heavy loads (almost prone) and 40–55° under light loads (mirroring sprint acceleration). Too upright shifts force vertically; too low loses leg-cycling space.

One load only stimulates a slice of the horizontal force-velocity curve. Comprehensive development requires rotating loads across blocks, with prescription accuracy depending on the propulsion angle the IMU actually measures.

The load-velocity curve in sled push is individual, and so is the load that maximises power. Cross et al. (2017) reported peak power at roughly 50% of unloaded velocity, typically corresponding to 70–90% body mass. Sprint capability, leg length, and neurological profile shift this range ±20%. The PoinT GO measurement protocol:

1. Push 10 m at 4–5 loads (30%, 50%, 70%, 90%, 110% BM), 2 trials each.
2. Auto-extract mean horizontal velocity from a pelvic IMU.
3. Fit a linear F-v regression: F = F0(1 − v/v0). Estimate F0 and v0.
4. Power-peak load = F0/2 or v0/2.

The logic mirrors 1RM estimation (see 1RM calculation methods) and is part of the broader load-velocity profiling paradigm. Training within ±10% of the estimated power-peak load consistently produces the largest power gains.

The following 8-week program produced a mean 18% gain in horizontal power in the PoinT GO cohort (n=22, rugby/football athletes).

The key is including both the heavy block (weeks 5–6) and the light block (week 7), avoiding the trap of single-load prescription. If the first push of any session shows trunk angle 5° steeper (more upright) than the 4-week mean, drop the next set's load by 10% immediately. Posture collapse correlates directly with injury risk. Complement with trap bar deadlift power and hex bar jump squat, which build a non-specific vertical-power base.

Without measurement, sled push posture drifts week to week. The PoinT GO operational rules: First, start with a single pelvic IMU. The most stable mount is the sacral region between PSIS markers. Second, compare mean trunk angle of the first three steps vs the last three of every set. A > 7° gap means either drop the set from analysis or reduce load on the next set. Third, treat an 8% drop in mean per-step horizontal impulse vs the 4-week mean as a deload trigger.

L/R symmetry usually runs < 8% per-step impulse difference. Above 12%, unilateral weakness becomes an injury liability and unilateral assistance work is required. Finally, the sled is a uniquely good schoolroom for measurement. The interaction of posture, load, and output is unusually visible, and data-driven prescription delivers fast, observable wins. For deeper adaptation monitoring see the athlete testing battery guide and why form breaks down on heavy sets.