Sled Push Load-Velocity Testing: A Field Protocol That Beats Guessing

Most sled sessions still get loaded the same way: someone stacks two plates on each side, tells the athlete it's about 40% of body mass, and calls it a training zone. Nobody actually knows what that load is doing to force or velocity output, because nobody measured either one. Cross et al. (2017, Int J Sports Physiol Perform) tested rugby athletes across a full range of sled loads and found that the load which maximized mechanical power varied by roughly ±20% of body mass between individuals, even though the group mean landed near 70-90% BM. A fixed percentage-of-body-mass rule can put one athlete right on their power peak and another 20-30% away from it, and there's no way to tell which without a load-velocity test. This guide walks through the field protocol: how many loads to test, how far to push, how long to rest, and how to turn raw velocity numbers into an F0-v0-Pmax profile you can actually use.

The sled push produces a near-linear relationship between the resistance it presents and the velocity an athlete can push it — heavier load, slower push, in a pattern that is remarkably individual. Two athletes of identical body mass can post nearly identical push velocity at 20% BM and then diverge sharply by 80% BM, because one of them is limited by maximal horizontal force output and the other by the rate at which they can express it. That difference is exactly what a load-velocity profile (LVP) is built to capture, and it is invisible if you only ever train at one fixed load.

The practical payoff is prescription accuracy. Once an athlete's force-velocity relationship is known, the load that maximizes mechanical power output — not force alone, not velocity alone — can be calculated directly rather than assumed from a body-mass percentage table. Pareja-Blanco et al. (2017, Scand J Med Sci Sports) demonstrated the same principle in barbell training: prescribing load from measured velocity rather than a generic percentage produced more consistent training stimulus across a squad of 63 resistance-trained lifters split across velocity-loss protocols. The sled push version of that logic simply swaps a barbell for a loaded sled and a linear position transducer for an IMU strapped to the athlete's pelvis or waist belt.

You need four things before an athlete takes a single push: a sled with a known, adjustable resistance, a flat surface you can hold constant across sessions, a way to add load in known increments, and a velocity-measurement device. A pelvis-mounted IMU sampling at 800 Hz or higher (PoinT GO's sled protocol runs at 800 Hz) is the most field-practical option; a laser or radar gun works too, provided it stays locked onto the sled or the athlete's center of mass for the full push distance.

Mark a straight 10 m testing lane on the chosen surface — turf, rubber track, or gym floor — and note the surface for every future retest, because friction alone can shift push velocity 10-25% at an identical load. Weigh the sled itself plus every plate combination you plan to test so that load percentages are calculated from true total resistance mass, not just the plates added. Before the first loaded push, have the athlete complete a 10 m unloaded calibration push at moderate effort so you can confirm the sensor is capturing a clean velocity trace before committing to maximal trials.

The protocol below uses five load conditions spanning roughly 20% to 100%+ of body mass, which is enough spread to fit a reliable linear regression without turning the session into an all-day event.

1. Warm-up: 10 minutes of general movement plus two unloaded 10 m sled pushes at increasing effort, finishing at a near-maximal push.
2. Load order: Test from lightest to heaviest load so fatigue does not contaminate the lighter, faster trials, which are the most sensitive to small technical breakdowns.
3. Trials: Two maximal-effort 10 m pushes per load, taking the faster of the two as the value for that load.
4. Rest: 3 minutes between trials at the same load, extending to 4-5 minutes once loads exceed 60% body mass, since heavier pushes recruit more total muscle mass and take longer to recover from.
5. Data capture: Record mean horizontal push velocity across the full 10 m for every trial, not just the fastest instantaneous point — the load-velocity regression uses average, not peak, velocity.

Five loads is a practical minimum. Dropping to three loads still allows a regression line to be drawn, but the fit becomes far more sensitive to a single mistimed trial, so five is the standard PoinT GO recommends whenever session time allows it.

Each load condition converts to an estimated horizontal force using the resistance mass, the surface friction coefficient, and the measured acceleration, then gets plotted against its matched mean velocity. A linear regression through those five points produces the relationship F = F0 (1 − v/v0), where F0 is the theoretical maximal horizontal force at zero velocity and v0 is the theoretical maximal velocity at zero resistance. Maximal mechanical power, Pmax, follows as F0 × v0 ÷ 4, and it occurs at exactly half of v0.

A worked example: an 85 kg athlete pushes a 17 kg sled (20% BM) at a mean velocity of 3.1 m/s, a 34 kg sled (40% BM) at 2.6 m/s, a 51 kg sled (60% BM) at 2.0 m/s, a 68 kg sled (80% BM) at 1.5 m/s, and an 85 kg sled (100% BM) at 1.1 m/s. Regressing those five force-velocity pairs yields roughly F0 ≈ 540 N and v0 ≈ 4.1 m/s, giving Pmax ≈ 553 W at a velocity of about 2.05 m/s — which lands close to the 60% BM load in this athlete, not the 40% or 80% load a generic table might have assigned. That gap between the calculated power-peak load and whatever the athlete happened to be training at is usually the single most useful number to come out of the test.

A test is only useful for tracking change if repeating it under the same conditions gives you a stable number. Cross et al. (2017) reported that F0, v0, and Pmax derived from a multi-load sled push protocol were highly correlated with direct force-plate measures across their rugby cohort, supporting the regression method as a valid substitute for lab equipment in a field setting. Reliability work on youth sled push profiling, including Cahill et al. (2020, Sports), reported good-to-excellent test-retest agreement for the same three metrics across separate sessions in a small squad of adolescent rugby league players, with the lighter-load velocity measures consistently more stable than the heaviest-load trials.

The practical takeaway from the reliability data is where your error bars actually sit: a session-to-session Pmax shift under roughly 8% is close to measurement noise, while anything larger is more likely to reflect a genuine change in the athlete's underlying capacity, provided surface, sled, and warm-up were held constant.

F0, v0, and Pmax describe three different things, and training decisions change depending on which one is limiting an athlete. A low F0 relative to body mass, paired with a comparatively high v0, marks a force-deficient profile — this athlete moves light sleds well but stalls quickly as load climbs, and heavy sled work (80-120% BM) is the higher-value zone. The reverse pattern, strong F0 but low v0, marks a velocity-deficient profile better served by lighter loads (20-40% BM) trained for speed of movement rather than raw resistance.

Pmax and its associated load are the most direct prescription number: training within roughly ±10% of the calculated power-peak load, alternated with occasional heavy and light exposures to keep the full curve stimulated, matches the broader force-velocity profiling approach Morin and colleagues have supported across sprinting and resisted-sprint work. A profile isn't a permanent label — retesting after a block should show F0, v0, or both shifting in the direction the program targeted. If they don't, that's a signal the loading zone needs to change before the next block, not after it.

A handful of recurring mistakes turn an otherwise sound protocol into noisy, untrustworthy data.

• Testing only two or three loads. Fewer points make the regression hypersensitive to any single off trial, and a bad heavy-load rep can swing the calculated Pmax load by 15% or more.
• Skipping the calibration push. A sensor reading noisy at light load won't magically clean up at heavy load — check the trace before committing to maximal effort.
• Changing surface mid-cycle. Moving from turf to indoor rubber between sessions can shift velocity at a given load by 10-25%, which looks like a real fitness change on paper and isn't.
• Under-resting between heavy trials. A tired athlete's second heavy-load push reflects fatigue, not true maximal capacity, and drags the whole regression down.
• Rounding load to the nearest available plate. Jumping from 60% to 85% BM because that's what the plates allow leaves a gap right where the power peak often sits.

Retesting too often wastes training time on assessment; retesting too rarely means training an outdated prescription for weeks after it stopped being optimal. A full five-load retest every 4-6 weeks, aligned with the end of a training block, fits most programs without eating into development time. In-season, a shortened three-load check (light, moderate, heavy) every 3 weeks is usually enough to confirm the profile hasn't drifted meaningfully.

Given the roughly 8% session-to-session variability reported for Pmax under controlled conditions, treat any single retest showing a 5-8% shift as a trend worth watching rather than a confirmed change, and treat anything beyond that — particularly if it holds up on a second check — as real. A rising Pmax load with a proportionally rising v0 typically reflects genuine speed-strength development; a rising Pmax load driven mostly by F0 without a matching v0 shift more often reflects a strength gain that hasn't yet transferred to sled-push velocity.

Key References

• Cross, M. R., Brughelli, M., Samozino, P., & Morin, J.-B. (2017). Methods of power-force-velocity profiling during sled-resisted sprinting. Int J Sports Physiol Perform, 12(8), 1069-1077.
• Pareja-Blanco, F. et al. (2017). Effects of velocity loss during resistance training on athletic performance, strength gains and muscle adaptations. Scand J Med Sci Sports, 27(7), 724-735.
• Cahill, M. J. et al. (2020). Sled-push load-velocity profiling in young athletes. Sports, 8(9).