Wheelchair Basketball Propulsion Power Test: A Rim-Torque Protocol for the 5m Sprint Start

A coach lines up two players for the same 5m sprint test: a class 1.5 player with limited trunk control, and a class 4.0 player with close to full trunk rotation. Both cross the line at 1.9 seconds. The spreadsheet records identical numbers, and on paper the two performances look the same. What it does not show is that the class 4.0 player produced that time with a powerful, trunk-assisted opening push, while the class 1.5 player recruited nearly everything available in the shoulders and arms just to match it, with far less left in reserve for the fourth quarter. Treating those two splits as equivalent is a common blind spot in wheelchair basketball testing, and an easy one to fix once you measure what actually produced the number.

The fix is not a faster stopwatch. It is measuring the torque applied at the push-rim during the first few strokes off a dead stop, then reading that torque against what a player's own functional classification would predict, rather than one fixed target for the whole roster. The protocol below builds that measurement from equipment most programs already own, walks through a standardized 5m stationary start, and shows the arithmetic that turns a velocity trace into a torque and power number two classification categories can be compared against on their own terms.

Why the Same Split Time Can Hide Very Different Propulsion Power

Wheelchair basketball classifies players from 1.0 to 4.5 points, built almost entirely around how much trunk and hip function a player retains. A 1.0 or 1.5 point player has little to no active trunk control and stabilizes against the backrest, so nearly all propulsive torque comes from the shoulders and arms. A 4.0 or 4.5 point player can flex and rotate the trunk through most of the push phase, adding a second, substantial torque source. Villacieros, Pérez-Tejero, Garrido, Grams, López-Illescas, and Ferro (2020) tested twelve elite male players split into Category A (classes 1.0-2.5) and Category B (classes 3.0-4.5) across sprint and ball-handling drills, and found a significant relationship between peak isokinetic moment at the shoulder and elbow and maximum sprint velocity (p < 0.05). Shoulder and elbow torque were not a side note to the sprint number; they were a direct input into it.

Vanlandewijck, Verellen, and Tweedy (2011) showed the seating side of the same mechanism. Fifteen participants sprinted on a wheelchair ergometer in three seating positions that progressively restricted trunk range of movement, and the most restricted position produced a drop in acceleration capacity that was both statistically and practically significant against the least restricted one. Two players seated in chairs that both look correctly configured can still start from meaningfully different torque potential, for reasons that have nothing to do with effort.

Equipment and Chair Setup

The test needs a standardized 5m lane, a way to capture velocity from a dead stop, and the athlete's own chair with its geometry recorded rather than assumed.

Chair setup is the variable most programs forget to log. Camber, seat height, and rim radius all change how a push translates into torque, so measure push-rim and wheel radius on each athlete's own chair rather than a manufacturer default, and re-measure after any adjustment.

Step-by-Step Testing Protocol

Total time per athlete, warm-up included, runs about 10-12 minutes once the lane and sensor are set.

1. Log the baseline numbers: classification class, push-rim radius, wheel radius, and total system mass (athlete plus chair), before the first trial.
2. Warm-up (8-10 minutes): easy propulsion, shoulder circles, then two submaximal 5m efforts building to roughly 85% effort.
3. Standardize the start: front casters behind a fixed line, both hands on the push-rim at a consistent clock position, no rocking or rolling start before the signal.
4. Familiarization trial: one submaximal start-to-5m effort to confirm hand placement, not scored.
5. Maximal trials: three dead-stop starts to the 5m line at full effort, 2-3 minutes of rest between each.
6. Capture the trace: record velocity through the 5m line via a wheel-mounted IMU or frame-by-frame video, marking the first three push cycles.
7. Valid trial criteria: discard any trial with a rolling start, an inconsistent hand position, or a caster past the line at the signal.
8. Scoring: use the shortest valid 5m time for the split, and that same trial's trace for the torque and power calculation below.

From a Velocity Trace to a Rim Torque Number

You do not need an instrumented push-rim to estimate propulsive torque. Basic kinematics gets you there, at the cost of some precision the estimate is upfront about.

Acceleration during the opening push phase: a = Δv / Δt, from the dead stop (v = 0) through the end of the third push cycle.

Net propulsive force: F = m × a, where m is total system mass (athlete plus chair). This treats the whole net force as propulsive, slightly overstating true push-rim force since some is lost to rolling resistance and bearing friction, an acceptable simplification for tracking one athlete over time rather than matching a lab force plate.

Torque at the push-rim: T = F × r_rim. Angular velocity is ω = v / r_wheel, and instantaneous power is P = T × ω. Average power over the push window is the change in kinetic energy (½ × m × v²) divided by elapsed time.

Worked example: an 80kg total system (70kg athlete, 10kg chair) reaches 3.2 m/s over the first three push cycles in 0.4 seconds. Acceleration is 3.2 / 0.4 = 8.0 m/s². Net force is 80 × 8.0 = 640N. With a 0.26m push-rim radius, torque is 640 × 0.26 = 166.4 Nm. With a 0.33m wheel radius, angular velocity is 3.2 / 0.33 = 9.7 rad/s, giving instantaneous power near 1,614W and average power across the window of roughly 1,024W. Normalized to total mass, that is about 2.08 Nm/kg of torque and 12.8 W/kg of average power, the two numbers to carry into the classification bands below.

What the Research Actually Shows

The Villacieros et al. (2020) findings above came with a limitation the authors were upfront about: twelve players is workable for a correlational isokinetic study but not enough to generate population-wide torque norms, and the relationship they found was between peak isokinetic moment and sprint velocity, not a direct field measurement of push-rim torque during a dead-stop start like this one. Their result supports the classification-based comparison used here; it does not hand over a ready-made cutoff table.

The Vanlandewijck et al. (2011) seating study carries a different limitation worth flagging: the fifteen participants were non-disabled volunteers modeling seating effects, not wheelchair basketball athletes with the classification-defining impairments the sport actually tests for. The direction of the seating effect on acceleration is well supported; the magnitude in an athlete with a genuine trunk impairment may differ.

A more recent field study adds a piece this protocol borrows directly. Brassart, Bakatchina, Alberca, Pomarat, Watelain, Weissland, and Faupin (2025), in Frontiers in Sports and Active Living, had thirteen wheelchair basketball and ten wheelchair rugby players from French national squads run six repeated 20m sprints, using wheel-mounted IMUs and an Instantaneous Symmetry Index to quantify left-right propulsion asymmetry. Asymmetry peaked at the start of each sprint rather than mid-effort, basketball players ran higher asymmetry than rugby players, and peak power output dropped across the repeated sprints. That points to the same acceleration phase this protocol isolates as where technique breaks down first, and is a reason to log left-right torque balance once the basic protocol is running.

Adaptive Bands: Reading Torque by Classification Category

Because trunk function changes how much torque a player's body can produce independent of training, comparing a Category A player's raw number against a Category B teammate's rewards classification, not preparation. The bands below split at the same Category A / Category B line used in the Villacieros research, expressed as average propulsive power per kilogram of total system mass over the first three push cycles.

Treat these as a starting point, not a validated normative table. No published dataset reports population-wide W/kg norms by classification category for this exact protocol, so the ranges are assembled from magnitudes typical of the wider handrim propulsion literature rather than one citable source. Rebuild the bands from your own squad's data once a season is logged, and use the categories to flag a player well below their own group rather than to make roster decisions alone.

Mistakes That Skew the Torque Number

Turning a Low Torque Score Into a Training Target

A player in the underdeveloped band for their category usually has a specific, fixable gap rather than a general fitness problem. Category A players tend to respond well to shoulder and lat strength work aimed at arm-only torque, since trunk assistance is not available: resisted start pulls, seated med-ball throws, short-lever rowing. Category B players with a low score more often have a timing gap between trunk swing and push rather than a strength deficit, and respond faster to video-cued start drills than to more gym work.

If the flag is asymmetry rather than a low overall score, remember the Brassart et al. (2025) pattern: it shows up hardest at the start, so check left-right torque balance on the opening push before assuming general fatigue. Retest every 4-6 weeks; torque and power move slower than a stopwatch split, and testing sooner mostly adds noise rather than signal.