Field-Based Assessment of Running Biomechanics Using Global Positioning System/Inertial Measurement Unit: Establishing Construct Validity in Elite Athletes
Journal of strength and conditioning research · 2026
GPS/IMU running-mechanics metrics are reliable and internally consistent in elite AFL players, but aren't yet validated against real biomechanics.
Part of: Training load and injury risk: what the evidence says
The paper
Methodological/reliability study, n=54 elite male AFL players, 3,943 repeated 40-m field running trials across one competitive season.
What they found
Velocity-adjusted between-session reliability was moderate to good-to-excellent across 12 GPS/IMU biomechanical variables (ICC = 0.65-0.92, 8 of 12 reaching ICC ≥0.75); a MARS model predicted running speed from these variables with R2 = 0.93 (RMSE = 0.20 m·s-1, ICC = 0.96), with negligible Bland-Altman bias (-0.01 m·s-1, LoA -0.243 to +0.223 m·s-1) and no proportional bias (p = 0.369).
Study design
A single-cohort, prospective observational reliability and construct-validity study of a wearable measurement system, not a clinical intervention trial, using n=54 professional AFL players and 3,943 trials (reliability subsample n=25, mean k=2.95).
Methodology
Strengths are the large trial volume, athlete-cluster bootstrap CIs, and athlete-stratified cross-validation guarding against overfitting when choosing among five candidate models. But 'construct validity' here means predicting speed from the same device-derived variables, not comparison against an independent criterion standard (e.g. 3D motion capture or force plates), so this is closer to internal consistency than true criterion validity; the authors state this themselves. Single sport, single sex, single team also limits generalisability, and funding/affiliation with the AFL club isn't disclosed in the abstract.
The appraisal
The reliability statistics are statistically solid and some variables clear clinically useful thresholds (ICC ≥0.75, MDC reported), but the headline R2 = 0.93 reflects how well the variables predict speed, which is a circular validation rather than proof the device captures true biomechanics; until it's checked against a lab-based gold standard, it is better read as 'this tool is internally consistent' rather than 'this tool measures real running mechanics accurately'.
The gap
No criterion validation against 3D motion capture or force-plate biomechanics, which the authors explicitly flag as pending; without that, it's unclear whether the variables reflect true running mechanics or just a reproducible proxy that happens to correlate with speed. It's also device- and population-specific: all data come from one commercial GPS/IMU system in elite male AFL players sprinting 20-40 m near maximal velocity, so neither the reliability nor the validity figures should be assumed to transfer to other brands or algorithms, female or youth athletes, or fatigued, match-realistic running.
Landmark context
This extends the broader GPS/IMU wearable-reliability literature (e.g. work validating GPS for distance/speed/acceleration monitoring in team sports) into the newer space of extracting kinematic/biomechanical variables from field wearables as a cheaper alternative to lab-based 3D motion capture, which remains the actual gold standard this paper has not yet tested against.
What to do Monday
Not yet actionable for individual technique or injury-risk decisions. It supports continued use of GPS/IMU units for tracking within-athlete trends in max-velocity sprint metrics across a season in elite, AFL-type running, but these figures are specific to one commercial system in elite male athletes doing 20-40 m maximal sprints, so don't assume they transfer to other brands, female or youth athletes, or fatigued match-running. For now, keep using clinical gait assessment, video, or lab/force-plate testing, not these derived 'mechanics' scores, when the decision is about an individual's technique or injury risk, until criterion validation against motion capture is published.
In practice
This is a monitoring-signal paper, not a technique-diagnosis paper: on the S&C floor, keep using the GPS/IMU unit's speed, load and HSR data as you already do, and treat any derived 'biomechanical' variables as a reliable number to trend within an individual across a season, not a validated technique score to rank athletes or make selection/technique calls on. In the clinic, for a hamstring or calf return-to-sprint pathway, don't substitute these device metrics for your usual sprint video review, clinical gait assessment, or force-plate/hop testing; use the GPS/IMU data for load progression and keep the technique call on direct observation until criterion validation against motion capture exists. Caveat: this is one commercial system in elite male AFL players doing flat 20-40 m maximal sprints, so be cautious extrapolating to other brands, female or youth athletes, or fatigued match-play running where mechanics likely diverge.
The five papers that matter — every week, free.
We read the whole sports-medicine, rehab & performance literature and appraise the papers worth your time.
Get the free weekly issueRelated appraisals
Read in your language: English · Français · Deutsch · Español · Italiano · Nederlands · Dansk · Svenska