**Background:** Traditional time-motion analysis in soccer uses speed-based metrics like total distance and distance in speed zones to quantify external load, but these parameters do not account for fluctuations in movement speed that directly contribute to energy cost. Accelerating is more energetically demanding than maintaining speed, yet conventional approaches treat acceleration separately from speed. The metabolic power model, which considers both speed and acceleration to estimate instantaneous energy requirements, may provide a more accurate assessment of high-intensity demands in intermittent team sports. Previous metabolic power research in top-level soccer was limited to the Italian Serie A (Osgnach et al., 2010), and no recent seasons or other top leagues had been analyzed despite evidence that match intensity has increased over the last decade.
**Methods:** Data were collected from three German Bundesliga teams and their opponents across the 2016/17 season using a semi-automatic camera match analysis system (25 Hz) permanently installed in all league stadiums. The system demonstrated high reliability (ICC ≥ 0.98) for metabolic power data and good validity for soccer-specific movements. A total of 1,345 discrete player data sets from 96 matches were gathered from 380 players. Players were grouped into six positions: wing-back (WB; 68 players, 230 matches), center-back (CB; 82 players, 394 matches), wide midfielder (WM; 88 players, 226 matches), central midfielder (CM; 79 players, 291 matches), striker (ST; 42 players, 158 matches), and central-attacking midfielder (CAM; 21 players, 46 matches). Goalkeepers and players who did not play the entire match were excluded. Speed, acceleration, and metabolic power variables were categorized using established thresholds (high speed >15.5 km/h, high acceleration >2 m/s², high metabolic power >20 W/kg). The constant for running on grassy terrain (KT) was set at 1.29. Statistical analysis used one-way ANOVA with Bonferroni post-hoc tests, partial η² for effect sizes, and paired t-tests for half-time comparisons with Cohen's d.
**Key Results:** Mean playing time was 93:55 ± 2:13 min:s with net playing time of 58:28 ± 4:12 min:s. CM and CAM had the highest values across nearly all parameters: total distance (CM: 11,445 ± 638 m; CAM: 11,494 ± 765 m), energy expenditure (CM: 59.61 ± 3.61 kJ/kg; CAM: 59.81 ± 4.21 kJ/kg), average speed (CM: 2.03 ± 0.12 m/s; CAM: 2.04 ± 0.15 m/s), average metabolic power (CM: 10.90 ± 0.62 W/kg; CAM: 10.93 ± 0.81 W/kg), and equivalent distance (CM: 13,223 ± 778 m; CAM: 13,271 ± 907 m). CB showed the lowest values for each parameter (total distance: 9,755 ± 615 m; energy expenditure: 51.65 ± 3.59 kJ/kg; average speed: 1.82 ± 0.16 m/s; average metabolic power: 9.27 ± 0.63 W/kg). The highest EDI was recorded for WM (1.17 ± 0.02), indicating more intermittent activity, while CB had the lowest anaerobic index (0.32 ± 0.04). For high-intensity classification, the metabolic power approach identified substantially more activity than speed-based analysis across all positions: 48.9% more time (10:17 ± 2:21 min:s vs. 5:02 ± 1:24 min:s), 66.5% more distance (2,464 ± 597 m vs. 1,638 ± 458 m), and 50.4% more energy (19.96 ± 4.35 kJ/kg vs. 10.05 ± 2.87 kJ/kg) at high intensity (all p < 0.001, η² = 0.92–0.94, large effect). Comparing halves, mean metabolic power (10.36 ± 1.03 vs. 9.59 ± 0.97 W/kg, d = 1.24, strong effect), total distance (5,260 ± 296 m vs. 5,127 ± 361 m, d = 0.04, no effect), total energy expenditure (28.39 ± 3.30 vs. 27.11 ± 4.29 kJ/kg, d = 0.48, small effect), and high-intensity energy expenditure (10.47 ± 2.23 vs. 9.52 ± 2.35 kJ/kg, d = 0.62, medium effect) were all higher in the first half.
**Clinical Implications:** The findings demonstrate that traditional speed-based analysis substantially underestimates high-intensity demands in soccer, as it fails to capture the energetic cost of accelerations that occur at submaximal speeds. The metabolic power approach provides a more comprehensive assessment of external load and should be incorporated into match analysis and training monitoring. Position-specific differences in metabolic power parameters confirm the need for tailored athletic training programs, particularly for central midfielders who exhibit the highest volume and intensity demands and center-backs who show more steady-state activity. The significant decline in metabolic power and energy expenditure from first to second half suggests fatigue occurs as matches progress, which may inform substitution strategies and conditioning programs. However, the metabolic power model has limitations—it does not account for vertical movements (jumping), limb movements (passing, shooting), or soccer-specific actions, and individual anaerobic thresholds should be used when available for more precise analysis.