Positional Differences in Pre-Season Scrimmage Performance of Division I Collegiate Football Players

This study aimed to describe the physical demands of American football players using novel performance analysis techniques. Heart rate (HR) and accelerometer-based activity levels were observed across two pre-season scrimmages in 23 Division I collegiate football players (age: 19 ± 1 y, height: 1.90 ± 0.06 m, weight: 116.2 ± 19.4 kg). Data were analyzed using a MATLAB program and inter-rater reproducibility assessed using inter-class correlations (ICC). Players were analyzed by side (offense/defense) and position (skill/non-skill). Performance variables assessed in bursts of activity included burst duration, HR_mean and HR_max (bpm), and mean activity (vector magnitude units [vmu]). Exercise intensity was categorized as time spent in % HRmax in 5% increments. The burst duration (8.1±3.9 min, ICC = 0.72), HR_mean (157 ± 12 bpm, ICC = 0.96) and mean activity (0.30 ± 0.05 vmu, ICC = 0.86) were reproducible. HR_mean (p = 0.05) and HR_max (p = 0.001) were greater on defense. Offense spent more time at 65-70% HR_max (p = 0.01), 70-75% HR_max (p = 0.02) while defense spent more time 90-95% HR_max and ≥95% HR_max (p = 0.03). HR_mean (p = 0.70) and HR_peak (p = 0.80) were not different between positions across both sides. Skilled players demonstrated greater mean activity (p = 0.02). The sport-specific analysis described HR and activity level in a reproducible manner. Automated methods of assessing HR may be useful in training and game time performance but ultimately provides support to coaching decision making.