Objectives: To compare fixed epochs (FIXED) and rolling averages (ROLL) for quantifying worst-case scenario (‘peak’) running demands during professional soccer match-play, whilst assessing contextual influences. Design: Descriptive, observational. Methods: Twenty-five outfield players from an English Championship soccer club wore 10-Hz microelectromechanical systems during 28 matches. Relative total and high-speed (>5.5 m∙s-1) distances were averaged over fixed and rolling 60-s to 600-s epochs. Linear mixed models compared FIXED versus ROLL and assessed the influence of epoch length, playing position, starting status, match result, location, formation, and time-of-day. Results: Irrespective of playing position or epoch duration, FIXED underestimated ROLL for total (~7-10%) and high-speed (~12-25%) distance. In ROLL, worst-case scenario relative total and high-speed distances reduced from 190.1±20.4 m∙min-1 and 59.5±23.0 m∙min-1 in the 60-s epoch, to 120.9±13.1 m∙min-1 and 14.2±6.5 m∙min-1 in the 600-s epoch, respectively. Worst-case scenario total distance was higher for midfielders (~9-16 m∙min-1) and defenders (~3-10 m∙min-1) compared with attackers. In general, starters experienced higher worst-case scenario total distance than substitutes (~3.6-8.5 m∙min-1), but lower worst-case scenario high-speed running over 300-s (~3 m∙min-1). Greater worst-case scenario total and high-speed distances were elicited during wins (~7.3-11.2 m∙min-1 and ~2.7-7.9 m∙min-1, respectively) and losses (~2.7-5.7 m∙min-1 and ~1.4-2.2 m∙min-1, respectively) versus draws, whilst time-of-day and playing formation influenced worst-case scenario high-speed distances only. Conclusions: These data indicate an underestimation of worst-case scenario running demands in FIXED versus ROLL over 60-s to 600-s epochs while highlighting situational influences. Such information facilitates training specificity by enabling sessions to be targeted at the most demanding periods of competition.