<p>This study investigated whether match outcome influences external training load in a rolling three-team small-sided game (SSG) format in semi-professional male football players. Twenty outfield players contributed valid GPS data across five training sessions consisting of repeated 6v6 rolling SSG bouts, where the winning team remained on the pitch and the losing team rotated out. External load variables, total distance (TD), accelerations (ACC), decelerations (DEC), and high-speed running (HSR) were measured using 10&#xa0;Hz GPS and analysed using repeated-measures ANOVA with Bonferroni corrections, considering both absolute and time-normalised (per-minute) values. Accumulated total distance (TD; F = 18.04, <i>p</i> &lt; .001, η²p = .52), accelerations (ACC; F = 6.05, <i>p</i> = .006, η²p = .27), and decelerations (DEC; F = 8.11, <i>p</i> = .001, η²p = .34) differed significantly between match outcomes, with values decreasing from win to draw to loss conditions. Accumulated high-speed running (HSR) did not reach statistical significance (F = 3.74, <i>p</i> = .055, η²p = .19). No significant effects were observed for any time-normalised variable (TD: F = 1.57, <i>p</i> = .223, η²p = .09; ACC: F = 0.49, <i>p</i> = .620, η²p = .03; DEC: F = 1.72, <i>p</i> = .196, η²p = .10; HSR: F = 0.45, <i>p</i> = .642, η²p = .03). These findings indicate that match outcome was associated with accumulated, but not time-normalised, external load in this rolling SSG format. The winner-stays rule should therefore be considered when interpreting accumulated external load in such training formats.</p>