
Although data and digital analytics are increasingly used in sport, their implications for coaching decision-making and expertise remain insufficiently understood. This study examined how data and digital analytics were incorporated into coaching judgement and decision-making and what these processes meant for the reconfiguration of coaching expertise. Semi-structured interviews were conducted with four coaches, three data and analytics specialists, and three athletes with experience in data-informed coaching in elite and professional sport. Data were analysed using reflexive thematic analysis. The findings showed that data did not replace experiential judgement but extended what could be perceived and traced over time, while providing additional grounds on which existing assessments could be reviewed. However, greater access to information did not make coaching decisions self-evident or necessarily more objective. Participants interpreted data's meaning and limitations contextually and selectively weighted diverse forms of evidence. Interpretive work and epistemic authority were distributed among coaches, analysts, and athletes, rendering decision-making more negotiated, whereas authority and accountability for final decisions remained largely with coaches. Taken together, the findings indicated that coaching expertise was being reconfigured around the capacity to orchestrate distributed knowledge, explain, justify, and revise professional judgements, and remain accountable for decisions made under uncertainty. These findings suggest that coach education should extend beyond technical data literacy to develop capacities for contextual interpretation, knowledge orchestration, and accountable professional judgement.
This study aimed to identify multidimensional kinematic differences among male 100-m sprinters during the 50–56 m section near maximal velocity and to inform technical assessment. High-speed videos from three competitions yielded 161 race observations from 147 sprinters classified as Level 3, Level 2, Level 1, or Elite. One stride cycle was recorded at 240 Hz. Primary outcomes were analysed using mixed-effects models, while secondary outcomes were analysed using mixed-effects, generalised least-squares, or ordinary linear models as appropriate, with Benjamini–Hochberg corrections applied within domains and globally. Performance level significantly affected step frequency and step length [F(3, 120.5) = 20.482 and F(3, 11) = 15.690; both Holm-adjusted p < 0.001]. Step frequency generally increased with performance level, whereas step length was lower in Level 3 than in the other groups, with no differences among Level 2, Level 1, and Elite. Twenty-seven secondary outcomes remained significant after false discovery rate correction. Compared with Elite sprinters, Level 3 sprinters showed longer support-related durations, greater touchdown distance, and distinct swing-leg, joint-posture, and centre-of-mass displacement patterns. Sensitivity analyses generally supported the stability of these findings. Overall, performance-level differences near maximal velocity were multidimensional, supporting individualised, multivariable technical assessment rather than reliance on a single universal kinematic target.
This study examined the GPS-derived match physical outputs of youth soccer players promoted to compete in a higher age category. A retrospective observational study was conducted using match data collected over three seasons (2022–23 to 2024–25) from an elite youth soccer academy (n = 221). Three questions were examined: (i) whether promoted players differed from their regular teammates when competing in their original age category; (ii) whether they differed from the regular players of the higher age category when competing alongside them; and (iii) whether the same promoted players changed their physical outputs between the two age categories. External load variables included relative total distance, high-speed running, sprint distance, high intensity accelerations and decelerations. In their original age category, promoted players did not consistently differ from their regular teammates, although promoted Under-19 midfielders covered greater high-speed running distance. In the higher age category, promoted players generally showed physical outputs similar to those of the regular players of that category, although promoted Under-19 wingers covered greater sprint distance than regular Under-23 wingers. Within-player comparisons, often based on very few players, revealed position- and category-specific changes, with significant effects for decelerations in Under-16 and Under-19 centre-backs, high-speed running and sprint distance in Under-17 full-backs and strikers, and total distance in Under-19 strikers. These findings suggest that promotion is not associated with a consistent change in match physical outputs, but rather with specific changes according to age category, playing position, and physical variable, supporting individualized approaches to player development.
This study compared physiological and neuromuscular responses elicited by the Carminatti's Test (T-CAR), the 30–15 Intermittent Fitness Test (30–15 IFT ), and the Yo-Yo Intermittent Recovery Test Level 1 (Yo-Yo IR1 ). Twenty-six collegiate male soccer players completed the three shuttle-run protocols over three weeks. Peak speed (PS), peak blood lactate concentration (peak [La]), peak heart rate (peak HR), and peak rating of perceived exertion (peak RPE) were compared among the protocols. Additionally, countermovement jump (CMJ) height and 10-m and 30-m sprint speeds were assessed before and after each protocol. The results suggest that the 30–15 IFT elicited higher values of PS and peak [La] (18.8 ± 1.4 km·h −1 ; 12.4 ± 4.1 mmol·L −1 , respectively) than T-CAR (15.6 ± 1.3 km·h −1 ; 9.7 ± 2.5 mmol·L −1 ) and Yo-Yo IR1 (15.7 ± 0.7 km·h −1 ; 9.3 ± 2.4 mmol·L −1 ) (p < 0.001). No significant differences were observed for peak HR (p = 0.390) or peak RPE (p = 0.961) across protocols. A significant main effect of time was found for CMJ height (p < 0.001), but not for sprint performance. Larger pre-to-post absolute and relative reductions in CMJ height were observed following T-CAR (Δ = –2.5 cm; −7.0%) and 30–15 IFT (Δ = –2.2 cm; −5.6%) compared with Yo-Yo IR1 (Δ = –0.9 cm; −2.2%). In conclusion, the 30-15 IFT enables players to reach higher PS and peak [La] values than both T-CAR and Yo-Yo IR1 . Moreover, T-CAR and 30-15 IFT appear to induce greater acute impairment of lower-limb muscle power than Yo-Yo IR1 , while sprint performance remains preserved regardless of the test employed.
Identifying and quantifying the different playing styles used by African youth soccer teams is important for objectively describing tactical approaches and team behaviours. This study aimed to identify and analyse team-match playing-style profiles during the Africa Cup of Nations (AFCON) Under-17 (U17) and Under-20 (U20) soccer tournaments. The AFCON U17 sample included 68 team-match observations from 34 matches, and the AFCON U20 sample included 60 team-match observations from 30 matches. Data were obtained from the Wyscout platform. Data were analysed using descriptive statistics, multiple factor analysis, Ward.D2 hierarchical clustering, and outcome-appropriate regression models with standard errors clustered by team and match. The pooled analysis identified four team-match profiles: (1) limited progression, (2) controlled possession and progression, (3) balanced counterattacking, and (4) direct high-transition duel. Pooled Clusters 2 and 3 had higher estimated goal rates than Clusters 1 and 4 after adjustment for multiple comparisons, and U20 teams had a lower estimated goal rate than U17 teams. No pairwise differences in winning odds between clusters remained significant after adjustment, and no robust association was identified between cluster membership and goals conceded. The findings, therefore, provide stronger evidence for an association with goals scored than with match outcome.