ECSS Paris 2023: CP-AP08
INTRODUCTION: Saber is the fastest weapon among the three fencing disciplines. Literature has shown that male players score most frequently using attacks, but there is no consensus regarding specific scoring target areas (1). Moreover, there is little understanding about the impact of score status (leading, even or trailing) and winning the 1st round on the outcome. METHODS: This study analyzed 120 men’s saber matches from the Budapest World Cup (Hungary), Orléans Grand Prix (France), and Seoul Grand Prix (South Korea) from 2021 and 2024, using notational analysis. We examined the relationship between techniques (used and scored) and match outcomes (in winners and losers), and further in different score status and target areas. RESULTS: Winners were ahead in the 1st round nearly 80% of the matches. Both winners and losers used attacks most frequently, followed by counter-attacks. However, winners had a higher proportion of attack compared to losers, while losers counterattacked more. In terms of techniques scored, both winners and losers scored most often with attacks, followed by ripostes. Both winners and losers demonstrated association between techniques, both used and scored, with score status, with both expanding their proportion of techniques scored compared to techniques used in leading status. However, winners exhibited a significantly higher proportion of leading status, contributing to the difference in win/loss outcomes. Notably, only winners could effectively perform ripostes while leading, highlighting their superior defensive capabilities. Both groups predominantly targeted the head and the non-dominant side’s outer area. For attacks, the dominant-side inner area and non-dominant-side outer area were targeted. Ripostes targeted the dominant-side upper arm and non-dominant-side arm, while counterattacks targeted the dominant-side forearm. Losers exhibited a more dispersed distribution in ripostes and counterattacks, showing a lack of optimal adjustment. CONCLUSION: The results of this study reveal that world-class saber fencers adjust their techniques based on score status and preferring specific target areas. These findings provide valuable insights for coaches to design situational training and offer tactical guidance during competitions.
Read CV Liu Ting YuECSS Paris 2023: CP-AP08
INTRODUCTION: In recent years, the integration of Artificial Intelligence (AI), Internet of Things (IoT), and sensor technologies has significantly advanced sports analytics. For instance, in team sports like soccer, players often wear sensors equipped with GPS and accelerometers to comprehensively measure movement metrics, such as distance covered and overall exercise intensity during practices and games. By analyzing tracking data from soccer matches, new insights have been uncovered, particularly in predicting injury prevention. This study focuses on visualizing the offensive and defensive tactics and team strategies of Japanese university elite soccer teams using network analysis. This approach is effective in identifying the characteristics of a teams strategic planning and quantifying the contributions and interactions of individual players. METHODS: We analyzed a top league match of domestic university soccer using an acceleration sensor 1, and applied the results to a unique algorithm to analyze and evaluate the network (player coordination). We investigated the changes in the acceleration of the players, derived random variables, and evaluated the entropy of the movement through statistical processing. This index is an index of synchronization between players, and sheds light on the characteristics of team performance (offense and defense) during a match. [1] Magome, T. et.al. The development of automatic training analysis using 3D accelerometer in male young elite soccer team, In Proceedings of the 12th (icSPORTS 2024), pp. 160–167. RESULTS: We analyzed official matches between university teams of comparable skill levels, obtained acceleration data, and performed network analysis. We statistically analyzed the coordination between each player using phase shift entropy. We found that in matches where team tactics were more effective, players coordination was better, and there was a significant difference in players coordination between periods when organized attacks were successful and periods when they were less effective (p>0.05). Furthermore, our results suggest that network assessment may be able to assess the performance and injury risk of individuals within a team structure. CONCLUSION: In this study, we were able to evaluate match performance based on the cooperation between players, visualize the quality of tactics, and obtain insights that can lead to injury prediction. These, combined with the data of individual players, can lead to an understanding of team performance. However, it is difficult to analyze all tactics in detail. Therefore, we aim to work with teams and players to visualize specific data that is key to offense and defense, and provide feedback to the entire team. In the future, we will explore the value of quantifying team performance to confirm and optimize high-quality movements. These are likely to have a ripple effect throughout the sports world and are useful.
Read CV takuya magomeECSS Paris 2023: CP-AP08
INTRODUCTION: Elite swimming performance is determined by a complex interplay of anthropometric, physiological, biomechanical and technical factors. While previous research emphasises the role of explosive power, physiological conditioning and technical precision in 100m Freestyle, ‘in-race’ factors discriminating World-class from international and national-level swimmers has not been examined. This study uniquely examined race-phase metrics associated with World-National level performance. METHODS: High-definition race-analysis software examined race performances of 204 male swimmers recorded at World-National level competitions (2019–2024). Race indices including stroke rate, stroke length, turn efficiency, underwater durations, and 5m interval velocities were extracted. Random forest models and feature importance methods distinguishing World- (0–2.5%), International- (2.5–5%), and National-level (5–10%) swimmers based on proximity to world record performance. RESULTS: Random forest modelling = 89.5% classification accuracy. Swimmer velocities at 65-70m, 70-75m & 60-65m were the most discriminatory features (relative importance = 0.086, 0.081, 0.053). Shapley Additive exPlanations (SHAP) and Partial Dependence Plots validated mid-race velocity importance as distinguishing factors across performance levels. CONCLUSION: Given the unique analysis approach, results overall identified how mid-race speed maintenance differentiated World-National level 100m. Findings offer practical implications for performance analysts and coaches, suggesting benefit from assessing race-phase velocity profiles and for targeted training within the 100m event.
Read CV Giovanni Luca PostiglioneECSS Paris 2023: CP-AP08