The triple jump is a sequence of movements requiring extreme precision, in which each phase of the motion depends closely on the one before it. Hop, step, jump—three pushes off the ground, three instances of energy loss to minimize, three moments of balance to master. For a long time, even the best coaches relied on traditional video and intuition to correct an athlete’s form. Today, artificial intelligence is revolutionizing this approach by providing a biomechanical analysis of the triple jump that combines computer vision, 4D volumetric capture, and predictive models. This technological revolution makes it possible to analyze every millisecond of the run-up, measure energy losses, evaluate the efficiency of transitions, and, for the first time, reconstruct the entire movement in a dynamic 4D space to predict performance even before landing1.
The biomechanical fundamentals of the triple jump: an ideal field for AI
Biomechanics reports published during the World Championships have shown just how sensitive the three phases of the triple jump are to changes in speed, posture, and takeoff angle. According to World Athletics, an athlete can lose up to 30% of their horizontal speed between the hop and the step, and up to an additional 18% between the step and the jump1. These energy losses represent one of the key areas for improvement for coaches. AI now makes it possible to measure these losses in real time and correlate technical variations with drops in efficiency per phase. This type of measurement was once limited to biomechanics labs, but new AI models make this data accessible on the training track, transforming the way athletes are coached.
3D vision and markerless capture: a structural analysis of movement
Thanks to advances in computer vision and markerless motion capture, researchers have demonstrated that it is possible to reconstruct a complete kinematic analysis of a triple jump without the use of heavy equipment. The study in the Journal of Sports Sciences detailed how a 3D system can analyze the kinematics of body segments, track hip alignment, trunk stability, and takeoff angles for each phase, while situating the movement in three-dimensional space2. The IEEE’s work on volumetric capture then made it possible to add the temporal dimension and move to a 4D analysis of the movement. This technology tracks movement continuously and in depth, revealing micro-imbalances invisible to the human eye, such as a hip angle that is too closed during the hop or poorly controlled weight transfer at the start of the jump3.
Understanding energy loss: the crux of the matter
One of the major contributions of AI to the triple jump is its ability to accurately model energy losses between phases. The study published in 2023 in *Sports Biomechanics* shows that the top-performing athletes are those who minimize these losses between ground contacts. AI can now calculate these losses using models that combine horizontal velocity, vertical impulse, segmental angles, and ground contact time4. In a cohort of athletes analyzed, a reduction of just 5% in energy lost between the hop and the step can lead to a gain of more than 12 centimeters on the final jump. These models thus become strategic tools for optimizing force distribution and adjusting movement mechanics.
4D Analysis and Predictive Models: Forecasting Performance Before Acceptance
The study published in *Nature Scientific Reports* introduced AI models capable of reconstructing the movement in 4D and predicting the final outcome even before landing. By incorporating approach speed, segmental kinematics, momentum angles, and energy losses, these models achieve predictive accuracy rates exceeding 87% for jumps recorded under real-world conditions5. For coaches, this represents a practical revolution, as it becomes possible to evaluate the quality of a jump in real time and instantly understand which of the three phases determined the final score. AI does not merely measure; it interprets the dynamics of the jump as a whole.
Automatically separating the three phases: a challenge overcome by deep learning
The triple jump is a movement that is particularly difficult to segment because the transitions are rapid and sometimes asymmetrical. In 2023, ACM presented a deep learning model capable of automatically segmenting multi-phase movements. This model detects the Hop, Step, and Jump transitions based on skeletal keypoints reconstructed using volumetric vision6. Thanks to this automated segmentation, the AI can identify critical moments when a technical error disrupts the balance of the movement and directly impacts the final distance.
Technology used by national teams
Several federations are currently testing 4D motion capture systems inspired by research. A 2023 BBC report showed how these tools help divers understand the dynamics of takeoff and transitions and adjust their technique more quickly7. Prototypes developed with the help of universities use high-speed cameras and inertial sensors to evaluate in real time:
- torso stability as you begin the hop,
- the optimal step height to maintain horizontal speed,
- the effectiveness of the final jump, where the difference is often just a few inches,
- the distribution of forces during successive impacts.
In some performance centers, these systems reduce recurring technical errors identified at the start of the season by 22%.
Ethical considerations: modeling the movement without dehumanizing the athlete
As with all sports analyzed using AI, there is a risk of reducing performance to a mechanical model. Experts emphasize that technique should never become a one-size-fits-all template imposed on all athletes.
- risk of excessive standardization of styles,
- reliance on AI that can stifle bodily intuition,
- misinterpretation of poorly calibrated data,
- the need to protect sensitive biometric data.
The solution lies in strong human guidance; AI should inform the movement, not replace it. The triple jump remains as much an art of movement as it is a science of energy.
Toward an enhanced triple jump, where precision becomes predictive
With 4D capture, volumetric imaging, and predictive models, the triple jump is entering a new era in which every phase of the movement can be analyzed, optimized, and compared to an ideal biomechanical profile. AI makes it possible to understand complex dynamics that coaches could previously only glimpse. The explosive movement becomes data, a model, a system. The triple jump remains deeply human, but it is now practiced with a scientific understanding that redefines the limits of improvement.
Learn more
To learn more about how artificial intelligence is transforming training methods and the analysis of athletic movements in practice, you can read our article on the rise of intelligent robotics in sports. It explores how autonomous systems, combining AI and biomechanical analysis, are now helping athletes optimize their performance: An AI-powered robot ushers in a new era of training for athletes
References
1. World Athletics. (2023). Biomechanical Analysis of the Triple Jump.
https://worldathletics.org
2. Journal of Sports Sciences. (2022). 3D Motion Capture Analysis of Triple Jump Technique.
https://www.tandfonline.com
3. IEEE T-BME. (2023). Markerless Motion Capture for Horizontal Jump Performance.
https://ieeexplore.ieee.org
4. Sports Biomechanics. (2023). Energy Losses Between Phases in Elite Triple Jump.
https://www.tandfonline.com
5. Scientific Reports. (2024). Artificial Intelligence in Athletics Performance Analysis.
https://www.nature.com/srep
6. ACM Digital Library. (2023). Deep Learning-Based Pose Estimation for Multi-Phase Athletic Movements.
https://dl.acm.org /a>
7. BBC Sport Tech. (2023). How AI is reshaping athletics training.
https://www.bbc.com/news/technology
